Computer Science HCI and UI – Assignment

Save Time On Research and Writing
Hire a Pro to Write You a 100% Plagiarism-Free Paper.
Get My Paper

https://www.youtube.com/watch?v=bszqofMRAoo

> Conduct a use case analysis of a business collaboration tool equipped with comprehensive functionalities which might be overwhelming for the employees.

> Your analysis should highlight what would be required to increase the usability of the tool and boost collaborative productivity for new and experienced employees.

Write 6-8 APA-formatted pages and cite at least 6 peer-reviewed articles. Need introduction and conclusion.

Save Time On Research and Writing
Hire a Pro to Write You a 100% Plagiarism-Free Paper.
Get My Paper

CHAPTER

Documentation and User
Support (a.k.a. Help)

•· We learn by example and by direct experience because there

,,

are real limits to the adequacy of verbal instruction .

Malcolm Gladwell

Blink : The Power of Thinking without Thinking, 200

5

•· Our life is frittered away by detail. Simplify, simplify. ”

CHAPTER OUTLINE
14.1 Introduction

14.2 Shaping the Content of the
Documentation

14.J Accessing the Documentation

14.4 Reading from Displays versus
Reading from Paper

14.5 Online Tutorials and Animated
Demonstrations

14.6 Online Communities and

Other

Avenues for User Support

14.7 The Development Process

Thoreau

485

486 Chapter 14 Documentation and User Support (a.k.a. Help)

14. 1 Introduction

Standardization and improvements in user interfaces have made computer
applications easier to use, but using new interfaces is still a challenge. First-time
computer users struggle to understand basic interface icons and actions as well
as their tasks. Even for experienced users, learning advanced features and
understanding novel task domains take commitment and concentration. Many
users learn from someone else who knows the interface; others learn by trial and
error, while others use the supp lied (typically online) documentation; and ye t
others use outside sources, be they online or externally produced books. Over
time, the reliance on formal user manuals, printed documentation, and tutorials
has been less frequent. They are being replaced by customizable requests for
help using the easier- to-use (and usually available) online search facilities. But
some users still want this type of documentation as can be seen in the success of
publishing groups \.Yith series such as For Dumrnies, Missing Manuals, and Teac

h

Yourself Visually.

Learning anything new is a challenge. Challenges can be joyou s and
satisfying, but when it comes to leanung about compu ter systems, many people
experience anxiety, frustration, and disappointment. Much of the difficulty
stems directly from the poor design of the menus, displays, or instructions or
simply from users’ inability to easily determine what to do next. As the goal of
providing universal usability becomes more prominent, online help services are
increasingly necessary to bridge the gap between what users know and what
they need to know.

Studies have show n that welJ-written and well-designed user manuals, on
paper or online, can be effecti, re, bt1t toda y’s users show little iI1terest m detailed
manuals. Today’s user mterfaces are expected to provide everythmg onlme,
supplemented by quick -start gu ides and interactive tutorials that serve user
needs for training and later on for contin ued reference. In fact, as displays
appear in cars, phones, cameras, public kiosks, mobile de, rices, and elsewhere,
ubiquitous and customizab le onlme help should be the norm. Increasing atten­
tion is bemg paid to improvmg user -interface design, but the comp lexity and
diversity of interactive applications are also growing. Evidence suggests that
ther e will always be a need for suppl emental materials that aid users in various
formats (though the use of prmted manuals seems to be fadmg).

There are diverse ways of providmg guidance to users online, rangmg from
a simple pop-up box (often called a tool tip, ScreenTip, or balloon help) to more
advanced assistants and wizards. Most user interfaces have frequently asked
questions (FAQs) and lively user communities that provide a more “grassroots”
type of he lp and support. These communities are often supported or spon­
sored by companies to enable users to solve problems and provide product

14.2 Shaping the Content of the Documentation 487

See also:

Chapter 5, Evaluation and the User Experience

Chapter 8, Fluid Navigation

Chapter 11, Communicat ion and Col laboration

Chapter 15, Information

Search

improvements. This user assistance may be available through formal online and
structured user communities and newsgroups or more informal e-mail, chat,
and instant messaging. A broad variety of formats and styles are used to meet
the ever-present need for documentation and user support (Earle et al., 2015).

This chapter starts by discussing shaping the contents of documentation
including writing for the web (Section 14.2) and then continues with accessing
the documentation (Section 14.3). Section 14.4 explores the differences found
reading from displays versus reading from paper. Next, tutorials and demon­
strations are discussed in Section 14.5, and online communities for user assis ­
tance and support (Section 14.6) are reviewed. The chapter closes with a brief
section on the de, relopment process for user documentation (Section 14.7).

14.2 Shaping the Content of the Documentation

Traditional training and reference materials for computer systems were paper
manuals, frequently left to the most junior member of the development team
as a low-effort task to be comple ted at the end of the project. As a result, the
manuals often were poorly written, were not suited to the users’ backgrounds,
were delayed or incomplete, and were tested inadequately if at all.
Documentation has been referred to as “The Cinderella of Information
Systems” (van Loggem, 2013). Today, managers recognize that designers may
not fully understand the users’ needs, that system developers might not be
good writers, and that it takes time and skill to write effective documentation.
They have also learned that testing and revision must be done before
widespread dissemination and that system success is closely coupled to
documentation quality. Users show little interest in reading manuals or
documentation from cover to cover; their interest is focused on gathering
information or completing a task (Redish, 2012). Content is now considered
important enough that it has morphed into its own field, with books and
guides about writing good content (Redish, 2012; Handle y, 2014). Writing
content exclusively for the web has its own gu idelines (see Section 14.2.2).

488 Chapter 14 Documentation and User Support (a.k.a. Help)

Many users are technologically sophisticated. There is no appeal to reading
or brow sing through many page s of documentation; users want to get going
quickly with their technology products. Users expect an easy learning curve and
the actions to perform the tasks at har1d to be easy to find and discover. Only
with the most sophisticated computer systems is there a 11eed or desire for
extensive training and long start -up times. Users want quick -start guides
(Fig. 14.1), easy -to-navigate material, and lots of examples (Novick and Ward,
2006a); they also want content on tl1eir preferred medium including, video and
audio, topic -based information, embedded help, animations, and 3-D graphics
(Hackos, 2015) Some of these attributes are difficult to produce in general
documentation for a wide-ranging audience.

Quick Start Guide
New to OneNote 2016 or upgrading from a previous version? Use this guide to learn the basics.

Quick Acc:ess loolblf
Kiecp .._._ OOll’ffl$nd5

p,ll’fNlllll’ltlf•iobla.

llq,lett dw ribbon
Stt “”-‘ OneHolle 2016 c.i do bef dicb,g ttwe ribbo

ft

w.wo~-•…,,.._’°°”-

DIK- conlf’XIU.1,1 tonWIWKh
Sdect,ny,-tcl … ~t.abteto _,fldcl .. ~~ 1CIOIII.

I] Office

Shareyoiwworkwi th~

$,gnlft..-itl’l-,O.,

,,.., ,. , ,, ,, — ,a::-e ..
~ . .J ,_ – ·- .. _ ·- …. .,…. , ,, ..

C -· .,, ·-
“‘:”‘ .. — , , 11-

••

%•£ • • )I. ………. , – -~ —

r
,a .. – •• –· —–·-

Nowt,oolq: Li~
Od:: ltle 1’11:Jtdiot,l( Nffle lo
-11:h bet-. IICllrboolcs Of

cick the FW’ icon to biep the —
~ lt

Gd; or d,.g h gnpper 1

0

Nlcftdal’ClletoMkcto,
r,,o,,,e11t, ot r,ght-dc:kc fot _,,.,,..

~o…–
Microsoft

FIGURE 14.

1

… …..

+ f•,!(“‘lf.’O!~P!ln ·-·-·–. -~——· . __ .. _ … ..,,_

* ~ ~–J’IM ___ ………….. ·•—, ___ … w,.._
Rnbot Nocit ( OM.-inrn

.,__ ___ CRglheedgetotNloenotec

fll the pogeor~thetnbre
f,-,ne tomo¥eit~

T~ any,t,fl«e on• ~

— Oneffote\ fled:’eQff4$ wi,c:~•” …. “”‘Y’
INIOCntr~-.

,
……. ;7

‘ ‘

-~-~~ … …… —·-

Hide 1he ,llbon

Need – –

?

Oid:the-lo
tum the nbbon on …..

•lt’tantJy f’,nd ~Y’hi119
s-m the wnent page 0, ..

notttioob .t ontc.-w:I nwi·
gMethefewlb…ihet1w.

Kold)ootthgci
Od:1hesetlhltoMotch
~P’flll5″‘the
curreo-.,-..11:-bot.A~

This particu lar quick-start guide for Microsoft OneNote 2016 points out common
actions for users. Its goal is to minimize the learning curve for users upgrading
from a previous version. The quick-start guide is four pages long; this is the first
page . Explanations are provided for the tool bar and the ribbon in addit ion to other
commonly used features.

14.2 Shaping the Content of the Documentation 489

14.2.1 Organization and writing style

Designing documentation is a chaUenging endeavor. The author must be knowl­
edgeable about the technical content; sens itive to the background, reading level,
native language, and intellectual ability of the reader; and skilled in writing
lucid prose. Assuming that the author has acquired the technical content, the
primary job in creating documentation is to understand the readers and the
tasks that they must perform.

A precise statement of the instructional objectives is an invaluable guide to the
author and the reader. The sequencing of the instructional content should be
governed by the reader’s current knowledge and ultimate objectives. Precise
rules are hard to identify, but the author should attempt to present concepts in a
logical sequence in increasing order of difficulty to ensure that each concept is
explained before being used in subsequent sections, to avoid forward refer­
ences, and to construct sections that contain approximately equal amounts of
new material. In addition to these structural requirements, the documentation
should have sufficient examples and complete sample sessions. These guide­
lines are valid for manuals and other documentation that is typicall y read
sequentially. When writing for sophisticated interfaces, where approaches are
not sequential, new techniques may need to be tried. A mantra of many techni­
cal writers is “write once, publish many places,” as they are aware that the con­
tent will be published in multiple formats and in multiple places. Companies
such as Acrolinx can help the writer with this process.

Before starting to write any documentation, it’s important to thoroughly
understand who the intended users are and how and where the documentation
will be used. Frampton {2008) sugges ts numerous questions that should be con­
sidered: Who is the intended audience for the documentation? What are the
market expectations for the documentation? What amount of the budget has
been allocated for it? Are some components of the documentation considered
essential or required and others supplemental or nice to have? How will the
documentation be used? Will it be used once and thrown away or used repeat­
edly over a long period of time? What is the reading level of the potential users?
Is the documentation written in their nati ve lang uage? What is the intended
user’s level of comfor t w ith techno logy? Following the user-centered design
process is a good way for the documentation authors to communicate and dis­
cuss requirements with the users, and this approach will yield better-designed
documentation ..

Redish (2012) encourag es authors to divide documentation int o topics and
subtopics. Topics can be organized according to time or sequence, tasks, people,
type of information presented, or questions people ask. Today’s online world is
one of agile information development, and technical information development
is und ergoing many changes (Hackos, 2015). Development cycles are short, and
competition is fierce. A trend in technical communication since the late 1990s is

490 Chapter 14 Documentation and User Support (a.k.a. Help)

to write documentation based on standards, such as Darwin Information Typing
Architecture (DITA). The DITA standard (http:/ /dita.xml.org) emphasizes
developing and organizing content based on three information types: concept,
task, and reference. These information types represent “chunks” of content that
can be reused and, importantly, published across multiple platforms (Fig. 14.2).
Adhering to a standard also enables sharing of content between different groups
within and outside an organization. For example, a training group can reuse
content from tl,e techni cal writing group.

Users interact with the documentation on several different cognitive levels.
They go to the documentation to find information that is relevant to accomplish ­
ing a task. They need to understand what the documentation is explaining, and
they then need to apply that understanding to the task that caused them to
consult the documentation in the first place. In this process, there are lots of
places for misunderstandings and increased cognitive load. Additionally, users
may already be in a stressful situation and frustrated because the interface is not
letting them accomplish their task.

The choice of words and their phrasing is as important as the overall structure.
A poorly written sentence mars a well -designed manual, just as an incorrect
note mars a beautifully composed sonata. The Elen1ents of Style (Strunk et al.,
2000) and On Writing Well (Zinsser, 2006) are two class ic references. Other more

Darwin Information Typing Architecture (DITA) Mapping

Data source topics

1

2

3

4

5

IT) i————- –4 … ►

FIGURE 14.2

Compiled
help

CFM

Website
HTML

Print

PDF

The documentation content is organized as single source topics created, during
the authoring process, with the intention of reuse. The various documentation end
produc ts (compiled help, website, print) are mapped from the various combinations
of the single source topics.

14.2 Shaping the Content of the Documentation 491

recently updated books on writing include Everybody Writes (Handley, 2014),
Letting Go of the Words (Redish, 2012), Every Page Is Page One (Baker, 2013),
Designing for Digital Reading (Pearson et al., 2014), Conversation and CommunihJ
(Gentle, 2012), Designing Information (Katz, 2012), The Essentials of Technical
Communication (Tebeaux and Dragga, 2014), Solving Problems in Technical
Con1n1unication Gohnson -Eilola and Selber, 2013), and The Sense of Style (Pinker,
2014). Specialized books such as The Global English Style Guide (Kohl, 2008) can
help with writing for global audie11ces, and ReaderCentric Writing for Digital
Media (Hailey, 2015) provide s a theoretical framework to improve online media
content. Style guides for organizations represent worthy attempts at ensuring
consistency and high quality (see Chapter 1). But, of course, no set of guidelines
can tum a mediocr e writer ir1to a great writer. Writing is a highly creative act;
effective content and documentation writers are needed.

There are numerous resources available for professional communicators,
with an emphasis on technical communication. Formal courses and degree
programs exist as well as specialized institute s and workshops. Books (as
mentioned above) exist to explain technique s for writing documentation (and
specifically web content) as well as formal pedagogy. The IEEE (through its
Professional Communication Society), the ACM’s Special Interest Group on
Design of Communication (SIGDOC), and the Society for Technical Communi­
cation (STC) pro vide theore tical publi cation s as ‘”‘ell as information of a mor e
practical nature. Redish (2010) discusses the intertwining of technical communi ­
cation and usability, how both fields can and do learn from one another , and
how the combination adds strength to both. Studies are also und er way to see
how crowd sourcing can be us ed to impr ove writing (Bernstein et al., 2015).

14.2.2 Writing for the web
When writing for the web, it is important to be aware of the differences from
conventional writing and reading practice s. When the web is utilized on mobile
devices, screen sizes can be small and screen real estate is critical, so the designer
needs to be aware and make every word count. In journalism, there is the ongo ­
ing discussion about the pros and cons of the inverted pyra1nid style of writing. In
the inverted pyramid the most important information or the main point comes
first, then the remaining information follows with the least important informa ­
tion at the end (Redish, 2012). This works well for the web since it has been
shown that users typically don’t scroll.

Gaining a better und erstanding of us ers’ reading patt erns in online environ ­
ment s is important. A study using eye-tracking and displaying where and how
users viewed a webpage clearly shows users read following an F-shaped pattern
(Fig. 14.3). This indicates that users do not read online text word by word. The
beginning paragraph s should contain the most important information; after
reading them, us ers tend to contint1e to scan down the left side of the pag e, so

492 Chapter 14 Documentation and User Support (a.k.a. Help)

FIGURE 14.3
In heat maps from the eye-tracking study, red indicates the area w here the use r
looked mos t , yellow indica tes fewer views, and blue ind icates the fewest views. Gray
is used for areas t hat were not viewed. The image on t he left is f rom an article in the
“About Us” section of a corpora te website, the center image is a product page on
an e-commerce website, and the image on the right is from a search engine results
page (https://www.nngroup.com/articles/f-shaped-pattern-reading-web-content/ and
Ja kob Nielsen).

the words they see there should also carry the important information content
(Nielsen, 2006).

Another approach coined by LesLie O’Flahavan (founder of ewriteonline.
com) is “a bite, a snack, and a meal.” This describes the varying amoun t of
content presented to the reader. A bite is the headline or main point. A snack is
a bit more detail and perhaps an article summary or some cogent points. A meal
is the entire article. The way content is read is changing, and that affects how
content is written. Now readers not only view the content but also want to easily
share the content with friends, coworkers, and so on. Web content has to be
fluid across platforms.

Box 14.l offers some design maxims to follow when writing for the web. It is
important to understand the reason the user is corning to the website. Users may
be coming for content; they are not there to reflect upon the design. Users come
for information to answer a question or to help them complete a task; that
information should be easy to find and easy to understand. Think about every
time the web content is accessed as the beginning of a conversation started by
the site visitor (user). Obtain as much information as you can about your users.

14.3 Accessing the Documentation 493

BOX 14.1
Helpfu l advice when writing for the web (Redish, 2012).

• Break your text into short sections with clear headings.

• Start with your key message.

• Write short sentences and short paragraphs.

• Use lists and tables-tables may not work well for mobile devices.

• Write meaningful links (do not use “click here”) .

• Illustrate your content.

