Tài liệu Web Application Design Patterns- P10 doc

Furthermore, to make it easy for users to fi nd their uploaded content, allow them to tag it see the TAGGING pattern next.. ALLOW USERS TO COPY FILES FROM OTHER ONLINE SOURCES In cases

when users reach the end of the carousel, disable the right arrow, and when users revert to the fi rst item, disable the left arrow Pagination cues may also be used to indicate that users have reached the fi rst or last item in the set ( Figure 8.48 ) Additionally, a partial image of the previous or next item in the carousel may be shown ( Figure 8.49 )

Related design patterns CAROUSELS use visual effects (ANIMATIONS/TRANSITIONS), such as slide-left, slide-right, slide-up, and slide-down, to allow users to maintain visual context between items in the carousel For linear carousels, use PAGINATION indica-tors to show users ’ location within the carousel (see Chapter 6)

USABILITY ISSUES INHERENT WITH RIAS

Like any other web application, poorly designed RIAs can undermine ity and must be tested for usability In fact, there are a few inherent usability issues with RIAs designers must be aware of These issues relate to the use of the “ back ” button and bookmarking (or favorites) functionality

The “ back ” button problem Users who are not used to RIA-style web applications may not be aware that it

is possible for part of a page to update, so when they see a piece of a web page

FIGURE 8.46 Flickr uses a carousel approach in their slideshow

FIGURE 8.47 Amazon presents items in the carousel horizontally with prominent buttons at each side

change, they may think they have navigated to a new page They may then try

to click the browser’s “ back ” button to return to the previous state of the

appli-cation, which takes them to the previous page in the browser’s history instead

of the previous state of the application Although users are trying to undo a

previous action, they typically fi nd themselves completely out of their task

con-text and could potentially lose data

A common solution to this problem is to allow users to undo their actions

on the same page However, it is more important to understand users ’ natural

behavior with the application and determine if the RIA approach is

appropri-ate for the task at hand A good example is Gmail, which uses RIA for lists or

emails (i.e., “ Inbox, ” “ Starred, ” “ Sent Mail, ” etc.) and when viewing

conversa-tions (i.e., chronological thread of email exchanges) but allows users to use the

browser’s “ back ” button to return from the conversation page to the list page

( Figure 8.50 )

Usability Issues Inherent with RIAs

FIGURE 8.48 Yahoo! Food uses left and right navigation arrows for browsing Also, because it

uses a linear carousel, it uses pagination-like indicators (as dots) to indicate which “ section ” the

user is viewing

FIGURE 8.49 Pandora shows partial images of previous and next songs to indicate the

presence of additional carousel items

The bookmarking problem Because the browser’s location/address bar stays exactly the same when users select functions and change the application’s state, turning bookmarking into specifi c application views is impossible Although some clever approaches are available to address the bookmarking problem by rewriting URLs, it is typi-cally not a big issue for web applications because users do not need to book-mark specifi c states For more information on URL rewriting, see AJAX: How

to Handle Bookmarks and Back Buttons, Brad Neuberg (2005) at www.onjava com/pub/a/onjava/2005/10/26/ajax-handling-bookmarks-and-back-button.html

FIGURE 8.50 When users click on a conversation in the list view (a), Gmail takes them to a separate page (b) This allows users to click the browser’s “ Back ” button to return to the list page from the conversation page

(a)

(b)

INTRODUCTION

Increasingly, web applications are designed to encourage user participation

and sharing User participation is typically in the form of user-contributed

con-tent, where users add their own content to the application (ADD/UPLOAD

CONTENT) and describe it using tags (TAGGING) Other ways for users to

participate is by providing RATINGS and REVIEWS of content offered by the

application Many applications also involve users in promoting items by letting

them vote for their usefulness and relevance (VOTE TO PROMOTE)

To ensure that user participation leads to a trusted online community, users

have to establish an account with the application and create a USER PROFILE

Although for products and services, trust may be established through RATINGS

and REVIEWS, for users it’s important that they achieve a high REPUTATION,

especially if they want to transact online or gain respect of other online

com-munity members One aspect of reputation is based on the size of users ’

social networks Social applications, therefore, facilitate users to connect

with others with shared interests, backgrounds, and experiences (DISCOVER

NETWORK MEMBERS) Once discovered, not only can they “ friend ” them

(FRIEND LIST) and/or “ follow ” their activities online, but they can also create

groups to discuss and share common interests (GROUPS/SPECIAL-INTEREST

COMMUNITIES) Social applications also facilitate interaction among friends

by allowing them to chat in real time, send messages to each other, and write

comments in shared areas (MESSAGING); for encouraging real-time messaging

it’s also important to convey users ’ online status (PRESENCE INDICATOR)