14.3 Accessing the Documentation

Studi es in the past have confirm ed that well-des igned documentation can be
very effective. In spi te of improvements, however, most users avoid user
manuals if they can. If users read the documentation, they “sa tisfice,” skip, scan,
and skim (Mehlenbacher, 2009). Users typically do not want to sift throu gh
volumin ous user manuals that can be difficult to navigate. Instead, they want
quick and easy access to instruc tions for the speci fic tasks they are trying to
carry out (Redish, 2012). Even when problems arise, man y users are reluctant to
consult written documentation and may do so only as a last resort. Although
guidelines have been applied to improve the design of online components to
take advantage of their unjque media, studi es still show low use of docwnenta­
tion (Novick et al., 2007).

Standard formats such as HTML Help, Java Help, and DITA have stimulated
development of a growing number of software tools to generate help files, such
as Adobe RoboHelpTM (http://adobe.com/products/robohelp.html) and
oXygen WebHelp (https:/ /www.oxygenxm l.com /xm l_author/webhelp.html).
These tools facilitate coordination among teams of authors in creating interactive
online help in multiple formats for multiple platforms.

Documentation is often placed online for good reasons. The issue becomes
one of making the best use of the online environment when accessing the
material. Multiple ways are available to search and traverse the online
information differently from paper documentation. One feature of online
documentation is the ability to offer pop-up help as well as customized help for
var ious user population s, such as users with dis abilities, international users,
and user s in varying age ranges.

494 Chapter 14 Documentation and User Support (a.k.a. Help)

14.3. 1 Online documentation

The low production and shipping costs of CD-ROMs initially encouraged
hardware suppl iers to produce online documentation that was an exac t
duplicate of the paper documentation and/ or manuals. Most manufacturers
toda y put their user documentation directly online and no longer create sup ­
porting paper-based documentation. Modern designs assume that online
documentation or web-based documentation will be available, usually with
standard browsing interfaces to reduce learning effort. For mobile devices,
small displays limit the possibilities, but providing helpful instructions on
the device to complement printed user documentation should still be a prior­
ity. To keep this information up to date, users are often referred to the manu­
facturer’s website, where downloadable manuals and other forms of
documentation, such as frequently asked questions (FAQs), are often readily
available.

A vital feature for online documentation-especially manuals-is a properly
desigr1ed table of contents that can remain visible to the side of the displayed
page of text. Selection of a chapter or other entry in the table of contents
(Fig. 12.5) should immediately result in the appropriate page being displayed.
Use of expanding or contracting tables of contents (the common use of the plus
and minus signs) or multiple panes to show several levels at once car1 be benefi­
cial. Being able to conveniently and easily navigate using hyperlinks through
large volumes of online documentation is vital for the user.

Although they are frequently generated from the same source document
(usually an XML or XHTML document), online documentation now tends to
differ from paper documentation, benefitting from all the physical
advantages, navigation features, and interactive services available online
(see Box 14.2). On the other hand, paper documents have traditionally housed
supplemen tar y local information that is of ten written in margins or included
on slips of paper stuck in at the appropriate pages. Some printed documenta­
tion remains pristine, neatly encased in its original shrink -wrapped packag ­
ing, but dog-eared, stained, and worn pages are often seen on well-used
documentation. Online documentation that allows for local annotations, syn­
onyms, alterna te phrasing, or tran slations has enhanced value . Addi tional
desirable services include support for bookmarking and automatic history
keeping, which allows backtracking. Designers will be most effective when
they design online documentation to fit the electronic medium and take
advantage of text highlighting, color, sound, animation, and string search
with relevant feedback. In some cases, provisions may be needed to provide
the documentation offline, if access is not allowed outside of internal net ­
works. Researchers caution that words should be treated as valuable com­
modities and used spar ingly in documentation, whether in printed form or
online (Redish, 2012).

14.3 Accessing the Documentation 495

BOX 14.2
Advantages of online documentation.

• Physical advantages: Available whenever on web-connected electronic
device, can ‘t get lost or misplaced, physica l workspace not needed, can
be updated rapid ly.

• Navigation features: Can provide inde x and other search facilities, can
link to other external materia ls and sources.

• Interactive services: Can bookmark, annotate, and tag; can inc lude
graphics, sound, color, and animation; screen read ers or other tools can
be provided for users with disabilities.

14.3.2 Online help
Users typ ically seek out help in solving spec ific probl ems and will wa nt to go
directly to the informa tion that is needed. The traditional approach with online
documentation is to have users type in or select a help-menu item; the system
then disp lays a list of alphabetically arranged topics that can be clicked on to
read a paragr aph or more of helpful information. This method . can work, but it
is often frustrating for those users who are not sur e of the correct terms to use to
search for information on the tasks they wish to accomplish . They may see
several familiar terms (search, query, select, browse, find, reveal, display, info ,
or view) but not know which one to select. Worse still, there may not be a single
command that accomplishes the task, and there is usually litt le information
about how to assemble actions to perform tasks (such as converting graphics
into a different format). Online help that offers concise descriptions of the inter­
face icons and actions (Fig. 14.4) is probably most effective for intermittent
knowledgeable users; it is likely to be less useful for novices, ,-vho have more
need for tutorial training.

Sometimes simp le lists-for example, of keyboard shortcuts, menu iten1s, icon
glossaries, or n1ouse shortcuts–can provide the necessary information. Each item
in the list might have an accompanying feature description. However, many
designer s recognize that such lists can be overwhe lmir1g and that users usually
want guidance for accomp lishing their specific intended tasks (for example,
printing on envelopes).

The online help and suppor t center found with most Microsoft products
offers man y ways of finding relevant articles, called topics. Users can browse
an organized table of contents that lists the topics hierarchical ly or search
the text of the articles. Finally, Microsoft’s current approach allows users to

496 Chapter 14 Documentation and User Support (a.k.a. Help)

• • Migrat ion Process

Tables Relationshi s TOs Value lists CFs

Toblos. 5 .

/

New Table Source Original Table ~,_,,. ____ ns_, ___ __,~ Contac~

FileMaker 14

Step 2:
Import Layouts from
FlleMaker Advanced
database(s).

Click lhe Layouts tab,
click the Layout Batch
Import bullon .

Step 3:
Click lhe Scripts tab.

Select all of lhe
SctiplMaker Scripts
within FileMaker
Advaneed, oopyto
clipboard.

Click lhe Add Scripts
button to copy the
scrip~ from the

Stop 5

Remap Objects

ln Prog,oss

Inc.

FIGURE 14.4

Members FileMaker 14
lbl Assets lbl assets FileMaker 14
tbl_Maintenance_Reoor

FIOlds. 70

ID
1
2
3
5
7

Original Fieldname
10
Model
Item
Category
f”.nc::

I

Step 2

~-·· –_,
-· -·

Got Layouts

Co~ted

Stop 6

+
j

Create CFs, Tabios ,
Table Occ urrences

Not Started __ ….,

FileMaker 14

)
PK Auto-Increment
1 1

0 0
0 0
0 0
n n

Step 3

+
Got Scripts

Co~e ted

Stop 7

XML

Create Value Lis ts ,
Scripts and Layouts

Not Started

Seri ts ~u
. Xl,S

Destination
FileMaker 14
FileMaker 14
fileMaker 1
FileMaker 14

GUI

Fields Recor ,
56
17
0

?

Next SN New Fieldnamo
5 id

model
item
category ..,…,

Step4

Got
Relationships

Co~eled

Stop 8

Create Data
Import Scripts

Not Started

This f igure provides a step -by -step approach to the activ ities in a database migration
applicat ion. The details invo lved in each step can be found on the left -hand side.
The actual action and activi ty are in the middle . On the bottom is a road map for the
steps involved; the steps are marked as to t heir status (comp leted, in progress, no t
started). For novices, it would be easy to insert links to more deta iled tutoria l-type
information (http://www.fmpromigrator.com).

type requests int o a search box using natural languag e statement s; the pro ­
gram then selects the relevant keywords and offers a list of topic s org aniz ed
into cate gories.

Interactive help The simplest way to take context into account is to moni ­
tor the cursor location and provide helpful information about the icon under

14.3 Accessing the Documentation 497

the cursor. This form of user-controlled interactive help is readily understand­
able to users and even fun to use. This information should have three quali­
ties: It should be available without interfering, succinct yet descriptive, and
nonintrusive (Sherwin, 2015). Users position the cursor on a widget (or other
visible interface icon) and then press a help key or hover the mouse over the
icon for a couple of seconds to produce information about the icon on which
the cursor is resting. In a common version of this technique, users simply move
the cursor to the desired location and hover over the icon, causing a small pop­
up box (often called a tool tip, Screen Tip, or balloon help) to appear witl, an
explanation of that icon (Fig. 14.5). Alternatively, all of the balloons may be
displayed at once so that users can see all of the explanations simultaneously
(Fig. 14.6). Anotl,er approach is to dedicate a portion of the display to help,
which is updated automatically as users hover over or selec t interface wid­
gets or icons. Yet another approach is to use two monitors, where one monitor
displays the user’s activity and the other monitor displays the help informa­
tion with videos and help pages (Matejka et al., 2011). User-controlled help can
also be used for more complex items such as con trol panels or forms. These
features provide a narrower window into the extensive volume of help that is
available to the user. By having the users control the action, it is less disruptive
to them. Designers can create a small overlay help window that allows users
to adjust the size as well as allowing them to minimize, close, or move the
window (Sherwin, 2015).

System-initiated help Another approach is to provide system-initiated
help, often called “inte lligent help,” that tries to make use of the interaction
history, a model of the user popu lation, and a representation of its tasks to
make assumptions about what users need. By keeping track of user actions,
some researchers believe that they can provide effective system guidance, such
as suggesting that users redefine their margins since they are indenting every
line. Research in computer-based intelligent user interfaces still continues to
have mixed results.

8 (‘) 6 – dtu l 6 chap 14 ,ev 10- BS K •’
~ -:aia, 13~ rs I), “1:i:· • . • . t: ~ frnil . u (._O.· l. Se.arch In Document

TEXT IN”l)(Hf • -.ITimu Ntw,.. I• .I 12 ~• I ~~~ ~!, ntng from one kxa11on and llJ>PIY — •= e = O · oal • A

• –

” Homo Uyo,n DocurMnt Ek’rn<'nl,; T.tblt-S Chuu •=- Rtvicw, ~ (I ... ,. . ..,., S1y1t, ln1t tt ""~·

ln– …… 1·1112 I·’. A• A· .~ – j ~ == • •= • ‘•= · J .:;; .ra Tl · r ~- ~-~-::,,. A.a “I:., • ,·- ,, i- ,, •- , , – ‘=-‘ · © AlflloCc[)dS AIHbCd)46 ® ….
00 1 \l •:’dliF. N A, .a. . ,a( • Ii},. – – -i ~ :~ • + • f; , TOO’ IIIOlNT WITT :0 Tell.1 lox “”‘ ·- ,,,._
-r,-: C….: , 8 … • II . • •JI . • . • .. • .. 4,, • • •• $1 …. .. .. A •

• •

I • • llol … _,’.,, .. ffllillarThe sim’nlc:sl “”~\’ to take conkxt into ‘acOOUJ1l ill to ~onitor ihc cunlOr localion and •

FIGURE 14.5
A small too l tip or fly-over help on Microsoft Word. Moving the mouse (cursor) over
a particular icon causes an explanation of the icon to appea r on the screen.

498 Chapter 14 Documentation and User Support (a.k.a. Help)

1’3,) 9 -=:-r-~ ,—,– 11 41 ‘,1-f 11 _ _,,, .’I – • –
eo… CanllOllbal Adderley ‘4!i…;”w””•


Import • “”9 r,,,.,, ;CJoud, (:),q,bo< \ \ c, rAtr Biti.i!'ooth or lt.Lan .,-/

Frank Rosollno


r-oft’l.ini * —— …. J

Sa-gi,ooi

n,;, ;, “” “‘”f _.liJI ,…d, with
“‘

0

Fai:nbool: ,..de,
l,rpo-l all !JW ~kroook pdfa •r,J io, ) \j
l’1’8 ,eadj b’ tha noo .. am teS.illOI\ I –

0oot,,,,.,,. mot/a

Rm,,.,,,, nni ~,l,10 i~, :m \/\/
dwrlood p:I~ and Mfo from ll •
rifemlli

FIGURE 14.6

Om.,, tinrcto’icffl ~~ J

~k.8 too arrotat(1\!l ~~ I

~ –
ii lb, Fl rrT7 I ‘..

&,r,g ;nR, ….in E,r,1 O”‘!J nHrii…,.

and oq>C

&,,j oo, moda Setup ops ,1,..,.
li:r !”‘li:r”””‘”

~ ~’ooze :” 0Jd”4o I 1~

Jeannine I Pia\! roo0Jd 0000 \ \

J ~n~ ~”J ~,xle , J

Reccrd d,nntj ~”!I imde ”

Mclr””‘”” .nd o11 … ,eff>”!J’ fir “” I , ,
,,. avg ;/

1 r-:– ~::—:.==–:—~• -=—-

Since this is a complicated music interface, interactive help is provided for many
of the fields. This figure shows all the help balloons open. Severa l standard icons
are used such as new, import, record, and play, but other specific icons explain
the various modes that are availab le while using the too l (http://www.saxopedia .
com/2013/12/07 /ca lypso-score-the-pe rfect-ipad-m usi c-score-reader/) .

14.4 Reading from Displays versus Reading
from Paper

The technology of printing text on paper has been evolving for more than
500 years. Tl-te paper su rface and color, the typeface, character width, letter
sharpness, text contrast with the paper, width of the text column, size of margins,
spacing between lines, and even room lighting all have been experimented with
in efforts to produce the most appealing and readable format.

Visua l fatigue and stress from reading computer displays are common problems,
but these conditions respon d well to rest, frequent breaks, and task diversity.
Even if users are not aware of visual fatigue or stress, their capacity to work with

14.4 Reading from Disp lays versus Reading from Paper 499

F GURE 14.7
Users read their favorite news sites on a
variety of media. Sometimes it may be
their phone, other times their computer.
The content needs to adjust (dynamic
resizing) to the chosen medium. Not e how
the display on the laptop is in a landscape
(horizontal) format, but the display on the
table is portrait (vertical). On the phone,
some of the materia l is not availab le on
the first screen but may be available as
users scroll or select a Ii nk to display
further information. It is important that
the format of the content is tested and
reviewed on the different-sized devices.

displays may be below their capacity to
work with paper documentation.

The interest in reading from
displays I-1as increased as mobile
devices, tablets, specialized electronic
book platforms, and web-based
librarie s have become more common.
Being able to download a customized
morning newspaper onto a pocket­
sized electronic device to read while
standing in a crowded subway or to
carry a full city guide on such a device
while touring are strong attractio11s
and often expected by today’s users
(Fig. 14.7). If users are to read large
amounts of material onl ine, high­
resolution and larger displays are rec­
ommended. Other studies recommend
that quick response times, fast display
rates, black text on a wnite back­
ground, and page-sized displays are
important considerations if displayed
text is meant to replace paper docu ­
ments. Some applications provide a
dedicated reading layout view that
limits the number of controls and

increases the space available for viewing the text. Dynamic resizing can take
into account the display size to facilitate paging through the document instead
of scro lling. Sans serif fonts (Fig. 14.8) should be used online, as they are crisper
and produce a less cluttered appearance. Sans serif fonts are also preferred for
low -vision readers. The term sans serif comes from the French and means
“without serifs,” where serifs are the ornamentations on each letter.

0 e oogle
FIGURE 14.8
In September 2015, Google changed its standard logo. Note the new logo (on the
right) uses a sans serif font. One reason for this change is to make the logo easier to
scale on small screens like mobi le devices. Sans serif fonts appear cleaner and less
cluttered and are eas ier to read on screens.

500 Chapter 14 Documentation and User Support (a.k.a. Help)

Large online libraries of books-whether they are made avai lable for free, as
by the Gutenber g archive s or the Librar y of Congress; by library subscription;
or for pay, a servi ce offered by num erou s publishers-promote efforts to
improve the reading experience. Publishers of newspapers and scientific jour­
nals are evolving to satisfy the intense demand for online access to articles while
struggling to ensure a way to recover their costs. Plasticity of documents
(responsive web design) is becoming a requirement. The ability to automatically
sense the correct orientation for reading a document has become a standard
feature. Con tent designers have to structure their ma terials so that they can be
read on small, medium, and large displays and at different font sizes to accom­
modate vision -impaired users. Designers need to be aware of their target
audi ence, including accommodation s for age and educational levels (Fig. 14.9).

Skip Navigation Help Search lffll ~——~
Home ‘ Health Topics A-Z J Videos A-Z I About Us I Contact Us

NIHSeniorHealth
Built wi th You in Mind

Resi1e TH t: A A A Change ContrastO Prilll ~ Sign Up~ Sllare a

Health and wellness infonnation for older adults from the Nationa l Institutes of Health.

Hea lth Top ics by First l etter

A B C O E F G H IJ KLM N OPQRST UVWXYZ ——–

Catego ries

• Bones and Joints

• Cancer
• Diseases and Cond itions

• Healthy Aging
• Heart and Lungs

Exerc ise Sto ries

People of al ages and physical conditions benefit
lrom exercise and physical acti’lily. These exercise

stor ies fea1ure older aduls and lhe dverse

actNities they enjoy.

FIGURE 1 .9

Fe.itured Topic

• Memory and Menta l Health

• Treatments and Therap ies
• Vis ion and Hearing

Psoriasis
• All Topics A-Z

Hea lth Videos

Many of our healh topics featu’e short videos ,
that complement lhe information in the topic.

The health videos olfer up.to-date medical
irf0

conditions or aging.

The Nationa l Institutes of Health’s site for seniors (http://www .nihseniorhealth.
gov) has controls to adjust the text size and adjust the contrast. The font used is
a sans serif font, and the font size is larger than the typ ical size used on the web
(being aware that the site is targeted at olde r adu lts). Several ways are provided to
navigate th rough the information (alphabetica l, grouping by category, etc.). There
are controls provided to resize the tex t as well as change the con t rast.

14.4 Reading from Disp lays versus Reading from Paper 501

14.4. 1 Extended reading on displays
Althou gh ear lier displays did not always provide the best resolution for
reading, the digital ink technology used on many of today’s reading devices
has changed that. Yet researchers are still trying to understand and see what
issues, if any, exist from extended reading on displays.

The Pew Research Center (2014) states that 50% of Americans have a
dedicated hand held device used for reading (Fig. 14.10). Another finding from
its research is that those who read ebooks read more books in all formats. People
like ebooks for their speedy access and portability. Print books are still preferred
in certain situations, such as reading to children. Often people prefer print to

FIGURE 14.10
A lthough conventiona l books
are relative ly easy to transport
and read in different places,
reading an ebook has become
very convenient as we ll (inc luding
reading in bed). The ebook devices
are lightweight, the size of the
characters can be easily adjusted,
and today’s technology has even
removed much of the glare, so the
ebook can even be read at a sunny
locat ion like the beach.

digital for sustained periods of reading
(Stone and Baker-Eveleth, 2013).

Johnson (2014) states that reading is an
artificial skill that must be learned by
instruction and practice. Designers should
be aware how poor information design can
make reading more difficult and disrupt the
process. Some guidelines for supporting
reading are:

• Don’t use uncommon or unfamiliar
vocabulary.

• Avoid difficult-to -read typefaces (ALL
CAPS is harder to read) .

• A void text on busy backgrounds.

• A void information buried in repetition.

• Don’t use centered text.

In a study on reading comprehension
performed with 10th graders in Norway
comparing reading texts in print versus
PDF files on a computer, it was found that
reading from the text s in print produced
better scores on reading comprehension.
Reading in print gives users a better merltal
representation of where they are on th e
page, providing fixed spatial cues (Mangen
et al., 2013). There are some concerns that
the way users read online is different from
how they read in print (see Section 14.2.2).
Users often skim information on the web.
In Western cultures, users read from left to

502 Chapter 14 Documentation and User Support (a.k.a. Help)

right, and the scanning across each line gets shorter as users move dov.rn the
page (Fig. 14.2).

No one yet has answers about what will happen to people as they continue
to receive such a large volume of information from so many different sources.
When Maryanne Wolf initially pub lished her work in 2007, the electronic
tsunami had just begun. It seems that as more reading content moves online,
studies seem to show a decrease in reading comprehension. Just like we listen
to “sound bytes,” we now read with ”eye bytes.” Studies seem to show read­
ers are losing some of their deep reading skills (Wolf as cited in Konnikova,
2014). Other studies are under way such as Chen and Chen (2014) that show
how a collaborative reading annotation system can enhance digital reading.
Baron (2015) adds that techt1ology is reshaph1g what it means to read, and
with that we need new innovative devices that support reading in the online
environment (Fig. 14.11).

FIGURE 14. 11
New ebook devices will be created that support the metapho r from physical books,
allowing the user to flip the pages to scan the contents of the book. In th is photo
t he user reveals a page tha t has previously been “bookmarked” by the simp le
act of resting his thumb on the margin of the page. But the tipping alone does not
actually nav igate back to the prev ious page in and of itself. The user has to confirms
the act ion by tapping t he thumb against the edge of the device, so as to actua lly
navigate back to the location of the bookmark. This serves to reveal the interaction,
while also ensuring that it is an overt act that is not like ly to be triggered by accident
when one simply rests their hands on the device (Yoon, et al. 2015).

14.5 Online Tutoria ls and Animated Demonstrations 503

14.5 Online Tutorials and Animated Demonstrations

An online tutorial is an interactive training environment in which users can
view explanatory descriptions of user-interface icons and actions, often tied to
realistic task scenarios. There are many approaches to the use of electronic
media to teach users how to master an interface. Depending on the complexity
of the interface and the amount of time users are ready to spend absorbing the
tutorial materials, they might be served well by an extensive computer-based
training module, an animated demonstration of features, or a recorded
welcome message by a familiar person. The challenge often is to prepare
materials that will satisfy users who want a three-minute introduction as well
as those who need a one-hour in-depth treatment. This section reviews a range
of online possibilities, from textual and grapl1ical tutorials to fully animated
demonstrations.

A more ambitious approach to training is based on a complex model of learn­
ing patterns tied to carefully designed educational tutorials that guide users and
correct their mistakes. These have demonstrated impressive outcomes, but the
success stories are based on years of development, testing, and refinement. The
successful designs provide clear challenges, helpful tools, and excellent feed­
back. This can be contrasted with the many YouTube tutorials available, many
by do-it-yourselfers with minimum expos ur e to any learning principles and
pedagogy.

14.5. 1 Online tutorials
One introduc tory tutorial for the Adobe Photoshop ® package displays the exact
steps users must make and then shows the actions being carried out using a
recorded demonstration. Users just keep pressing the space bar key to speed
through the demonstration. Some users find this guided approach attractive;
others are put off by the restrictive sequencing that prevents errors and
exploration. Automated tutorials can be created using Autodemo ® and Show
Me How Videos TM_ Autodemo has arrangements with several wor ldwide
companies and provides specific instructions on navigating the various
websites. Additional research into online tutorials is ongoing, with excellent
contributions from Autodesk TM_ Four emerging directions are gamification,
community input into tutorials, leveraging transfer effects, and distribution of
practice (Cockburn et al., 2014).

The opportuni ty for carrying out practice tasks during online tutorials is one of
their greatest strengths. Getting users to be active is one of the key tenets of the
minimal -manual approach, and it applies especially well to online tutorials.

504 Chapter 14 Documentation and User Support (a.k.a. Help)

Another attractive variant is the start -up tip: Each time users start the
interface, they get a pop-up box displaying a brief explanation of a feature. Some
systems monitor user behavior and show start-up tips only for features that
are not used by this particular user. Of course, the user sl1ould always be give11
the option to tum off these tips at any time.

Creators of interactive tutorials must address the usual questions of
instructional design and also the novelty of the computer environment. A
library of common tasks for users to practice is a great help. Sample documents
for word processors, slides for presentation software, and maps for geographic­
information systems help users to experience the applications. Repeated testing
and refinement are highly recommended for tutorials.

With the variety of tutorials available (from both high-quality companies and
do-it-yourself selections), it is l1ard for the general user to assess the quality of
the tutorial. There is limited information in the research literature about a set of
quality metrics. A single rating does not seem to distinguish quality, so multiple
criteria (learning, coolness, ease of following, enjoyment, writing style, error
prediction, image helpfulness) need to be used (Lount and Bunt, 2014). Today
tutoria ls are not just a sing le point of entry into learning the software. Most
users will also go to the web and see what others say about the tutorial. Having
a set of Tagged Comments (Fig. 14.12), where users can tag their comments and
pin them to particular locations in the tutorial, is helpful (Bunt et al., 2014).
Another approach is to provide for community-enhanced tutor ials, which
continuously improve and evolve as more users work with them (Lafreniere et
aJ., 2013).

Using some automated tools can improve the quality of tutorials. MixT auto­
matica lly generates step -by-step tutorials from user demonstrations and actions
(Chi et al., 2012). Another product is DemoCut, a video -editing system that
improves the quality of amateur instructional videos for physical tasks (Chi et
al., 2013). EverTutor provides simplified tutorial creation by generating interac­
tive tutoria ls on smartphones based on user demonstrations (Wang et al., 2014).
Designers need work to continue to improve the quality of tutorials (Wakkary et
al., 2015).

14.5.2 Animated demonstrations and multimedia

Animat ed denionstrations have become a modern high-tech art form (see
Section 12.4). Manufacturers originally designed them mostly to attract
potential users of software or hardware by showing off system features using
the best animations, color graphics, sound, and information presentation that
advertising agencies cou ld produce. Those demonstrations focused on
building a positive product image. Demonstrations and videos have become a
standard technique to train users as they work. These have been enhanced by
the use of augmented reality (Mohr et al., 2015 and Section 7.6). The focus is

14.5 Online Tutoria ls and Animated Demonstrations 505

Qui ckly Build an Abstract
Bac kground of Color ed Bars
Oftenbmff -··· a nffd , … q..ocl< but - .S.sq,. In - lu!ONI, l'I M -..ng you a short but attrKIIVt way to «•ate a g,Ol)lloc fo, l>Kl

Step 1

Fwst off, r.t’s sta ~ > Oouds),

FIGURE 14.12

F~tor 131 commenis by eatego,y

filte, comments
,.., ClOfflfflfflCS ..,

Pinned con1111e.nts Add~ COffW’l’tilrjt) OoH

°”‘ 18/10.’1010 ) l lt 4J PM. 9J •M l • ~
(214) I Md us,r,g 1M r-. fill., wO

• • ,,,,,. 0. , • …,.,

°”‘Ol/07 / 2010 U•ZI • > …… M n od
(210) u how 10 JSI try ,O.,S and 11 w­
w,U, Vfl’f good : ) good ~ –

• • ..,) ,o. ….. J

ubmil a comment anonymously

… ..i.ct. — !hot-· – yov,
OfT’lffllf’t.

This is a specialized interface cal led TaggedComments. The comments section can
be seen to the rig ht of the images. In t he upper right corner, the user can filter the
comments by category. Co lor coding is used to distinguish between the categories.
There is a place to pin a comment and a comment can be posted anonymous ly.
(Bunt et al., 2014).

on demonstrating step-by-step procedures and explaining the results of the
actions (Woolf, 2008). Automatic pacing and manual control satisfy hands-off
and hands-on users, respectively. Use of standard playback controls allows
users to stop, replay, or skip parts and add to their acceptability.

An animated demonstration can be prepared as a slide show, a screen-capture
animation, or a video recording of a person using the device. A slide show might
be appropriate for form fill-iI1 or menu-based interfaces, but animation is prefer­
able to demonstrate direct-manipulation interactions such as drag-and -drop
operations, zoom boxes, or gestures. A screen-capture animation is easy to
produce with standard tools such as Camtasia Studio ® and Flash . These
recordings can then be saved, possibly annotated or narrated, and replayed
automatically by users. In our own explorations, we found that users appreci­
ated recorded voice explanations, which make the demonstrations livelier and

506 Chapter 14 Documentation and User Support (a.k.a. Help)

lead to more compact demonstrations; however, providing scripts and subt itles
is necessary to address the needs of users with disabilities. Also, a video of a
person using the interface can help clarify how special hard ware is to be used­
for example, to demonstrate the two-handed operation of a drawing system or
the unfolding of a telephone keyboard accessory.

No vice users are sometimes overwhelmed by tl1e complexity of today’s
interfaces. Supplying excellent documentation can aid their understanding, but
some times the interfa ce needs to be simplified . Usil1g a multi-layered interface
design that can unfold and further challenge and encourage users as they become
more accomplished is a good approac h . Experts can jump to the deeper layers
quickly, but novices initially encounter a simplified interface (Hwang and Yu,
2011). Computer-game designers deserve credi t for advancing the art of the
animated demonstration, with lively introductions and introductory trailers
that show samples of how the games are played. Demonstrations and previews
have to explain the game ai1d make it seem appealing and challenging, all within
30 seconds . Gamifica tion including compe tition can be used in designing
tutorial s (Li et al., 2014).

Autodesk ai1d Adobe both produce products that cai1 create animations.
There are other products as well including Cinema 4D, Toon Boom Studio, and
Blender. The market is often changing with some products disappearing and
new ones appearing. Specific uses often con trol the best choice.

14.6 Online Communities and Other Avenues for
User Support

Instead of natural language conversations with compu ters to get help, interac­
tion with other people online is pro, ring to be effective and popular. This com­
munal approach may employ e-mail, chat, or instant messaging for question
asking and responses (Novick et al., 2007). Questions can be sent to a designated
help desk or staff member or posted on a discussion board. Responses can be
received in seconds or, more typically, minutes or hours. Appl ications (e.g.,
Oracle’s Right Now) exist that can aggregate information from e-mai ls, chat,
texts, user forums, and so on, into information that can be shared with custom ­
ers to resolve product issues and answer product questions, letting users solve
their problems with the he lp of other users.

The existence and proliferation of online communities (see Chapter 11) are
increasingly appea ling because of the minimized cost to software -maintenance
organi zations and help-desk staff. Many respondents get a sense of satisfaction
from being able to help others and demonstrate their abi liti.es. Some are
motivated to achieve prominence within a community in the hope of gaining

14.6 Online Communities and Other Avenues fo r User Support 507

consu lting contracts. Of course, the downside of broadcasting appeals for
assistance is that users must publicly expose their lack of knowledge and risk
getting incorrect advice. The upside is that a specific problematic case may be
solved by an expert user.

Many websites now provide e-mail contact information or chat facilities (see
Chapter 5) as opposed to written addresses or phone numbers. Also, to prevent
basic questions from tying up staff resources, managers of help desks often record
common questions and answers into files of FAQs. This enables newcomers to
browse typical problems that have been discussed in the past. These files are often
searchable and organized by type of issue or some other hierarchical scheme.
Other alternatives for advice can be personal sharing activities. (Fig. 14.13).

Today, going to the Web for informa tion of all types is considered standard
practice. The simple interface of the ubiquitous Google search box is commonly
used. The online choices out there are many and varied, but users need to be
aware that not all information is correct and valid (see Chapter 15).

Although companies may provide a host of online facilities for information and
advice, people are still very comfortable with asking the office “guru” for help
(Novick and Ward, 2006b). Human -to-human communication removes some of
the barriers found with more traditional documentation: Lack of understanding

-‘

J.

FIGURE 14.13
Peop le like face -to-face help as ev idenced by frequent occasions when you find
severa l users working toget her on a shared device.

508 Chapter 14 Documentation and User Support (a.k.a. Help)

can be handled quickly and v.rithout consequences of errors, and the human-to­
human interface offers interactivity and other cues that enhance understanding.

14.7 The Development Process

Recognizing the difference between good and bad user documentation, regard ­
less of format, is necessary for producing successful documentation on time and
within a reasonable budget. Production of any documentation, like any project,
must be managed properly, handled by suitable personnel, and monitored with
appropriate milestones (see Box 14.3).

Getting started early is invaluable. Including the technical writers in creating
the product functional specifications provides several benefits including the fol­
lowing: early involvement of people who represent the user, better written func­
tional specification, and reuse of the functional specification content into product
documentation. If the documentation-writing process begins before the inter­
face is built, there also will be adequate time for review, testing, and refinement.
Furthermore, the user documentation can act as a more complete and compre­
hensible alternative to the formal specification for the software. Implementers
may miss or misunderstand some of the design requirements when reading a
formal specification; well-written user documentation and other supporting
material may clarify the desigi1. The co11tent writer becomes an effective critic,
reviewer, ai1d question asker who can stimulate the implementation team. It1
the time before the software is completed, the documentation may be the best
way to convey the designers’ intentions to potential customers and users as well
as to implementers and project managers. Over the course of development, a

.sOX 14.3
Development process guidelines.

• Seek professional content writers and copywriters.

• Prepare user documentation early (before implementation).

• Set up guidelines documents and coordina te and integrate across all
involved departments.

• Review drafts thoroughly.

• Field-test early editions .

• Provide feedbac k mechanisms for readers .

• Revise to reflect changes regularly.

Practitioner’s Summary 509

large number of documents may be generated. It may be worthwh ile to assign a
document specialist to a large project (van Loggem, 2013).

Informal walkthroughs with users are usually an enlightening experience for
software designers and writers. Potential users are asked to read tlrrough the
documentation and to describe aloud what they are seeing and learning as well
as what they think migllt be missing. Field trials with moderate numbers of
users constitute a further process for identifying problems with the user docu­
mentation and the software. Field trials can range in scope from half an hour
with a half-dozen people to several months ,-vith thou sands of users. One effec­
tive and simple strategy is for field-trial users to mark up the documentation
while they are using it, allowing them to rapidly indicate typos, misleading
information, and confusing sections . The use of collabora tive review tools also
provides a history trail and encourages all stakeho lder s to participate in the

.
review process.

Software and its accompanying documentation are rarely truly completed.
Rather, they undergo a continuo us process of evol utionar y refinement. Each
version eliminates known errors, adds refirlements, and extends the function­
ality. If the users can communicate with the writers, there is a greater chance
of rapid improvement. When possible, keeping logs of the use of help
materials and help-desk calls will determine which parts of th e system need
modification.

Quite often, this development work is performed by different groups of peo­
ple who may be geograp hicall y dispersed in the same company, or some of the
work may be “contracted out.” However, it is important that the user sees a
smooth and integrated view . This means attention to common colors, logos, ter­
mino logy, and style. Standardized guidelines must be estab lished and adhered
to (Section 3.2), and the documentation, the associated software, and all the
packaging must represent an integrated system. It is important to know the tar­
get audience for the documentation. For examp le, conte nt writers would want
to know if they were writing documentation for developers or for the users.
Knowing if the user has any special needs that should be considered is impor­
tant. This can include age, reading level, language proficiency, cross-cultural
issues, or any other special charac teristics. It is important while developing any
documentation to be aware of lessons that can be learned from usability testing
(Redish, 2010 and Chapter 5).