Participation and interaction are further enhanced when users can share

pho-tos, news stories, videos, bookmarks, and other content — commonly referred

to as social objects — with their friends and trusted colleagues (SHARING)

or work together to coordinate activities and events or co-create content

(COLLABORATION)

Social Applications

CHAPTER 9 CHAPTER 9

ADD/UPLOAD CONTENT

Problem Users need to transfer content fi les, such as music, photos, presentations, and

so forth, from their own computers to those of the application providers to share them with other users

Solution Provide users a way to upload one or more content fi les In addition, if appro-priate, allow users to describe (and tag) the content and indicate their prefer-ences for who can view it ( Figure 9.1 )

Why Making uploading of fi les easier is essential for web applications that rely on users to provide content Furthermore, to make it easy for users to fi nd their uploaded content, allow them to tag it (see the TAGGING pattern next) How

In most cases, users will have the content fi les on their computer Therefore, make it easy for them to upload them from their computer When users are likely to upload several fi les at a time, such as when uploading photos, allow them to select multiple fi les and upload them all together ( Figure 9.2 )

ALLOW USERS TO COPY FILES FROM OTHER ONLINE SOURCES

In cases where users may already have uploaded their fi les (e.g., photos) on sites such as Picasa or Flickr, make it easy for them to transfer fi les directly from

FIGURE 9.1 SlideShare allows users to upload content (i.e., presentations) as well as provides a way to describe and label it to make it easy to fi nd and share In addition, to facilitate uploading of content, SlideShare offers several upload options: bulk upload, single upload, URL upload, email upload, and browser plug-in

those accounts ( Figure 9.3 ) rather than make them search for the fi les on their

computers or download them from the original site to their computers before

uploading them again

ALLOW USERS TO REMOVE SELECTED CONTENT

FILES FOR UPLOAD

It is possible for users to select incorrect fi les to upload or change their minds

about certain fi les after they have selected them Allow them to remove such

fi les ( Figure 9.4 )

Add/Upload Content

FIGURE 9.2 Flickr allows users to select and upload multiple photos at the same time

FIGURE 9.3 MyFolia allows users to import photos from their Flickr, Picasa, or Gravatar

accounts

ALLOW USERS TO SET PRIVACY PREFERENCES

Users may not want to share uploaded content or may want to limit sharing

to certain users Offer them options to specify such privacy and sharing ences ( Figure 9.5 )

KEEP USERS INFORMED ABOUT THE UPLOAD PROGRESS

Allow users to monitor the progress of content uploading by providing a ress indicator (see the DELAY/PROGRESS INDICATOR pattern in Chapter 8) This makes it easier for them to judge the time it will take to upload the fi les In addition, users can interrupt the upload if they feel that it may take longer than they had anticipated or they realize that they selected an incorrect fi le to upload

CONFIRM SUCCESSFUL UPLOAD OF CONTENT FILES

Acknowledge to users a successful upload of fi les Once fi les have completed uploading, either take users to the page where they can manage the uploaded

FIGURE 9.5 Flickr allows users

to defi ne privacy settings for uploaded photos

FIGURE 9.4 Flickr allows users to remove fi les selected for upload

fi les or keep them on the same page with the option to upload more fi les If

users would benefi t by tagging content or providing descriptions, suggest

appropriate next steps to them

Related design patterns

Because uploaded fi les may be large, especially when adding media fi les, use

of the DELAY/PROGRESS INDICATORS (see Chapter 8) pattern is relevant and

should be considered In addition, most applications that support

user-gener-ated content require that users describe them using tags (TAGGING)

TAGGING

Problem

Web applications that allow users to add content (e.g., bookmarks, photos,

music, videos, and so forth) may also want to allow them to categorize or label

their uploaded content to make it easier to fi nd later However, with all the

different ways users may want to label content, it may be impossible to

antici-pate and make available all potential labels and variations (or categories and

subcategories) For example, users may want to label personal photos with

labels such as the names of people, occasions, locations, ages, emotions, and

so forth

Solution

Allow users to label (i.e., tag ) content with any descriptive information they

desire so as to make it easy for them to fi nd it later ( Figure 9.6 ) The labels

used for tagging content should not be restrictive except when they might be

offensive to other users of the application; for example, an application might

not want to include profanity in labels

Why

The use of open-ended tags encourages a personally meaningful and

natu-ral vocabulary This makes it easy for users to fi nd items later and allows them