Practitioner’s Summary

Sufficient personnel, money, and time should be allocated to producing sup ­
port materials, whether they be online or on paper. Excellent content is crucial
to cu.sterner success, and organi zations should cons ider moving this effort from

510 Chapter 14 Documentation and User Support (a.k.a. Help)

a cost-center allocation to part of the development activities. Documentation
and online help should be developed before the implementation to help the
development team define the interface and allow adequate time for testing. All
documentation and online help should be tailored to the specific user communi­
ties where possible. Instructional examples should be realistic, encourage active
exploration with exercises, use consistent terminology, and support error recog­
nition and recovery. Animated demonstrations should be used \,vhen possible.
Online guidance can lend a ht1man touch if it con tain s supporting information by
real users. Social media activities (including crowd-sourcing activities) through.
newsgroups, listserv s, online communities, e-mail, chat, blogs, and instant
messaging provide powerful low-cost support and easy access mechanisms and
are widely accepted. Where possible, find a content or documentation specialis t
to create the documentation.

Researcher’s Agenda

The main advantage of online materia ls is the potential for rapid retrieval and
traversal, but little is known about how to offer this advantage conveniently
without overwhelming novice users. Cognitive mod .els of how animated,
int egra ted demos facilitate learning require better und ers tanding to guide
designe rs. Users’ navigation of online help systems should be recorded and
studied to gain a better understanding of what characterizes effective help
strategies. Better strategies for integrating he lp directly in the user interface are
needed. Mu lti-layered designs in which users can selec t their level of exper­
tise seem helpful , but further testing and refinement are necessary. Better
understanding of reading patterns using electronic documents is also needed,
as are further research and understanding regarding special populations and
the ir specific design requirements.

Most of today ‘s documentation advice is based on research from early inter ­
actions with computers. Computers did not have sophisticated interfaces, and
today they are used by a variety of users with differing backgrounds (includ­
ing some non-technical). Developing modem documentation guidelines for
toda y’s software is needed. Software today is not always u sed iI, a prescrip­
tive sequential method, creating the need for innovative approaches to software
documentation development. Metrics to rate the quality of the documentation
are needed. This is a qualitative change moving from micro-HCI to macro-HCI
desigt,s. More work is needed to design new reading devices that work witl,
and assist our cognitive re sources so that reading does not become a shallow
activity. Attention also needs to be given to low-liter acy users.

Discussion Questions 511

WORLD WIDE WEB RESOURCES

www. pearsonglobaleditions .com/shneiderman

• ACM SIGDOC: http://sigdoc.acm.org/
• Communication Design Quarterly: http://writethedocs .org

http://www. ag i lemodel in g .com/essays/a g i leDocu me ntati on BestPractices. htm
http://www.smash i ngmagazi ne.com/2012 /07 /
writing-effective-word press -documentation/

• IEEE Professional Communication Society: http: //pcs.ieee.org/

• Society for Technical Communication: http://www.stc.org/
• Usa bi I ity .gov: http://www. usa bi I ity .gov/how-to-and-tools/methods/writing

-for-the-web.html

• Shirley Kaiser: http://websitetips.com/webcontent/
• Leslie O’Flahavan: http://ewriteonline .com/

Discussion Questions

1. Your documentation team has developed an online help supp ort for instant
messaging clier1t. You are hiriI1g a usability testing firm to eva luat e the on­
lme help and recommend impro vemen ts. Prepare a contract that specifies
how you want the usability test to be performed and what reports you would
like to receive. Your sched ule gives them one week to prepare the stud y, one
week to run it, plus one week to wr ite up the final report. In your cont ract,
specify the details of your requirements, includmg such information as the
number of subjects required, test plans, and the types of reports.

2. What are the advantages of online documentation?

3. Defend the use of online communities for user assistance. Ensure you r argu­
ment is well-rounded by pointing out both the weaknesses and strengths.

4. Describe the role user documentation and online help play in the lifecycle of a
piece of software . Decide when these help ma terials should be crea ted in the
softwar e design cycle.

5. v\lhat is meant by the “inverted pyramid” style of writing? Why is this style
used in writing for the web?

512 Chapter 14 Documentation and User Support (a.k.a. Help)

References

Baker, Mark, Every Page Is Page One: Topic-Based Writing for Technical Communication and
the Web, XML Press (2013).

Baron, Naomi, Words on Screen: The Fate of Reading in a Digital World, Oxford University
Press (2015).

Bernstein, Michael S., Little, Greg, Miller, Robert C., Hartmann, Bjorn, Ackerman,
Mark S., Karger, David R., Crowell, David, and Panovich, Katrina, Soylent: A word
processor with a crowd inside, Co1n1n1111ications of the ACM 58, 8 (2015), 85- 94.

Bunt, Andrea, Duboi s, Patrick, Lafreniere, Ben, Terry, Michael, and Cormack, David,
TaggedComments: Promoting and integrating user commen ts in on line application
tutorial s, Proceedings of the ACM Conference on Hu,nan Factors in Con1puting Systen1s,
ACM Press (2014), 4037-4046.

Chen, Chih-Ming, and Chen, Fang-Ya, Enhancing digital reading performance with a
collaborative reading annotation system, Contputers & Education 77 (2014), 67- 81.

Chi, Pei-Yu (Peggy), Liu, Joyce, Linder, Jason, Dontcheva, Mira, Li, Wilmot, and
Hartmann, Bjorn, DemoCut: Generating concise instructional videos for physical
demonstrations, Proceedings of the Syrnposium on User Interface Software & Technologi;
(UIST), ACM Press (2013), 141- 150.

Chi, Pei-Yu (Peggy), Ahn, Sally, Ren, Amanda, Dontche va, Mira, Li, Wilmo t, and Hartmann,
Bjorn, MixT: Automatic generation of step-by -step mixed media tutorial s, Proceedings of
the Sy1nposiu1n on User Tnterface Software & Technology (UlST), ACM Press (2012), 93-102.

Cockburn, Andy, Gutvvin, Carl, Scarr, Joey, and Malacria, Sylvain, Supporting novice
to expert transitions in user interfaces, ACM Con1puting Surveys 47, 2, artic le 31
(November 2014).

Earle, Ralph H., Rosso, Mark A., and Alexander, Kathryn, User preferences of softwa1·e docu­
mentation genres, Proceedings International Conference on Docun1entation, ACM Press (2015).

Frampton, Beth, Use as directed: Deve loping effective operations and maintenance
manuals, Intercam, STC Oune 2008), 6-9.

Gentle, Aru1e, Conversation and Cornrnunity: The Social Web for Docunzentation, 2nd Edition,
XML Press (2012).

Hackos, JoAnn, Changing times, changing skills, Camn1unication Design Quarterly 3,
2 (Febn 1ary 2015).

Ha iley, David, ReaderCentric Writing for Digital Media: Theory and Practice, Baywood
Publishing Company (2015).

Handley, Ann, Everybody Writes: Your Go-To Guide to Creating Ridiculously Good Content,
Wiley (2014).

Hwang, T. K. Philip, and Yu, Horng-Yi, Accommodating both expert users and novic e
user s in one interface by utilizing multi -laye r int erface in compl ex function products,
in Rau, P. L. P. (Editor), Internationalization, Design, HCJI, LNCS 6777, Springer-Verlag
(2011), 159- 165.

Johnson, Jeff, Designing with the Mind in Mind, 2nd Edition, Morgan Kaufmann (2014).

Johnson -Eilola, Johndan, and Selber, Stuart A. (Editors), Solving Proble111s in Technical
Con1mu11ication, University of Ch icago Press (2013).

References 513

Katz, Joel, Designing Infornzation: Hurnan Factors and Cornnzon Sense in Information Design,
Wiley (2012).

Kohl, John R., The Global English Style Guide: Writing Clear, Translatable Docun11:ntationfor
a Global Market, SAS Institute (2008).

Konnikova, Maria, Being a better online reader, The Ner.v Yorker CTuly 16, 2014).

Lafreniere, Ben, Grossman, Tovi, and Fitzmaurice, George, Community enhanced
tutorials: Improving tutorials with multiple demonstrations, Proceedings of the ACM
Conference on Hu1nan Factors in Con1puting Syste1ns, ACM Press (2013), 1779-1788.

Li, Wei, Grossman, Tovi, and Fitzmaurice, George, CADament: A gamified multiplayer
software tutorial system, Proceedings of the ACM Conference on Human Factors in
Computing Systen’IS, ACM Press (2014).

Lount, Matthe, and Bunt, Andrea, Character izing web-based tutorials: Exploring
quality, communi ty, arid showcasing strategies, Proceedings International Conference
on Docurnentation, ACM Press (2014).

Mangen, Anne, Walgermo, Bente R., and Br0nnick, Kolbj0m, Reading linear texts on
paper versus computer screen? Effects on reading comprehension, International
Journal of Education Research 58 (2013), 61-68.

Matejka, Justin, Grossman, Tovi, and Fitzmaurice, George, Ambient help, Proceedings of the
ACM Conference 011 Hu1nan Factors in Computing Syste,ns, ACM Press (2011), 2751- 2760.

Mehlenbacher, Brad, Multidisciplinary and 21st century communica tion design, Proceedings
International Conference on Docu1nentation, ACM Press (2009), 59-65.

Mohr, Peter, Kerble, Bernhard, Donoser, Michael, Schmalstieg, Dieter, and Kalkofen,
Denis, Retargeting technical documentation to augmented reality, Proceedings of the
ACi\1 Conference 011 Human Factors in Computing Systems, ACM Press (2015), 3337- 3346.

Nielsen, Jakob, F-shaped pattern for reading web cont ent, Jakob Neilsen’s Alertbox
(April 17, 2006). Available at http:/ /www.useit.com/ aler tbox/reading_pa ttern.html.

Novick, David G., Elizalde, Edith, and Bean, Nathaniel, To,-vard a more accurate view of
when and how people seek help wi th computer applications, Proceedings international
Conference on Docurnentation, ACM Press, Ne,-v York (2007), 95-102 .

Novick, David G., and Ward, Karen, What users say they ,-vant in documentation, Proceedings
International Conference on Docurnentation, ACM Press, New York (2006a), 84-91.

Novick, David G., and Ward, Karen, Why don’t people read the manual, Proceedings
International Conference on Docurnentation, ACM Press, New York (2006b), 11-18.

Pearson, Jennifer, Buchanan, George, and Thimbleby, Harold, Designing for Digital
Reading, Synthesis Lectures on Information Concepts, Retrieval, and Services, #29,
Morgan & Claypoo l (2014).

Pew Research Center, £-Reading Rises as Device Oivnership Jun·1ps (January 2014).
Available at http:/ /pe,-vinternet.org/Reports/2014/E -Reading -Update.asp.

Pinker, Steven, The Ser1Se of Style: The Thinking Person’s Guide to Writing in the 21st
Century, Viking/Penguin Group (2014).

Redish, J.C ., Technical communication and usability: Intert wined strands and mutual influ­
ences, IEEE Transactions on Professional Com1nunicati.on 53, 3 (September 2010), 191- 201.

Redish, Janice (Ginnyt Letting Go of the Words: Writing Web Content That Works, 2nd
Edition, Morgan Kaufmann (2012).

514 Chapter 14 Documentation and User Support (a .k.a. Help)

Sherwin, Katie, Pop-ups and adaptive help get a refresh , Nielsen/Norman Group (March
15, 2015). Available at http:/ /www.nngroup.com/ articles/pop-up-adaptive-help/.

Stone, Robert W., and Baker-Eveleth, Lori, Students’ expectation, confirmation, and continu ­
ance intention to use electronic textbooks, Co1nputers in Hu1nan Behavior 29 (2013), 984-990.

Strunk, Jr., William, White, E. B., and Angell, Roger, The Elements of Style, 4th Edition,
Allyn & Bacon, Ne,v York (2000).

Tebeaux, Elizabeth, and Dragga, Sam, The Essentials of Technical Co,nmunication,
3rd Edition, Oxford University Press (2014).

van Loggem, Brigit, User documentation: The Cinderella of information sys tem s, in
Rocha, A., et al. (Editors), Advances in Information Systems and Technologies, Springer ­
Verlag (2013), 167-177.

Wakkary, Ron, Schilling, Markus Lorenz, Dalton, Matthew A., Hauser, Sabrina,
Desjardins, Audrey, Zhang, Xiao, and Lin, Henry W. J., Tutorial authorship and hybrid
designers: The joy (and frustration) of DTY tutorials, Proceedings of the ACM Conference
on Hunzan Factors in Conzputing Systenzs, ACM Press (2015), 609-618.

Wang, Cheng -Yao, Chu, Wei-Chen, Chen, Hou -Ren, Hsu, Chun -Yen, and Chen, Mike
Y., EverTutor: Automatically creating interactive guided tutorials on smartphone s by
user demonstration, Proceedings of the ACM Conference on Hun1an Factors in Computing
Systerns, ACM Press (2014), 4027-4036 .

Wolf, Mary Anne, Proust and the Squid: The Stan; and Science of the Reading Brain, Harper (2007).

Woolf, Beverly, Building Intelligent Interactive Tutors: Student-Centered Strategies for
Revolutionizing £-Learning, Morgan Kaufmann, San Francisco, CA (2008).

Yoon, Dongwook, Hinckley, Ken, Benko, Hrvoje, Guimbretiere, Fran<;ois, Irani, Pourang, Pahud, Michel, Gavriliu, Marcel, Sensing Tablet Grasp + Micro -mobility for Active Reading. Proceedings of the 28th Annual ACM Syn1posiun1 on User Interface Software & Technology (2015), 477-487.

Zinsser, William, On Writing Well, Thirtieth Anniversan; Edition, Harper Collins,
New York (2006).

This page intentionally left blank

/
//

••

/

••

CHAPTER OUTLINE

15.1 Introduction

15.2 Five-Stage Search Framework

15.3 Dynamic Queries and Faceted Search

CHAPTER

•· The gods of the earth and sea
Sought through nature to find this tree.

But their search was al l in va in:

There grows one in the human brain. ”

William Blake
Songs of Innocence and Experience , 1789

15.4 Command Languages and 11Natural” Language Queries

15.5 Multimedia Document Search and Other Specialized Search

15.6 The Social Aspects of Search

517

518 Chapter 15 Information Search

15.1 Introduction

The way users conduct searches has undergone dramatic changes over the past
two decades (Hearst, 2009, 2011; Nudelman, 2011; Wilson, 2011; Russell-Rose and
Tate, 2013). Searching vast archives is now feasible for a broad spectrum of users,
ranging from children preparing school reports to researchers looking for up-to­
date results or experts to consult. This chapter focuses mainly on web or database
search of text and multimedia collections, but many principles described here also
apply to searching within a document or an address book.

The terminology swirl in this domain is especially colorful. The older terms
information retrieval (often applied to bibliographic and textual document sys­
tems) and database n1anagement (often applied to more structured relational data­
base systems wi th orderly attributes and sor t keys) are being pushed aside by
newer notions of information seeking, filtering, collaborative filtering, sensemaking,
and visual analytics. Users often alternate these several strategies for finding
information, a behavior that has been called berry picking. Computer scientists
now focus on the huge volumes of available data and talk about data mining or
deep learning. Information seeking differs from re-finding something that has
been seen before as users might prefer navigation over search (e.g., following
links , navigating a file structure or history), especially when users do not
remember spec ific keywords to use in a search (Chapter 8).

This chapter reviews interfaces appropriate for first-time or intermittent ver­
sus frequent computer users and also for task novices versus experts. Providing
powerful search capabilities without overwhelming novice users remains a
chalJenge, usuall y addressed by providing a simp le search interface with a link
to the advanced search interface (Figs. 15.1 and 15.2). The simple interface con­
sists of a single field in which to enter terms and a button to start the search. As
users gain experience with the interface, they may request additional

features

by adjusting control panels. Knowledgeable and frequent users may need a
wide range of search tools with many options that alJow them to compose, save,
replay, and revise increasingly elaborate query plans as they continue their
information gathering over hours or days.

See also :

Chapter 8, Fluid Navigation

Chapter 9, Expressive Human and Command Languages

Chapter 16, Data Visua lization

Many figures showing search interfaces in other chapters

15.1 Intro du ction 519

~ LIBRARYOF
~ CONGRESS

j ASKA UBRARIAN J ( OIGITAt.COtLECTIOHS J ( LIBRARYCATALOGS ‘ 5earch Search Loe.gov m

LIBRARY OF CONGRESS
ONLINE CATALOG

)’, I✓.( .~ j \
I .• – ~

,, I :\ · >’) … ”/. ‘ ‘ r… ‘
0 LC Onliot.c,tllogJ:lomt
o About thl C1ta!oa
o FmuonUv Atktd Quutiom
o Search/Browse Hero
0 PrlnttSava(EmaH HtlP

Search

o-..
0 Myanopd Syd,

0 Kevwgcd Starch

Your Account

0 Apgp,otlnfg

o Aern!/11! Hilo

FIGURE 15.l

Library of Congress Online Catalog

‘–1 L_C 0r1ne __ ea_ 1a_1og_ 0uld< __ s.._,.,, ___ _.l _s.a_,.,,_.I 8towM Adva"Cfd Sa.,.,, t

C-ontalns 18 mtllon ea1a10g reootds f0t bOokS. serills. ~s. maps. musle, reootdings, Images, and e1ec:ct0nle
ret

AddlUonal Catalogs & Research Tools

The LC On1lne Catalog Is lhe main access point oo lhe Ubrary’s collec:tlons. CIiek on the links below to use spedalzed calalogs and
IOols tl\at provide 80Cefl tc adclilional LC l’eSOl.#eeS:

e CqoyrSghl; Office Cataloo
U.S . Copyright regi54rations and ownerstiip
document&,. 1978-presen

LC AUthorit!n
LC authOtilY hNdit’lgs Sol’~. namtS, lilel,

~ and nameltillo combinations

llm, ~ E·BISQLIWI Oollna CalalA9 1″‘~ Subscription and frte databas.i, ejoumals. and
~ eboo’ks aoceu ible at the Library

The home page of the U.S. Library of Congress On line Catalog (https ://catalog .loc
.gov) show s the simple search box pro mine ntly placed at the top of the page and
provides alterna t ive means of finding items of interest in the dive rse co llecti ons.
Advanced search interfaces are provided to accommodate experienced searchers.

Web search engines have greatly improved their performance by making use
of statistical rankings, the information latent in the web’s link structure, and
user’s interactions (Liu, 2009). Thanks to the redundancy of informa tion on the
web, resu lts almost always return some relevant documents, and they allow
users to find answers directly or by following links. For example, to find an
expert on information retrieval, users might first find papers on that topic, lead­
ing to iden tifying a major journa l publication, the edito rs of the journal, and
their personal webpages. Da tabase searches are widespread as the general pub­
lic turns to the web to reserve travel packages, shop for groceries, search digital
libraries of children’s books, and more. Specialized databases also help lawyers
find relevan t court cases and scie11tists locate the scientific da ta they need. Con­
tinuing evalua tion will lead to further progress (Schuth et al., 2015).

To facilitate discussion, a few terms need to be introduced. Objects of interest,
such as movies to rent, airline flights to reserve, flowers or books to purchase, or
reviews to read, are stored in struc tured relational da tabases, textual documen t
libraries, or multimedia document libraries or mentioned in unstructured web
documents. A structured relational database consists of relations, which have items
(also called records), and each item has multiple attributes (often called fields),
which each have attribute values.

520 Chapter 15 Information Search

~ LIBRARYOF
~ CO NGRESS

ASKA LIBRARIAN ( DIGITALCOUECT IONS ‘ ( UBRARYCATALOGS search Search Lot.gov

LC Online Catalog

Advanced Search

Se:arch

I user Qe rface j ( as a plwase ; with in

)=-====_.===- cG”‘K=EY””)===~ ,—~

® ANO O OR 0 HOT

video ( al of these ; within
f:=-====_.=== (G=K,;EY=)==== ,’; ___ _,

0 ANO O OR C!JNOT

~ ywood I ( •• or these ; ‘ with in
~-:==;’ _.;=:;:==(:::::GK;;:EY~) ==== ,,—-‘

C Remove UrnilS

Ynr Publis h•dl”Cn ato’d

0 Yetar ( Al Vear, ; 1 0 From To I
Location In the Library

M Locations
.General Coleclions
,Roferenee Colltetion5 , AU
African Reference Col ectlon
Afriea,VMiddlc ENtem

Place of PublteatJon

—~ •

!~~=g~=;=:-________________ ·1
Tyi,9 of Materia l

Al Types
Al Text (Boot

.. Lang­
Engfoh
Abkhat
Achinese

“”‘”

Reieord• per page: ( 25 ;

Advanced SH tch Tips

,,

1. Ent.of your soarc:h tiomi(s) in !ht Soarch box(os):

• C.pilalizMion dots riot “”1tlr .

FIGURE 15.2

BnzwH Advaoctd StflJih Kl)’WP{d S11rsh

Seard”I Hi5tory I Account lrlo I Help I LC Authoritiesd:J

I Clear I ! Search I

The advanced search interface of the U.S . Library of Congress On line Catalog
(https://catalog.loc.gov). The entire page is now dedicated to search controls and
tips. Using check boxes, text fields, and menus, users can compose Boolean queries,
restrict the search scope to a subset of the collections, and app ly fi lters based on
metadata. Regular users sign up for an account to save results and keep a search
history to facilitate re-finding.

15.1 Introduction 521

A libran; consists of a set of collections (typically up to a few hundred collections
per library) plus some descriptive attributes or n1etadata about the library (for example,
name, location, owner) . Each collection also has a name and some other descriptive
attributes (for exan1ple, location, media type, curator, donor, dates, and geographic
coverage) and a set of items (typically 10 to 100,000 items per collection). Item s in a
collection may vary greatly, but usually a superset of attributes exists that covers all
the items. Attributes may be blank, have single values, ha, re multiple values, or be
lengthy texts. A 1nultin1edia library consists of collections of items that can contain
images, scanned documents, sound, video, animations, datasets, and so on. A collec­
tion is typically owned by a single library, and an item typically belongs to a single
collection. Digital libraries are generally sets of carefully selec ted and ca talog ed collec­
tion s, while digital archives tend to be more loosely organized. Directories hold meta­
data about the items in a library and point users to the appropriate locations (for
example, the NASA Global Change Master Directory helps scientists locate datasets
in NASA’s many archives). Items in unstructured collections (like the web) have no or
very few athibutes. Th ese may includ e only format or date created. Tools are appear­
ing that extract features automatica lly, but as this dynamically created metadata
becomes available to interface designers, accuracy is often an issue.

Tasks can be decomposed into browsing or searching (Russell-Rose and Tate,
2013). Tasks can range from specific fact findjng (or known-item search), where th ere
is a single readi ly identifiable outcome, to more extended fact finding with uncer­
tain but replicable outcomes. Relativel y unstructured tasks include exploration of
the availability of information on a topic, open-ended browsing of known collec­
tions, or complex analysis of problems and are also referred to as exploratory searches.

H ere are some examples of tasks:

• Specific fact find ing (known-item search)

\A/hat are the houses available for sale near Annapolis, Maryland, with three
bedrooms?
H ow to change a fla t tire?

• Extended fact finding

\A/hat videos are by th e author of the book Peace Is Every Step?
How do Maryland and Virginia counties compare on the un emp loymen t rate
since the last election?

• Exploration of availability

\A/hat genealogical information is available in the Nationa l Archives?
Are there recent survey papers on voice recognition in the ACM digital Library?

• Open-ended browsing and problem analys is

Does the Mathew Brady Civil War photo collection show the role of women
in that war?
Are there promising new treatments for fibromyalgia that might help my patient?

522 Chapter 15 Information Search

Once users have clarified their information needs, the first step toward satis­
fying those needs is deciding where to search. The conversion of information
needs (stated in task-domain terminology) to interface actions is a large cogni­
tive step. Once this is dorte, users can express these actions in a query language
or via a series of selections.

Supplemental findin g aids such as description of the content of collections,
tables of contents, and explanations of the indexes and subject classifications can
help users pursue their information needs. Careful understanding of the task
and of previous and potential future search requests helps provide hot-topic
lists and useful classification schemes. The U.S. Congressional Research Service
maintains a list of approximately 80 hot topics covering current bills before Con­
gress and has 5,000 terms in its Legislative Indexing Vocabulary. The National
Library of Medicine maintains tl1e Medical Subject Headings (MeSH), witl1
more than 27,000 items in a 12-level hierarchy, and the Gene Ontology Database
has more than 15,000 genes organized in a 19-level hierarchy, with many genes
appearing at m11ltiple nodes.

Section 15.2 presents a five-stage search framework to guide search interface
design with many examples. Section 15.3 focuses on dynamic queries and fac­
eted search. Section 15.4 reviews search in multimedia documents and other
spec ialized cases, and Section 15.5 covers the social aspects of search .

15.2 Five-Stage Search Framework

In designing the advanced search interface, a five-stage search framework may
help to coordinate design practices and satisfy the needs of first-time, intermit­
tent, and frequent users. The five stages of acti.on, illustrated more fully in
Box 15.1, are:

1. Form.ulation: Expressing the search

2. Initiation of action: Launching the search

3. Review of results: Reading messages and outcomes

4. Refinement: Formulatit1g the next step

5. Use: Compiling or disseminating insight

Information seeking is an iterative process, so the five-stages can be repeated
many times until users’ needs are met. Users may not see all the componen ts of
the five stages, but if they are unsatisfied with the results, they should be able to
have additional options and change their queries easily.

15.2 Five-Stage Search Framework 523

BOX 15. l
Five-stage framework to clarify search user interfaces.

1. Formul ation

• Use simple and advanced search.

• Limit the search using structured fields such as year, media, or location.

• Recognize phrases to allow entry of names, such as “George Washington;’

• Permit variants to allow relaxation of search constraints (e.g., phonetic
variations).

• Control the size of the initial result set.

• Use scoping of source carefully.

• Provide suggestions, hints, and common sources.

2. Initi ation of action

• Explicit actions are initiated by buttons with consistent labels (such as “Search”).

• Implicit actions are initiated by changes to a parameter and update results
immediate ly.

• Guide users to successful or past queries with auto-complete.

3 . Revi ew of res ults

• Keep search terms and constraints visible.

• Provide an overview of the results (e.g., total number).

• Categorize results using metadata (by attribute va lue, topics, etc.) .

• Provide descriptive previews of each resul t item .

• Highlight search terms in results.

• Allow examination of selected items .

• Provide visualizations when appropriate (e.g ., maps or timelines) .

• Allow adjustment of the size of the resu lt set and which fields are displayed .

• Allow change of sequencing (alphabetical, chronologica l, relevance ranked, etc.).

4. Refinement

• Guide users in progressive ref inement wi t h meaningful messages .

• Make changing of search parameters convenient.

• Provide related searches .

• Provide suggestions for error correction (wi t hout forcing correc ti on).

5. Use

• lmbed actions in results when possible.

• Allow queries, settings, and results to be saved, annotated, and sent to
other appl ications .

• Explore co llecting explicit feedback (ratings, reviews, like, etc.).

524 Chapter 15 Information Search

15.2.1 Formulation
The formuJation stage includes identifying the source of the information (i.e.,
where to search). Users may want to limit their searches for flights to certain
travel sites, or limit a text search to help documents and not the entire web
(Fig. 15.3), or search only women’s clothing on a shopping website (e.g., wom­
en’s shirts). Also called scoping, this limitation of the sou rce can lead to better
results but also lead to failures when the constra int remains active and users
forget about it. Clearly displaying the source of information is important.

Even if it is technically feasible, searching all libraries or the entire web may
not always be the best approach. Users often prefer to limit the search to a spe­
cific library or collection in a library ( e.g., within the manuscript collections of the
Library of Congress or within Wikipedia). Keywords or phrases can be specified,
and structured fields such as year of publication, volume number, or language
can be used to further limit the search scope. A text box and a few menus may be
enough in most cases, but form fill-in (Section 8.6) allows users to specify more
detailed searches in databases (e.g., search for nonstop flights between three local
airports and New Orleans across a range of possible dates).

In database searches (e.g., Fig. 15.2), users often seek items that contain mean­
ingful short phrases (“Civil War,” “Environmental Protection Agency,” “carbon
monoxide”), ar1d multiple-entry fields can be provided to allow for multiple
phrases. Searches on phrases have proven to be more accurate than searches on
individual words. Phrases also facilitate searching for names (for example, a search
on “George Washington” should not turn up “George Bush” or “Washington,
DC”). If Boolean operations, proximity restrictions, or other combining strategies

A Hor-e Mail Search News Soorts Finance Weather Games Ans\•,ers Screen fl,ck.r Mot> ‘e flore…,

YAHOO!
HELP boolean queries Seuch Hetp Search Web !] Sign In ir’iii Mail

Yahoo Help Community Forums: Join the discussion
Have a question . need an answer , or just want to join in on discussions about Yahoo product s? lhe Yahoo Help Community Forums and Yahoo Answers
Fon.ms are where the Yahoo oomm.m ity and Yahoo agents are active(y answering questions just like yours.

Ads in Yahoo Search results
Algori thmic Web searches. Provider content that Yahoo beiews is reievant to your qJety . Content from advertisers who pay to appear in Yahoo results.

Report abuse or spam on Yahoo
In Aiclcr, under the •Additional info· section of the image in queslion . dick Aag photo In Yahoo Mal , from the email in question, click Spam and select a …. …,_

FIGURE 15.3
The Yahoo! help search box has two buttons of different colors to search two
different sources of information: purple for searching the help information and
blue for searching the web. Pressing the purple button “scopes” the results to the
help information only and shows results below a purple banner. Searching the web
jumps to a different page (the normal search) that reuses the blue button color,
helping users keep track of which source of information they are searching.

15.2 Five-Stage Search Framework 525

are specifiab le, users should also be able to express them. Users or service pro,rid­
ers should additionally have control over stop lists (which typically filter out from
the search terms common words, single letters, and obscenities).

Whe11 users are unsure of the exact value of the field (tl1e terms to be searcl1ed
for or the spelling or capitalization of a name), the search constraints can be
relaxed by allowing variants to be accepted. In a textual-document search,
advanced search interfaces may give users control over variant capitalization
(case sensiti vity), stemmed versions (the keyword “teach” retrieves words with
variant suffixes such as “teacher,” ” teaching, ” and “teaches”), partial matches
(the keyword “biology” retrieves “sociobiology” and “astrobiology”), phonetic
variants from soundex methods (the keyword “Johnson” retrieves “Jonson,”
“Jansen,” and “Johnsson”), synonyms (the keywor d “cancer” retrie ves “malig ­
nant neoplasm”), abbreviations (the keyword “IBM” retrieves “International
Business Machines,” and vice versa), and broader or narrower terms from a the­
saurus (the keyphrase “New England” retrieves “Vermont,” “Maine,” “Rhode
Island,” “New Hampshir e,” “Massachus etts,” and “Connecticut”).

v\lhen searching il1 a simple list of items (e.g., searching a name in a list of con­
tacts), the result list can be displayed as users type. The list shrinks rapidly and
users can select the wanted item without finishing typing the name. When the
collection is large, auto-completion can be app]jed to the search term instead, reveal­
ing most common search phra ses that match the text already typed (Fig. 15.4). The

.. , ATIT • fl J4 ‘”

TOP NAME MATCHES

Bettway Plua Hardware …

PladeHe Jean

Plaisant Alain

Plaisant Alexis

Plalsant Cothcrinc …
a z e r t y u o p

q s d f g h j k I m

owxcvbn’@

123 • i ··*·

Al. hehrj

helmet in AJ Departments

helmet in Outdoor Recreation

ht lmet In Toys & Games

helmet in Automotive

motorcycle helmet

stormtrooper helme1

foo!bal helmet

helmet light

welding holme1

helmet camera

(bl (a) (c)

FIGURE 15 .4
Auto-complete suggestions can speed data entry and guide users toward successful
queries.

(a) In a mobile phone address book, typing one character filters t he list to all names
that contain that character, and the list is updated continuously as users type.

(b) Typing “helm” in Amazon’s search box shows suggestions for “he lmet light”
or “weld ing helmet” but also suggestions to narrow the scope of the search to
relevant departments.

(c) On the Adobe website, suggestions include products (e.g., typing the beginning
of the word “video” suggests several video-editing tools).

526 Chapter 15 Information Search

auto-complete list is updated as users continue typing, which helps them
recall terms of interest, limits misspelling, and speeds up the query initiation
process.

Mobile applications may use contex t information such as location to narrow
down the auto-completion suggestions. For example, searching on a map can
narrow the history of previous searches to the ones relevant to the current loca­
tion. This may allows users to find the exact location of a doctor’s office for a
follow-up visit even if they don’t remember the name or exact address of the
doctor – an impressive combined use of location, history and au to-completion.

For regular users who want additional control, advanced command lan­
guages can be offered to search databases (e.g., SQL; see Section 15.4). On the
other hand, new users will benefit from the reading of typi cal phrases (which
can be placed next to the search box), direct links to often searched items (e.g.,
sales or popular topics), and carefully designed tips. The seamless integration of
search with navigation and browsing will allow users to switch to menus of
choices when they are not able to come up with search phrases or to review
sample materials to better understand what is available even before tl1ey start
composing their query (see also Section 15.3 and Fig. 12.1 of NASA’s Earthdata
search interface).

15.2.2 Initiation of action
The second stage is the initiation of action, ,,vhich may be explici t or impl icit.
Many systems have a searcl1 butto11 for explicit initiation. A magnifier glass has
become the de facto standard icon for search when space is limited, but pressing
the Enter key on a keyboard or pausing during spoken interaction may be the
only thin g needed to initiate the search.

An appealing alterna tive is implicit initiation, in which each change to any
component of the formulation stage immediately produces a new set of search
results. Dynarnic queries in which users adjust query widgets to produce con­
tinuous updates (Shneiderman, 1994) led the way in demonstrating the benefits
of implicit initiation, and they have been widely adopted in faceted browsing
(Section 15.3).

15.2.3 Review of results
The third stage is the reviezv of results, in which users review resu lts in textual
lists (Fig. 15.5) or on geographical maps (Fig. 15.6), timeline s (see HIPMUNK in
Fig. 1.7), or other specialized visual overviews of results. If no items are found,
that failure should be indicated clearly. When messages are worded carefully
(see error messages in Section 12.8) and useful sugges tion s are provided, users
are less likely to abandon their search (e.g., leave a shopping website to never
come back).