Tagging

FIGURE 9.6 YouTube asks users

to add tags when uploading videos

to explore and interact with content in a myriad ways (Marlow et al., 2006) For example, by allowing users to label emails (and add multiple labels to the same email), Gmail permits users to not only use tags that describe the content

of the email but also to describe actions and priorities (e.g., “ to do, ” “ important, ” “ urgent, ” and so forth) In addition, users don’t have to force-fi t items into a cat-egory/subcategory combination; they can place them in many virtual categories

at once

Application developers also benefi t from tagging because they do not have to address the whole categorization scheme (i.e., taxonomy) in advance They can rely on users ’ tags to continually create a dynamic, evolving taxonomy (also

referred to as folksonomy) 1 and use it to supplement the high-level taxonomy to facilitate navigation

Finally, tagging can encourage user participation and sharing since it can help create communities with shared interests and allow users to explore content that is tagged similar to theirs

How Adding tags to a content item should be straightforward To tag an item, let users enter keywords separated by a space or a comma (or another delimiter)

in a text fi eld Using space as a delimiter may be problematic when users want

to enter multiword tags Therefore, consider use of commas, semicolons, or other special characters as delimiters In addition, allow users to tag both the content they are adding and the content that already exists ( Figure 9.7 )

1 Thomas Vander Wal (2007) coined the term folksonomy and described it as follows: “

Folks-onomy is the result of personal free tagging of information and objects (anything with a URL) for one’s own retrieval The tagging is done in a social environment (usually shared and open to others) Folksonomy is created from the act of tagging by the person consuming the information ”

FIGURE 9.7 Flickr allows users

to add tags to photos they upload

KEEP TAGGING OPTIONAL

The main purpose of tagging is to allow users to provide some descriptive

information about content to facilitate fi nding it in the future Because the

primary user task is to add content, tagging (or providing other descriptive

information) should be optional However, users should be permitted to add

tags later

ALLOW USERS TO TAG SEVERAL ITEMS TOGETHER

For content such as photos, users may want to add the same tags to several

items Allow them to select items that will share the same tags and apply tags

to them in “ bulk ” or “ batch ” mode ( Figure 9.8 )

SUGGEST TAGS TO MINIMIZE VARIABILITY

One of the problems with tagging is that items may be tagged using seemingly

similar labels caused by typos, plurals, or minor differences in spellings (e.g.,

color versus colour) For example, one user may label an item as “ web site, ”

another as “ website, ” and yet another as “ web_site ” or “ websites ” By

suggest-ing tags and lettsuggest-ing users pick from them, the application can minimize

redun-dancy and unnecessary distinctions in tags

In addition, suggesting tags may also make users consider alternative ways

to describe content and avoid conservative labels from users new to tagging

Suggestions may be in the form of the following (Smith, 2007):

■ Recommended tags. Tags the user should consider based on popular tags,

recently used tags, and other factors

To make it easy to add suggested tags, allow users to select from a list ( Figure 9.9 )

In addition, while entering tags, suggest tags using the AUTOSUGGEST/

AUTOCOMPLETION rich-interaction pattern (see Chapter 8)

Tagging

FIGURE 9.8 Flickr allows users

to apply tags in a “ batch ” mode Users can batch photos that they want to tag and then click “ Add Tags ”

to add descriptions

to all the items in the batch

ALLOW USERS TO CHANGE AND DELETE THEIR TAGS

Users may want to change their tags because they made a mistake or have found other tags that better describe the content Also, if users have tagged

a content item to describe an action they are going to take (e.g., labeling an item “ to do ” or “ urgent ” in Gmail), they may want to remove those tags if they are no longer relevant To accommodate such needs, allow users to remove, change, or add tags to an existing item ( Figure 9.10 )

Managing tags should be possible in batch mode as well — that is, users should

be able to change or delete tags for multiple items at the same time If it would help users, allow them to replace one tag with one or more tags as well

FIGURE 9.9 Delicious both recommends tags and lists popular ones for users to consider when adding a bookmark and tagging it To use one or more tags, users just have to click on them, and those tags are populated in the “ tags ” text fi eld

FIGURE 9.10 Delicious allows users to change or delete tags by clicking “ edit ” next to the bookmarked item

Related design patterns

When tags are used for labeling items, TAG CLOUDS are usually offered as a

way to navigate and explore content (see Chapter 5)