15.2 Five-Stage Search Framework 527

\Alhen results are presented in a list, it is common practice to return only
about 20 results , but larger initial sets are preferable for those with high band­
width and large display s. Previews consisting of carefully selected text samp les
(or snippets; see Fig. 15.5), human-generated abstracts, photo s, or automatically
generated summaries help users select a subset of the results for use and can

Go gle

FIGURE 15 .5

HCIL twinlist

Web Maps News Shopping Images More • Search tools

About 560 results (0.84 seconds)

Twinlist and Medication Reconciliation Interfaces
www.cs.umd.edu/hcll/sharp/twln llst/ • University of Maryland, College Park •
Novel User Interface Design for . ..• June 2015 : Story on our Twlnllst prototype for
medication reconciliation appears in User Experience, the Magazine of the User
Experience Professionals Assoc iation …. Those interlaces were built on the substratum
of a novel medication reconciliation …
You’ve visited this page many limes . Last visit: 10/19:15

Ben Shneiderman, Catherine Plaisant and HCIL Twinlist …
https://www.cs.umd.edu/ .. ./ben -shn … • University of Maryland, College Park •
Nov 21, 2013 – Ben Shneiderman , Catherine Plaisant and HCIL Twlnllst Team receive
Distinguished Paper Award at the American Medical nformatics …

Many Lists
www.cs .umd.edu/ hc i l/manylists/ • University of Maryland, College Park •
We applied the design concept from our recent comparison tools: Twinlist. It aims to
meet the following three goals: 1. Support the comparison of at least four …
You’ve visited this page 4 times. Last visit: 9/8/15

Novel user interface design for medication reconciliation: an …
www.ncbi.nlm .nih.gov/ … • National Center for Bi.:>technology Information •
by C Plaisant – 2015 – Related articles
Feb 8, 2015 • (2)Human.Computer Interaction Lab and Department of Computer …
www.cs.umd.edu/hcll/sharp/twlnllst) compared to a Control interface …
You’ve visited this page 3 times. Last visit: 9/29/15

TwinList – AMIA 2011 demo – YouTube

ft

https:l/www.youtube.com/wa tch?v=YoSxlK IOpCo
Oct 28, 2011 – Uploaded by Ca:herine Ptaisant
Twinlist is a user interface prototype developed at the University
of … on this ONC Sharpe (NCCD) project see:
www.cs.umd.edu/hcll/s harp …

• –

A Google Search result list. A summary is provided at the top (the total number of
results). Each result inc ludes preview information (or a snippet). Search terms are
highl ighted, includ ing “Human-Computer Interact ion Lab,” which is the expanded
variant of the search te rm “HCIL.” The name of the top- leve l organization was added
(he re ” Nat ional Cente r for Biotec hnology Info rmation”) to he lp users j udge the
t rust iness of the information .

528 Chapter 15 Information Search

• •

•• •

. ,

I • •’ .. ‘ ,.,
.,…,; . . .

• • .1

\ -• —

-\.
i

·~ . .,,.

••

• •

FIGURE 15.6

•’ -.. -···
.., •• • • . .,


·1
i’

. \ . ‘ .. ~ • •• • • •

• •
• • •
“• ..

“”””””‘
Annapolis MD Real Estate + ..,…..,.,,……,,..

~ “”-‘:–….,..eu ….., … o,.c-..i ..
~ • -• ,oas.-.u

S32S,S8<)

ISJ.Jlillll – ‘bdS • 2 bl • L97’$Qll

U. M•Mk, Aftn …. l …. WI>
• 110US1 ,oas.-.u
S275,000

..J’IJ ks. O..C-C..,tt»TM2,,._..,-L
• COICIOHHIMIJi
S249,000
Jbds · 2 b.J • 1,328 sqft

Searching for Annapolis on the real estate website Zillow returns a list of houses and
dots displayed on a map. The two windows are coordinated; when the cursor hovers
over a house in the resu lt list, the location of the house is ind icated on the map.
A cl ick on the house wou ld bring al l the detai ls disp layed in an overlapping window.

help them define more productive queries as they learn about tl1e contents of the
items (Greene et al., 2000). Translations may also be propos ed. Allowing user s
to control how results are sequenced (e.g., alphabetical, chronological, relevance
ranked, or by popularity) also contributes to more effective outcomes. If users
have control over the result set size and which fields are displayed, they can bet­
ter accommodate their information-seeking needs.

Highlighting the searc h terms iI1 the snippet or other preview helps users
gauge the value of results. Previous visits are noted. For websites, the URL is
partially visible, and the name of the organization is provided-when avail­
able- helping us ers gauge the tru stiness of the information. In databa se search,
the preview information might indicate which collection the item belong s to or
include a photo and important attributes (Fig. 15.6).

Additional previeiv and overview surrogates for items and collections can be cre­
ated to facilitate brow sing of results. Graphical overviews can indicat e scope,
size, or structur e and help gauge the relevance of collections (e.g., with maps,
timelines, or diagrams). Previews consisting of samp les from collections entice
users and help them define productive queries (Greene et al., 2000).

When the number of result s is very large and metadata is available, a useful
strategy is to provide an overview of the numb er of items in available categories
(see also faceted search in Section 15.3). For examp le, when searching a library
catalog, the number of books, journal articles, or news artic les can be indicated
(Fig. 15.7) and allow user s to filter the results. When no metadata is available, one
strategy is to automatically cluster the results based on content analysis (see, for

15.2 Five -Stage Search Framework 529

NCSU LIBRARIES
Summon· huma n co mputer in teraction

I He lp I About 11 chat now! Q I (§ ,c.,flsh=—- -‘-:

* ii¥ IHJ AdV8ncod Scl,rdi

@ Koop search rennemonts 0 Naw &earch

So.arch Results: Your search for human comp uter Interaction retume:t 905,538 results

Refi ne your search
I 11ems wilh fu11 text onllne
I Limit to articles from peer•

reviewed publications
l Add results beyond your

library’s collection

Y Content Type
Ally

r.., Journal Artlcle (709,418)
L S- / •B- (!196,491)
r Dissertation (318,738}

Conforenco Pl’0088ding
(187,383)

·~ S- Chapter (144,577)
Nowspap,,, Miele (92,884)

r;, Magazine Mielt (33,686)
£-‘ Trade Publjcatlon Article

(30,821)
v Journa l / eJoumal (98)

mote …

T Subject Terms
I« Ally
C’ amc:fe(111,0 78)

studios (92,536)
rosoorch (70,551)
onafy$1s (64,586)

I humans (56,493}
f computer science (54,938)

more …

‘f’ Publlcatlon Date
Any

r

Relevance : ij Et

i RecommendatJon: We foond one or more specialized collections that might heJp you.

• ACM Digital Library – CollecliDn of citations and full text from ACM joumal and newsletter
articles and conference proceedings

ACM transactio ns on computer-human Interaction
by Association for Computing Machinery
ACM transactions on ocmputer-human Interaction. ISSN 1557-7325, 1998
Humar’KX)lTlputer intera:;tlon

S: eJoumal : Ful Text Onine

If Joomal: Check Avu- lly
I”” GJournal: Ful Text Onine

Human-computer lnteractfon
Huma~pute r intera::tion, ISSN 1532-7051
System design, Compu·era, Human-machine syst ems, Psycholog ical aspects
fl”” cJoumol : Full Text OnM’,C

(I”” Journa l: Check Avallabl llty

Advances in human -computer interaction

tl
– -~’ . : ,,.

, /

Advances In humarMX>mputer Interaction, ISSN 1687-5€93
Human-computer Interaction
{I” eJOUT1al: Full Tex1 Onllne
ti eJot.Wflal: Full Text Onllne

lntemationa l joumal of human-computer Interaction

(£ Joumal: Check Avallabillty
(I’ &JOU1lal: Full Tex1 Onllne

2015 NCSU Ubraries: Summon I Powered by ·; : Surrmon ,. Personalized Search Cl, Saved Items (0)

FIGURE 15.7
A search for “human computer interaction” powered by Summon TM for a university
library cata log returns a very large number of resu lts. On the left, users can see
the number of results for categories organized by content type, subject terms, or
publication date . The box provides an overview of the resul ts, reveals how the
search was done (e.g ., here the default search does not returns dissertations), and
facilitates further refinement of the search. The menu at the upper right allows users
to sort results by relevance or by date. Help is avai lab le with a “Chat now” button,
which allows users to chat with a librarian (http: //www. lib .ncsu.edu).

example, Yippy at http :/ /www .yipp y.com). This allows users to navigate a tree
of hierarchically organized topics, but the quality and app ropri ate labeling of the
clusters are often probl emati c, so this techniqu e is losing pr opon ents.

To help user s identify items of interest , access to the full document is tisuall y
necessar y, with highli ghtin g of the terms used in the search. For large docu­
ments, autom atic scrolling to the firs t occurr ence of the keyword is helpful , as are
mark ers placed along the scrollbar to indi cate the locations of other occurr ences

530 Chapter 15 Information Search

(see Fig. 9.8). A common issue when re, ,iewing results within a document is to
have the search box hide the location of the terms found in a document.

15.2.4 Refinement
The fourth stage is refinement. Search interfaces can provide meaningful mes­
sages to explain search outcomes and to support progressive refinement. For
exampl e, corrections can be proposed, such as asking, “Did you mean fibromy­
algia?” when a term is misspelled. If multiple phrases were used, items contain­
ing all phrases shou ld be shown first and identified, followed by items containing
subsets. Progressive refinement, in which the results of a search are refined by
changing the search parameters (e.g., search phrases but also time range, loca­
tion , etc.), sho uld be made convenjen t by leaving the search terms active- along
with an easy way to clear them-instead of asking users to start from scratch
every time.

15.2.5 Use
The final stage, use of tl1e results, is where the payoff comes . Results may be
merged and saved, disseminated by e-mail, or shared in social media. Users may
want to feed the results to a bibliographic tool or be notified when new results
become available. Sometimes direct answers or actions can be embedded directly
in the result lists (Fig. 15.8), but most often search is only one of many compo­
nents of a more comp lex analysis tool. For example, powerful environments are
available for lawyers to review previous lawsuits and assemble supporting mate ­
rials for their cases. Intelligence analysts might use tools such as nSPace from
Uncharted Software (Fig. 15.9) to prepare evidence-based reports. Multiple
searches can be specified at once, and names, dates, places, and organizations are

M ., emergency, eall 911 Ol y(ll,I’ ioe.1 erne,geney number ifflmedietely
from any wtecl or W,refess pt,one . NI emergency la WIY sltueCIOn U’l81
require. lmmedlll’lte ~noo ln:lfn Ille poliOe, ftN cte,~rt!Mft1 o,
a!Ttlulance. Exa”l)IN Include.: A n,. ,

911.gov I Whan to Cati 911
– .91′.oo,;.._._… html

FIGURE 15.8

90 ~ts f

Sorted 0V ( It “-Mlldlo , )

” •– s:!ad LiK

I- “” OMCrlpl!Of’I

• Gr1p11 ~ –• Organic imc o

!!! • r o la G~• 0-n …., …. 0rg_.~

• … ‘-‘”‘I Pl’lct ….. -……. {5’.tell&) …..
……. (5′ … 11.8) ….

When possib le (and important), provide information or simp le actions without
requiring users to leave the search results page. On the left, Google Search users get
the answer to their safety-critical question at the top of the result list. On the right,
Peapod shoppers looking for groceries can specify quantity and buy directly from
t he list of results after a search on “grapes.”

~ wH~!!!d to tl>t! boafd t,, M.tvo,’ 11 lvll
c 11cil aiooort. + H11 i. excltli’d

1-•n.,-ce.u11dlarted.,clt..,..-ons-4.2..0/letdl?
• (Offl’i9dltl t~b7Mll6″f.’f17,1Qgtd-Aldwwoocl
M;tyot Rex Lutbor ttik p.-.t MoQUy nlqht to

~ ‘”‘ “IL !~'”‘”””INII.MCIIMbvniditto ~.ICIPW’lRMd
1n,ru11 to tti, tty ~,..,,, corn,n ‘”‘°”

1UP5· l/…,o nso,ce.un.h~rted.sdlw•e{an$ ◄..Z.Oll’ctdl
1• CO

h

hcliday, + tut alO!.lt tbe oi..- SU(lOOrt N’IOel toi..nd I, ber
ameffllffl Mayor Ra L11tbot

://&o!’no-nsp,ot..u.,cflalted.taftw.wefan””4..l.O/fd’.d’I?
COd7S9dX I l 9(9b7t0J6XV7

• 91 _,,, ,tf, k 1″11M.-M

.us- ,’f ~ .. ..,._ u .,_lwtl!d ••un.-.. 2 0/1″‘-th
1’•Y.,rtld1’ ) 10,:0l.\7fll ~ ’71 Nld

· dline,. .ccordi1111 to Alderwood Milycw Ile•
~flt , MCIOfdnO to Mdlr’wood ~vor 1tt11 l,J,ldlor. • M~

, …tlo last YHt filed for efitction
:/,–tHtsP)C…ufldllrted …… QCl’M-4.2.0/ff(dl’

•Y5COd759dX I l 9c9b7t036XV7CU&cd•Aldlrwood
• I’, Mol'(OI’ Re• Luthor, Or. Orlwin ~o•Hlenon, ♦

1, Hay,w 11#, lutN;lr, 0, Ol.>t,,,n SWl1XUO’I, ♦ 0,.,…,
)’ntgn, Or. ~.¥1~0 VOnlty~er. 0d
P4 //dllfl~ 11fldrl,8rtl!d tdl’w.lrf’/(111’1-‘t 2.0/ll'((tl)

d• l’5,dld/~9dlc: I t 9C9b7e036XY7b1 ;>&ccfe Allill’wood
•bo •ierv •’ttiYe it, Mt d1urdi.« ~id Re.a

~ ~vHYKt.-111\.,.diutth • co11k!Rn: Lu”‘o,,Mlw’loood
., ‘l)f ..cl eMC:tt d th~ f’Vl!IIL luthor
‘ ., ……. ~ u CNl10Ml.idtw¥.,Qll..-. l..0/fllh;.tl1

, i:Odr.19dX119t:9b7eo36XY1aJckol'”~
.c~eor •

:h• • • • O;il<1 P\dJldi.tto w• l/1/200◄ + ♦ + ♦ + M,111)f • lU1"'4>f _..Clllllnce-d tllidl1
-· _.._.._, .. mattM ..,_…,….,.__., ,…,..,..,

FGURE 15.9

15.2 Five-Stage Search Framework 531

~pie – ,. ,. . C X

• II I Mayor au; Ludlor

ii 62 Coundlnl1n >ctl1t TOtd’I

• t6 c-dlra•n :mi s.-ducd

Iii lO Cout!cill’Hn AJtll

16 AJdetwoocl O.ily News.

16 dtv council

1oms:r eounc:1
9 O!ambtr d COll,neru

SlolclliJ,n,,ood@ sd~
7 ~tOft Ubor’ltO!iff

6 Alder’wood Wftln’l~fl” COOytW!t M~k Prffl . .,.,.

Mdp – – – ,, ,. . X

, ……..

• •

,. m eountv
I l 11\lfriM er
9 Asilt(.dat y

•s..mne •·
9 Washlnr L
6 Sou;:;
‘i Cfftt.Ml’l lal s.;iart

s v»

nSpace TRIST from Uncharted Software is used by ana lys ts to produce evidence­
based reports. For this (fictitious) criminal investigation, a user is reviewing a
collection of documents (shown as icons at the top). The search history on the left
shows three searches. Names, places, and organizations have been automatically
extracted. Here the search term “Laboratory” and the person “John Panni” are
selected. Snippets are shown on the bottom left. Analysts use TRIST to define
dimensions of interest and to quickly identify documents of int erest for use in a
Sandbox (Fig. 15.1 O) (http://u ncharted.software/nspace; Chien et al., 2008).

extracted automatically. Analysts review documents and export information into
a Sandbox (Fig. 15.10) to organize the evidence they found, mark it as supporting
or refuting hypotheses, and then generate reports.

Many searches are related to searches done in the past. Users often need to
find the same information again or to continue a search started the previou s
day. A search history, bookmarking, tagging, and indication of past visits in the
results (e.g., “You have visited this page 2 times, Last visit 1/6/2016”) all con­
tribute to helping users re-find information. Keeping the visited links visually
distinct reminds users of where they have been.

Designers can apply the five-stage framework to make tl1e search process
more visible, comprehensible, and controllab le by users. The five stag es are
often repeated many times until users needs are met.

532 Chapter 15 Information Search

Q

S’ Network of Key Players

? Scandal in City Council

)
heavUy vested

FIGURE 15.10

nSpace Sandbox ® f rom Uncharted Software™ allows mu ltiple analysts to organize
and present the ev idence gat hered from research. A variety of too ls such as node
and link diagramming, automatic source attribution, recursive evidence marshaling,
and timeline construction provide support for analysis and reporting.

15.3 Dynamic Queries and Faceted Search

While form fill-in interfaces can lead users to waste time specifying queries tha t
have zero -hit or mega-hit resul t sets, early work on dynan1.ic queries and query
previeivs demonstrated the benefit of applying direct-manipulation principles to
queries (Shneiderman, 1994; Greene et al., 2000). \A/hen metadata is available,
dynamic query interfaces provide (1) a visual representation of the possible
actions (e.g., menus, sliders, or but tons to represent choices for each field), (2) a
visual representation of the objects being queried (e.g., a list of items, a map or
any other visual overview), and (3) rapid, incremental, and reversible actions
and immediate feedback. The dynamic query approach is appealing as it pre­
vents errors and encourages explora tion (Fig. 15.11). Early work on query pre­
views used bar charts to show the distribution of attribute va lues for each field. It
eliminated zero-hit queries as users could only select values leading to some

15.3 Dynamic Queries and Faceted Search 533

~@~@~ .. . . .. L!ll t1 l.!.J t1 L!ll HOTELS rLIGHTS CARS PACKAGES TRIPS Lo9,n :::

◄ Go to list v,ew

Chicago, IL

Oct 22 – Nov4

Stats

*****
✓ ****

elsJand

Change

Stlow all :I

z

• MAGH. CEHT
MILE

w :,, • • )!’

Al”P’ NO. H ,I

Popular a,eas © f””I

Cil Sightseeing

•. “II

fl Shvpp1ng

I
SI

y Nghtt fe

NIIV)’ Pier,l!I

✓ ***

$488

$185

$44 11

1111

I ihb? ~ ‘>I,

w111re, •
, . • 1:: I c::::::mll] – j

✓ ** r $116

‘ ,

S44

N ZI<" b - •

J; t, i111ndi~1!-IAilrl
!!? Piopt91,n Inc

, ~

0
• frw FAST S/OE

• The, C~;o Theatre

CHICAGO OOP ~ llD
WESrtOOP El

· • • I

Mi!lm 1m
.•,vt

!:I :::2 Cl a , M!t’ All ln!iti lUft’
‘lrf Cl> «t;iO

Review Score

6!, Excellent

6!_ Good

✓ Ol

Chicago
✓ Mediocre

Poiee

$44

r Free Breakfast

Free canoellatlon

~ Free Internet

Call

S894

cl&ar an

$44

$46

$44

FIGURE 15.11

e, o ,,

• ,.,
• Sillnt IQ1’18tiU$

Coll~f>rtP

• •

– M~ t Salelll16

Google

Cl –0- (:I Grnr,1 F11rk


~ • l:l • G)

PAINT£A’S ll’OW

! •

1111

I

\
\ G)

0 ‘S hN!dAqu

TheFttld 1,111:.tlffll * Ad o!t Pl~n .. 1,w uri, •

e
El

N
M O I

t
< m I

The hotel search interface of the Kayak travel website. After a form fill-in is used to
provide the location (Chicago) and dates, results are displayed in a traditional list
or on a map. The map provides an overview of the location of the hote ls and can
be zoomed to narrow the results. It is also augmented with a visualization of the
popular sightseeing areas. On the left, menus are avai lable to narrow down the
categorical values, and sliders are provided for numer ical values. Price is impor tant,
so the average price is provided for each category value.

resu lts and also laid the foundation for faceted browsing (that typically uses the
more compact numeric counts instead of bar charts to provide preview informa­
tion about availability). Preview information does not necessarily need to be
limited to the total coun t of items; for example, Fig. 15.12 shows an example
where the preview is tl1e average price, but users can still tell whether items are
available.

The specif ication of fields’ values can sometimes be simplified by using visual
representations of the possible values. For example, selecting dates on calendars
or using an airplane layout to select among available sea ts is useful. For house
hunters seeking sale information near the city of Annapolis in Maryland, a loca­
tion search box is needed to get started, but when a map of the area is displayed,

534 Chapter 15 Information Search

FIGURE 15.12

Take-off Chicago (CHI)

$592

$444

$296

$148

6a Ba 11a 1p 4p 6p 9p

Thu 10:24a – 2:50p

A preview of the price of available flights guides users to nar row down the time
range fo r takeoff. The preview el iminates empty result sets and avo ids high
expenses (htt p://w ww .kayak.com).

it allows the selection of neighboring cities without having to know their names.
Visual-search interfaces provide context and help users refine their needs by
providing information about categorica l or numerical values of attributes of the
resu lts. They are attractive and can reduce error messages such as “data out of
range” while providing information about data avai lability and a feeling of
thoroughness to users.

Faceted search was first demonstrated in the tool Flamenco (Yee et al., 2003.)
It tightly integra tes category browsing with keyword searching (i.e., naviga­
tion and search) . Faceted sea rch makes use of hierarchical faceted metadata
presented as simultaneous menus (Section 8.2.4) and dynamically updated
counts as a preview of the results . It allows users to navigate explicitly along
multiple conceptual dimensions that describe tl1e items and to progressively
narrow or expand the scope of the query while browsing. It shows the struc ­
ture as a starting point, organizes results in a recognizable structure, and gives
control and flexibility over the order of metadata use and over when to navi­
gate and when to searcl1. Counts are upda ted dynamically, so emp ty resul t sets
are avoided. Terms can be added to tl1e search at any time. Items can belong to
multiple categories, but all numerica l values are binned into a small number of
categories.

In the library example of Fig. 15.7, users can search for “human computer
in teraction” starting with books only and then expanding to books and disserta­
tions or narrowing down to the most recent items. In the shopping examp le of
Fig. 15.13, users can search for “REI tents” and review the results. They then can

C1tt4orlts

Sletpin9 C1p,clty

-0 1•1)trtK)n (5J
~ 2·person (7J
~ 3•per-8Ql’I (1)

• ‘·per90tl (1)

&and

D AEl (9t

Seasons

0 3 • 4•MEIWI ( I)

□ 3-~,on(8)

Best Use

features

Prlct

0 S100.00 10 ‘198.99 QI
0 $200.00 lo $409,GQ (7)

W•l9ht (Lbs)

FIGURE 15.13

15.3 Dynamic Queries and Faceted Search 535

Results for “REI tents” (9 ma1che$)

1-GWINI DGiHS:F FEAUII

Son by ( Relevance :] View: 30 60 90

••••
****” (19″‘) ****” “” ****1 .. ,
REI Pau. 2 Tttrt RO Halt Oo.e z PIiis T en.t iEI Ouarttr OOll’le 2 Tffll

$ 159.00 S219.00 1299 .00

I …….. I I ,_ .. I I ,,_. I

Faceted search interface of REI. Here users searched for “REI tents” and then
browsed different tents by selecting va lues for mu ltiple categories. The selected
fi lters are clearly indicated at the top with a black background, making it easy for
users to review the constraints and remove them.

narrow on tents for three people, narrow further by choosing backpacking, then
widen the query to show three- and four-person tents because only one tent for
three is available, and then switch brands, all the while staying in the flow and
focusing their attention on the images of the ten ts.

In some interfaces, only a subset of the characteristics of faceted browsing is
included. For example, some searches allow refinements in only one menu at a
time (e.g., the International Children’s Digital Library, http:/ /www.icdlbooks.
org) as opposed to more than one menu simul taneously, which is possible with
most shopping websites. Some interfaces reset all filters when the search terms
change in order to adapt the categories to the items being returned. Even the
counts might disappear when space is very limited (e.g., on small mobile
devices). Fig. 15.14 illustra tes how a mobile search interface can resemble faceted
browsing by including a menu of filters that slides open and partially overlaps
the result list. When users select a category va lue (e.g., condition being new),
they can see the result list underneath being updated.

536 Chapter 15 Information Search

••• oo AT&T 9

< -- golbeck

43 Results

12:0 7

amazon ….,_ ….,,;p,,,,,.

Filter by:

Book Format
Paperback

Promotion

* 72%111[}

• •

Filter 3) v

Clear All

V

V

Avg. Customer Review

*

*

*

* &Up

*** &Up

** &Up

* &Up

i International Shipping

– …. —

V

FIGURE 15.14
After searching for “Golbeck” in the Amazon app, users can scrol l through the
results or use the filter menu, which slides to the left and partially overlaps the result
list. “Fi lter (3)” indicates that three filters have applied (e.g., “Paperback” as book
format), reducing the results to 43.

The lack of cons istency between search interfaces means that users have to
discover how to search each time they use a different application. An analogy to
the evolution of automobile user interfaces might clarify the need for standard ­
ization of search interfaces. Early competitors offered a profusion of controls,
and each manufacturer had a distinct design. Some designs – such as having a
brake pedal that was far from the gas pedal – were dangerous. Furthermore,
if yo tt were accustomed to driving a car with the brake to the left of the
gas pedal and your neighbor’s car had the reverse design, it might be risky to
trade cars. It took half a century to achieve good design and appropriate consis­
tency in automobiles; let’s hope we can make the transition faster for search user
interfaces.

15.4 Command Languages and “Natura l” Language Queries 537

15.4 Command Languages and ”Natural”
Language Queries

Form filling, dynamic queries, and faceted search alJow users to specify fairly
complex queries. A subset of Boolean queries is possible (ORs between attribute
values and ANDs between attributes), but some users may want even more
control over their queries. Regular expressions allow users to specify patterns of
allowed variants (e.g., typing “*terror*” to return documents with “terrorist,”
“terrorism,” or “anti-terrorism”) (see also Fig. 9.8). The Structured Query
Language (SQL) remains a widespread standard for searching such structured
relational database systems and often is the underlying query mechanism
hidden under a more accessible front end. Using SQL, expert users can write
queries that specify matches on attribute va lues, such as date of publication,
language, or pub lisher. For example, an SQL-like command might be:

SELECT DOCUMENT#
FROM JOURNAL-DB
WHERE (DATE >= 2004 AND DATE<= 2008)

AND (LANGUAGE= ENGLISH OR FRENCH)
AND (PUBLISHER= ASIST OR HFES OR ACM)

SQL has powerful features, but using it requires days of training to reach profi­
ciency and even then users make frequent errors for many classes of queries.

Filtering with comp lex Boolean queries is made possible in commercial infor ­
mation-retrieval systems such as ProQuest Dialogr M® and OCLC FirstSearch ®.
They permit complex Boolean expressions with parentheses, but their wide­
spread adoption has been inhibited by their difficulty of use. Numerous propos­
als have been put forward to reduce the burden of spec ifying complex Boolean
expressions, but a great part of the confusion stems from informal English usage.
For example, a query such as “List all employees who live in New York and
Boston” would typically result in an empty list because the “and” would be
interpreted as an intersection; only employees who live in both cities wou ld
qualify! In English, “and” usually expands the options; in Boolean expressions,
AND is used to narrow a set to the intersection of two others. Similarly, in the
English expression “I’d like Russiarl or Italian salad dressing,” the “or” is exclu­
sive, indicating that you want one or the other but no t both; in Boolean expres­
sions, however, an OR is inclusive and is used to expand a set.

Web search with “natura l” language queries (for example, “How do I fix a
flat?”) is appealing to users, particularly when using spokerl irlteraction or
searcl1ing the web (see more discussion of human language technology in
Chapter 9). The computer’s capacity for understanding such queries is limited,

538 Chapter 15 Information Search

but the availability of extremely lar ge corpora enables search engines to find
answers not by understanding the meaning of the question but by using query
expansion, tracking large numbers of user interactions, and using statistical
methods. Most often the semblance of a natural language query is achieved
simply because other users have already asked the same question and the
ans,-vers provided by humans can be retrieved (see more discussion in Section
9.4). For example, thousands of web users have asked, “How do I fix a flat?”
and the answer can be located easily. Even if a user does not have the vocabu­
lary to form the query adequately, it is likely that a long “natural” language
query will provide useful answers. Hearst (2011) provides a nice example: “If a
searcher needs a device to connect both a Wii and a DVD player to a TV but
does not know what that device is called, a ke)rword query could fail. But the
query ‘How do I connect wii and dvd to my tv’ turns up a nearly perfect matcll
on a question -answering site, with the solution being a product called either
‘video selector’ or ‘two-way A/V switcher.”‘

Specialized corpora and sys tems restricted to narrow application domains can
also lead to acceptable results. For example, searching in legal documents with a
query such as “Find cases of tenants who have sued landlords unsuccessfully for
lack of heat,” the system can parse the text grammatically, provide synonyms
from a thesaurus (“renters” for “tenants”), deal with singulars versus plurals,
and handle other problems such as misspellings or foreign terms. Then the ana­
lyzer separates the query into standard components-such as plaintiff, defen­
dant, and cause- and finds all meaningfully related legal citations.

See Chapter 9 for more discussion of human language technology.

15.5 Multimedia Document Search and Other
Specialized Search

Interfaces to structured databases and text collections have greatly improved
and allow users to achieve great feats, but search interfaces in multimedia­
document collections are only starting to become more success ful. To locate
items such as images, videos, sound files, or animations, most systems depend
primarily on text searches in descriptive documents or searches on keywords,
tag s, and metadata. For example, searches in photo libraries can easi ly be done
by date, photographer, medium, location, or text in captions, but without
captioning or tagging, finding a photo of a particular ribbon-cutting
ceremony or a specific flower remains very difficult. Collaborative tagging of
multimedia documents is dramatically changing how users search for photos,
videos, maps, and webpages, but many important collections remain untagged.
Automatically generated metadata is not as accurate as human-generated data
but is often preferable to no metadata at all, as it can be useful to have comput­
ers perform initial filtering of results. Multimedia-document search interfaces

15.5 Multimedia Document Search and Other Specialized Search 539

that integrate powerful annotation and indexing tools, search algorithms to
filter the collections, and media-specific browsing techniques for viewing the
results lead to successful outcomes for users. Types of searches might include
the following:

15.5. 1 Image search
Finding images of things su.ch as the Statue of Liberty might be simplified when
they have already been tagged but ,,ery difficult to accomplish based solely on the
pixels of the photo. Image-analysis researchers describe this task as query by image
content, or QBIC (Datta et al., 2008; Heesch, 2008). Lady Liberty’s distinctive profile
might be identifiable if the orientation, lens focal length, and lighting were held
constant, but the general problem is difficu lt in large and djverse collections of
photos. Promising approaches are searching for distinctive local features, such as
the torch or the seven spikes in the crown, or for distinctive textures or colors, such
as red, white, and blue to locate an American flag. Of course, separating out the
British, French, and other similarly colored flags is not simple, either.

More success is attainable with searches based on similarity, where users pro,ride
an image and retrieve items wit h similar features (e.g., Google Images) or even a
sketch (e.g., Retrievr, http:/ /labs.systemone.at/retrievr/). Results are often mixed,
but errors might also help users broaden the scope of their search. When a photo
matches closely a famous painting or a unique mosaic, the results may seem quite
satisfying. Working with limited collections such as images of glass vases or blood
cells leads to good results with specialized search algorithms. Even using screen ­
shots of error message dialog boxes may be useful to find help (Yeh et al., 2011).

Photo tagging first appeared in commercial tools such as Adobe Photoshop
Album but is now widely used in online tools such as Facebook, Flickr, and
Google’s Picasa TM_ Cameras now provide location, and automatic tagging of
general im age categories is becoming practical (e.g., Fig. 15.15), as is face rec­
ogni tion. The analysis of very large repositories of photos plus text data can
also help make hypotheses about the conten t of photos (e.g., if thousands of
people mention “Statue of Liberty” in social media or webpages next to very
similar photos, it becomes possible to tag the photos with that phrase). Explor­
ing how and why photos are shot, shared, and used can further aid tagging
(Sandhaus, 2011). Overall, automatic tagging with human confirmation – and
manual tagging when automation fails – can lead to successful applications.
Rapid browsing of the results and smooth zooming are important.

15.5.2 Video search
Video content has increased dramatically, fueled by the ease of recording video
with cellphones and the availability of video-sharing services such as YouTube
and Vimeo. Many \rideos are short and have a narrow focus, so searching using
the text of the title is often effective, but identifying videos that include objects,

540 Chapter 15 Information Search

flickr You Explo,e c,eat, ~ Photos. people, or groups

Animal

Arch’itecturo

Food

Plant

– Rower 1014

– Follage 349

– Tree 37’91

Other 502

Sty!o

Text

Vehlcle

Other

camera Roll

t’J ShOwlnro

3 selected

Photostrcam Albums Favorites

Dote taken • I Magic vtew

Groups

Friends

Family

8 Prhl&Cy I!, Eolt ;¢ Share G Add to album * OowT11oad

FIGURE 15.15

CtcatJons

n 1,1
More•

••••• ::: ..

Oeusetectlon

ft ln$10U Uploodr

1i1J Delete mz:, t(/J

The ” Magic View” of flickr automatica lly genera tes top ic tags for each photo. Here
users selec ted the photos w ith flowers. Three pho tos are selected and ready to be
shared. The privacy setting is vis ible and can be changed wit h a menu .

actions, or events of interest and analyzing them remains a challenge (Snoek,
2008; Schoeffman, 2015). Video analysis builds on the advances in image analysis
but adds the challenge of tracking a person or object between scenes and recogni­
tion based on information found in multiple frames. Analysis of the text in the
scenes and speech-to-text transcripts help make large vo lumes of digital video
more searchable. Finally, once results are avai lable, automatic textual summaries
describing the features found in the video might be useful (Xu et al., 2015), but
most likely users will need to quickly review the video itself. Longer videos can
be segme11ted into scenes or cuts to allow scene skipping. Researchers are explor­
ing novel ways to browse result lists by video similarity as well (Fig. 15.16).

15.5.3 Audio search
Music-information retrieva l systems can now use audio input, where users can
query with musical content (Schedl, 2014). Users can sing or play a theme or
hook from the desired piece of music, and the system returns the most similar
items (e.g., with Shazam or Soundhound; Fig. 8.7). It is becoming possible to
recognize individual performers, such as “find Madonna.” Finding a spoken
word or phrase in databases of telephone conversations is still difficult, but it is
becoming possible, and speaker identification (also called voice biometrics) is

15.5 Multimedia Document Search and Other Specialized Search 541

.timeline

F GURE 15.16
The ForkBrowser of the MediaMill semantic video search engine (de Rooij, 2008),
which allows the user to browse the v ideo collection along various dimensions to
explo re different characteristics of the collection.

starting to emerge with systems provided by Nu ance or LexisNexis (see also
Chapter 9 on speech recognition and human language technology).

15.5.4 Geographic information search

Geographic information is increasingly used to inform search (De Sabbata et al.,
2015). Sensors on the ground or onboard vehicles pro vide the information for que­
ries such as finding all businesses within 10 minut es of an airport. Mobile devices
and digita l persona l assistants can use current location and direction of travel to
inform searches such as “Where is the closest gas station?” User interfaces that
provide map displays allow users to consider results in the context of additiona l
geographic information and knowledge, as a nearby outdoor restaurant may not
be the best option if it is too close to a noisy highway or accessible only via a toll
road. Using gazetteers to deal with names that change over time and taking into
consideration spatial synonyms are important (Samet et al., 2014), but user­
generated geographic information (such as the terms used to describe places in
collections of photos or tweets) is now making spa tial search more “natural” by
establishing the terms people use to describe places that do not have precise
boundaries: dow11town, the West End, and so on. Geographic information is

542 Chapter 15 Information Search

[ Tlm ellne l Search ]

j;Ti I ◄

rme
~,-~ l

H I I H
)

• • • • • •
• ~ . ~ ‘

Clear ( Search l

FIGURE 15.17

The Even tFlow graphica l search interface allows users to spec ify a seque nce of
events by plac ing point events or intervals ico ns on a t imeli ne. Icons are exp lained
in a separa te color legend. They can speci fy t he absence of events (which are shown
crossed out) or add tempora l constraints (http://www.cs.umd.edu/hc il/eventflow).

complex, and many challe11ges need to be addressed: deciding wl1at to show 011
the map, designing dynamic legends tha t could summarize results (Dykes et al.,
2010), and improving interaction with maps (Willett et al., 2015).

15.5.5 Multilingual searches
In some cases, users want to be able to search multilingual collections (Oard, 2009).
Current web search engines merely provide translation tools, but prototype systems
allow users to search multilingual collections of speech and/ or printed documents in
languages they do not know and provide specialized browsers for browsing results,
(e.g., http:/ /www.2lingual.com). The goal of translation systems may be to identify
documents that justify the cost of high-quality professional translation.

15.5.6 Other specialized searches
Many other search interfaces are being designed to tackle specia lized data types
such as event sequences, graphs, structure document layouts, engineering dia­
grams, and so on. The graphical search box shown in Fig. 15.17 can be used by
analysts to find patient records or student records that contain sequences of
events of interest (Monroe et al., 2013).

15.6 The Social Aspects of Search

Evans and Clu (2010) define social search as “an umbrella term describing search
acts that make use of social interactions with others. These interactions may be
explicit or implicit, co-located or remote, synchronous or asynchronous.” Users
might explicitly search restaurant rev iews in Yelp, filter results by ratings provided

15.6 The Social Aspects of Search 543

by thousand s of reviewers, or ask for direct advice from friends and colleagues
using social media sites like Facebook or Twitter (e.g., “I’m in Chicago, what jazz
venue do you recommend?”). The aggregated use by prior users can also be pre­
sented in the form of explicit filters; for example, buttons for “Most viewed” and
“Most shared” are helpful and comprehensible. Social bookmarking and ranking
allows websites such as Diggs, Reddit, or Delicious to collect recommendations
from thousands of people (i.e., the zvisdo1n of the crowd) and explicitly rank items
by popularit y.

Example s of implicit use of signals left by others include the use of page rank,
time spent on a page, mouse trails, or even social media connections by algo ­
rithms that select results or make similarity-based suggestions without the user
necessarily realizing how the social aspects of search are coming into play. Early
forms of personalized search relied on users having to create a profile with a set of
search terms automatically applied to streaming information, such as e-mail
messages, newspaper stories, or scientific journal artic les. In contrast, current
personalization relies on automatically building t1sers’ profiles based on their
shopping, posting, or interaction history (ofte11 shared between websites), which
can be aggregated and compared among groups of users to make personalized
suggestions. This is a form of in1plicit search as results are presented without any
query being specified. Corporate advertising in biogs or search portals might be
seen by some as a form of personalized implicit search, as well as news spread
by robots in socia l media (Lokot and Diakopoulos, 2016). The danger of such
filtering results may be the creation of a filter bubble that closes off new ideas,
subjec ts, products, and important information (Pariser, 2012). In general, users
are more satisfied if they can tell what information was used to arrive at the rec­
ommendations, making recommendations from friends in social media sites
more acceptable.

Collaborative filtering and recomn1ender systen1s allow groups of users to com­
bine their ra tings to help one another find interesting items in large collections
(Ekstrand et al., 2011). Each user rates items in terms of interest. The system can
then recommend new items based on similarity. For example, if Joe rates six
movies highly, the algorithms match him with other people who rated the same
six mo vies highly and recommends other movies they liked (e.g., in Netflix). A
slightly different social recommendation method is to follow the purchase of a
product with the recommendation of another product based on the fact that
other users have bought both products (e.g., in Amazon). Those strategies have
an inherent appeal, and their mechanism is understood by most users, with the
help of short explanatory sentences, so they are used extensively.

Music recommendation systems such as Pandora or Last.fm illustrate the
appeal of combining personalization and recommenders. For example, Last.
fm (Fig. 15.18) builds a profile of each user’s taste by recording the musical
tracks users listen to . User-generated tag s categorize artists and track s and
help the site generate recommendations of similar artists or tracks. While the

544 Chapter 15 Information Search

Marin Marais – Les volx humalnes

Marin Marais

v cplaisant •

Home Live Music Events Features

Passacaille
20,857 llsteners

(B.:

SCR088L£S LISTENERS

635.SK 80.9K
You have !<:robbled this aniSt 2 times.

uverv1ew I racKS Albums 1-‘lctures Similar Artists t vents taograpny :>houts

Share this artist: 0 0

Jean-Philippe Rameau (25 September 1683 • Jan Pieterszoon Sweelirck (April or May,

12 September 1764)w asone of the most 1562-0 ctober 16, 16211 was a Dutch

FIGURE 15.18

Last.fm is an examp le of onli ne radio using playlists created automatica lly . The process
starts by users select ing a start poin t (e.g., a song or artist they like); then users provide
feedback on the suggestions by clicking on the heart or skipping the track.

newly suggested tracks are playing, users provide feedback explicitly (with likes)
or implicitly (by skipping tl,e track). By combining personal preferences, whicl,
serve as starting points (such as a song or an artist), and collaborative filtering to
suggest tracks, the list quickly grows, and music can be generated all day. The
classic search form interface vanishes in favor of a more fluid approach to search.
Challenges remain when recommenders appear as a black box, leaving users
son,etitnes puzzled by the recommendations, and researchers are exploring
means to let users better guide the recommendation process (Harper et al., 2015).

Personalized search and recommendations are coming together in mobile
personal assistants (see Chapter 9) as they try to answer common questions
before users even ask them and trigger aler ts such as “It is time to leave for your
next appointment in Baltimore.”

Human-poivered question answering, such as in Yahoo! Answers or Ask.fm, allows
users to enter questions and thousands of other users to propose answers . Voting
on the quality of the answers allows the best ans,<1Ters to bubble up to the top.

Collaborative search corresponds to situations were users work together to con­
duct a search task. For examp le, remote family members might collaborate to
plan a vacation. Collaborative search is an active research area bt1t has not

Researcher’s Agenda 545

reached commercial success yet, so users are typically searching in groups with­
out the benefit of specialized tools for collaboration. Shared documents (e.g.,
Google Docs) might facilitate collecting and organizing results and can be useful
wlten comp lemented by e-mail, texting, or social media communication.
Researchers are exploring ways to manage the division of labor (e.g., one person
might be doing general search and triage, the second reviewing results and orga­
nizi.J.1g leads or evidence, and a third researching leads). Improving users’ aware­
ness of each other’s progress and making results, insight, and search histories
persistent across sessions and among collaborators are a challenge (Morris, 2013;
Shah, 2014). Future research is needed to assist teams of analysts in tracking
criminals or understanding food poisoning outbreaks. As teams grow to address
larger tasks, the social discovery framework (Shneiderman, 2011) may be useful
to suggest how to pool the collective efforts to build better thesauri or indexes, to
tag documents or objects, and to combine multiple search result suggestions.

Practitioner’s Summary

While social media is changing how information reaches users, search interfaces
remain a crucial component of many applications. Improved user interfaces to dig­
ital librari es and multimedia databases have spawned appealing new products.
Flexible queries against complex text, soW1d, graphics, image, and video data­
bases are emerging, while collaborative tagging and recommenders can eliminate
the need for a search box. Faceted search and direct-manipulation approaches to
query formulatio11 effectively combine search and browsing. Very large reposi­
tories of questions and answers are making “natural” language queries a more
effective specification method. Search and recommendation are evolving toward
each other with personal digital assistants. Advanced search interfaces provide
add itional controls and even powerful command languages for users who learn
to master them.

Researcher~Agenda

Although the computer contributes to the information explosion, it is potentially
also the magic lens for finding, sorting, filtering, and presenting the relevant
items. The need to search in structured documents, image libraries, and sound or
video files still presents grand opportunities for improved user interfaces. Better
understanding of the benefits and limitations of social and personalized search is
needed. Automatic generation of metadata is improving, but solutions that facili­
tate human confirmation and manual tagging when automation fails will likely
lead to more successful applications. Finally, providing collaborative search
interfaces will allow teams to produce richer results and generate better insights.

546 Chapter 15 Informa t ion Search

WORLD WIDE WEB RESOURCES

www. pearsonglobaleditions . com / shneiderman

• Library of Congress Catalog: https: //cata log.loc. gov; exampl e collection:
www.loc.gov /pictu res

• Dan Russell’ s blog on search research: http: //searchr esearch1.b logspot.com

• Marti Hearst’s book on search interf aces: http ://searchuserinterfaces.com

• Newsstand for news search with map query interface: http: //newsstand
.umiacs.umd. edu/

• Multilingua l search example: http: //www.2 lingual.com

Discussion Questions

1. Describe th e challenges first-time users face when using an infor mat ion­
explora tion sys tem. Propose how these challenges can be overcome.

2. Argue whether textual search interfaces should keep details of how the search
is performed hidden from the users. Decide which approach will allow the
user to ge t more accura te results.

3. Describe a framewo rk, of thr ee to seven phases, tha t will help to coordina te
design practices to satisfy the needs of first-time, intermittent, and frequent
users who are accessing a variety of textual and multimedia libraries.

4. Explain the strategy commonly used for searching multimedia archives. List
the limitations associated wi th this strategy.

References

Chien, L., Tat, A., Proulx, P ., Khamisa, A., and Wright, W., Grand Challenge Award
2008: Support for diverse analytic techniques , nSpace2 and GeoTime visual analytics,
IEEE Visual Analytics Science and Technologi; Conference (2008), 19- 24.

Da tta, R., Joshi, D., Li, J., and Wang, J. Z., Image retrieval: Ideas, influences, and trends
of the new age, ACM Co,nputing Suroey 40, 2 (2008), 1-60.

Dykes, J., Wood, J., and Slingsby, A., Rethinking map legends with visualization, IEEE
Transactions on Visualization and Co,nputer Graphics 16, 6 (2010), 890-8 99.

Ekstran d, M., D., John T. Riedl, J. T., and Kons tan, J. A., Collabo rative filtering recommend­
er systems, Foundations and Trends in Hurnan-Coniputer Interaction 4, 2 (2011), 81- 173.

Evans, B. M., and Chi, E. H., An elaborated model of socia l searc h, Tnformation Processing
& Managernent (2010), 31 pages.

References 547

Greene, S., Marchionini, G., Plaisant, C., and Shneiderman, B., Previews and overviews
in digital libraries: Designing surrogates to support visual information-seeking,
Journal of the A1nerican Society for Inforn1ation Science 51, 3 (2000), 380- 393.

Harper, F. M., Xu, F., Kaur, H., Condiff, K., Chang, S., and Terveen, L., Putting users in
control of their recommendations, Proceedings of the 9th ACM Conference on Recom1nender
Systen1s, ACM (2015), 3-10.

Hearst, M.A., Search User Interfaces, Cambridge University Press, Ne,v York (2009).

Hearst, MA ., “Natural” search user interfaces, Comrnunications of the ACM 54, 11 (2011), f:IJ-67.

Heesch, D., A su rvey of browsing models for content based image retrieva l, Multi1nedia
Tools and Applications 40 (2008), 261-284.

Liu, T.-Y., Leaming to rank for information retrieval, Foundations and Trends® in
lnfonnation Retrieval 3, 3, Now Publishers (2009), 225-331.

Lokot, T., and Diakopoulos, N ., News bots: Automating news and information dissemi­
nation on Twitter, Digital Journalism (2016).

Monroe, M., Lan, R., Morales del Olmo, J., Plaisant, C., Shneiderman, B., and Millstein, J., The
challenges of specifying intervals and absences in temporal queries: A graphical language
approach, Proceedings of the ACM Conference on Human Factors in Cornputing Systenis, ACM
Press, New York (2013), 2349-2358.

Morris, M. R., Collaborative search revisited, Proceedings of the 2013 ACM Conference on
Co,nputer Supported Cooperative Work (2013), 1181-1192.

Nudelman, G., Designing Search: UX Strategies for eCon1111erce Success, Willey (2011).

Oard, D. W., Multilingua l information access, in Bates, M., and Maack, M. N. (Editors),
Encyclopedia of Library and lnforrnation Sciences, 3rd Edition, Taylor & Francis (2009).

Pariser, E., The Filter Bubble, What the Internet is Hidingfronz You, The Penguin Press (2012).

de Rooij, 0., Snoek, C., and Worring, M., Balancing thread based navigation for targeted
video search, Proceedings ACM International Conf linage and Video Retrieval, ACM
Press, New York (2008), 485-494.

Russell-Rose, and Tate, T., Designing the Search Experience: The Infonnation Architecture of
Discovery, Morgan Kaufn,ann (2013).

De Sabbata, S., Mizzaro, S., and Reichenbacher, T., Geographic dimensions of relevance,
Journal of Docunientation 71, 4 (2015), 650-666.

Samet, H ., Sankaranarayanan, J., Lieberman, M., Adelfio, M. D., Fruin, B., C., Lotkowski,
J. M., Panozzo, D., Sperling, J., and Teitler, B. E., Reading news with maps by exp loit ­
ing spatia l synonyms, Con11nunications of the ACM 57, 10 (2014), 64-77 .

Sandhaus, P., and Boll, S., Semantic analysis and retrieval in personal and social photo
collections, Multimedia Tools and Applications 51, 1 (2011), 5- 33.

Sched l, M., Gomez, E., and Urbano, J., Music information retrieval: Recent developmen ts
and applications, Foundations and Trends® in Inforn1ation Retrieval, 8, 2- 3, Now
Pub lishers (2014), 127- 261.

Schoeffmann, K., Hudelist, M., and Huber, J., Video interaction tools: A survey of recent
work, ACM Con1puting Surveys 48, 1 (2015), Artic le 14.

Schuth, A., Hofmann, K., and Radlinski, F., Predicting search satisfaction metrics with
in terleaved comparisons, Proceedings of ACM Conference on Research and Developrnent
in Tnforn1ation Retrieval, ACM Press, New York (2015), 463-472.

Shah, C., Capra, R., and Hansen, P., Special issue on collaborative information seeking,
IEEE Computer 47, 3 (2014).

548 Chapter 15 Information Search

Shneiderman, B., Dynamic queries for visual information seeking, IEEE Software 11, 6
(1994), 70- 77.

Shneiderman, B., Social discovery in an information abundant world: Designing to
create capacity and seek solutions, lnforrnation Services & Use 31 (2011), 3- 13.

Snoek, C. G. M., and Worring, M., Concept-based video retrieval, Foundations and
Trends® in Inforn-zation Retrieval 2, 4, Now Publishers (2008), 215-322 .

Willett, W ., Jenny, B., Isenberg, T., and Dragicevic, P., Lightweight relief shearing for
enhanced terrain perception on interactive maps, Proceedings of the ACM Conference on
Hun1an Factors in Con-tputing Syste·ms, ACM Press, New York (2015), 3563-3572.

Wilson, M., Search User Interface Design, Morgan & Claypool Publis hers (2011).
Xu, S., Li, H. , Chang , X., Yu, S., Du, X., Li, X., Jiang, L., Mao, Z., Lan, Z., Burger, S., and

Hauptmann, A., Incremental multimodal query construction for video search, Proceed­
ings of the 5th ACM on international Conference on Multimedia Retrieval (2015), 675-678.

Yee, K.-P., Swearingen, K., Li, K., and Hearst, M., Faceted metadata for image search
and browsing, Proceedings of the ACM Conference on T-lu1nan Factors in Con1puting
Systerns, ACM Press, New York (2003), 401-408.

Yeh, T., White, B., Pedro, J. S., Katz, B., and Davis, L. S., A case for query by image and text
content searching computer he lp using screenshots and keywords, Proceedings of the 20th
international Conference on World Wide Web, ACM, Hyderabad, India (2011), 775-784.

Still stressed from student homework?
Get quality assistance from academic writers!

Order your essay today and save 25% with the discount code LAVENDER