RATINGS

Problem

With an abundance of content accessible on the Web, users are faced with the

problem of identifying relevant and useful content; this is made even more

dif-fi cult with user-contributed content that has not been reviewed or edited for

quality Applications also face a similar problem because their personalization

and recommendation algorithms often rely on user interests that are the same

when they try suggesting content to users

Solution

Allow users to indicate their likes and dislikes by rating items (e.g., movies,

music, videos, restaurants, hotels, and so forth; Figure 9.11 ) Ensure that rating

an item is quick and does not take too long or interrupt users ’ main task

Why

It is impossible for users to weed through all the available content to separate

use-ful and relevant content In addition, when buying products and services,

decid-ing among the available options can be paralyzdecid-ing (Schwartz, 2004) Therefore,

to help make decisions, users typically depend on others ’ experiences, which are

conveyed through ratings and reviews (see the REVIEWS pattern that follows)

Ratings in the form of star ratings are useful for two reasons: (1) providing

rat-ings is relatively straightforward; and (2) they provide quick, at-a-glance

infor-mation about the usefulness or quality of products, services, content, and other

items as judged by other users This makes it easy for users to at least fi lter out

content at a gross level Ratings can be used in a variety of ways For example,

■ eBay, a marketplace for buying and selling products, uses ratings to create

a detailed feedback profi le of sellers

■ Amazon, an online retailer and marketplace, uses ratings for its products

as well as its sellers

■ NexTag, a comparison-shopping application, uses ratings to show the

qual-ity of products as well as the reliabilqual-ity of sellers

An indirect measure often used to judge items is its popularity based on the

number of purchases, downloads, wish list additions, and so forth However,

this method indicates users ’ actions and behaviors, but not their satisfaction

Ratings

FIGURE 9.11 Amazon allows users to rate items using a fi ve-star rating system They also

make it clear that rating the item will help Amazon provide better recommendations

and experience with the product or service after they have purchased or rienced it Popularity and quality are not the same For example, a best-selling author may sell many copies of a newly published or to-be-published book even before anyone has had an opportunity to read and review it In addi-tion, the ratings and reviews obtained from other users of a product have been found to be useful and trustworthy to help users make purchase decisions Gauri et al (2008) concluded:

It is interesting that of all the attributes, positive customer reviews have the greatest impact on repurchase intention This is consistent across all categories (i.e., books and magazines; DVDs and videos; and fl owers and food) Even more impressive is the fi nding that number of years on the Web has the least impact on repurchase intention It suggests that stores would attract more customers by having positive customer reviews Another interesting fi nding is that it is not the total number of reviews that infl uences customer repurchase intention, but the percentage of positive reviews

How Web applications that attempt to capture users ’ feedback relatively quickly employ a star-rating approach with one star representing the lowest rating and

fi ve stars representing the highest rating; some applications allow increments

of half stars, thus increasing the range of the scale

Two types of interaction approaches are commonly used when using star ratings:

1 Separating the user ’ s ratings from the average rating. With this approach, users are shown an item ’ s average rating separately from their own rat-ing To rate an item, users are shown a set of fi ve “ empty ” stars As users hover over the stars, stars refl ecting the corresponding rating are high-lighted Users then click to assign and save a rating Users are then shown their ratings in a different color from the average rating ( Figure 9.12 ); they are also offered an option to either remove their ratings or change previously assigned ratings

2 Combining the average rating with the user ’ s ratings. By default, users are shown stars in red to indicate the current average item rating As users hover over the stars, the red stars are replaced by yellow stars (or another color) As users click to assign and save a rating, the stars are shown

in the yellow state to indicate that the item has been rated by the user ( Figure 9.13 )

Although it makes intuitive sense to separate the current rating from the users ’ ratings, both of these approaches are widely used In some instances, both approaches are used within the same application For example, Blockbuster separates the aggregated average rating from the users ’ rating on the movie

details page but combines them on other listing pages; it may be using the

“ combined ” approach on listing pages to save space and minimize clutter

ANCHOR THE RATINGS

Anchor each rating with a verbal descriptor to clarify to users what each star

rating means ( Figure 9.14 )

Ratings

FIGURE 9.12 Blockbuster shows users “ empty ” stars before they have rated an item (a) and

then in color after they have entered their rating (b)

(a)

(b)

FIGURE 9.13 Netfl ix shows the current rating using red stars (a) As users hover over the stars,

the fi ll color changes to yellow to indicate their rating (b) Once a desired rating is noted, the stars

are fi lled yellow to indicate that users have rated the item (c)

INDICATE TO USERS IF THEY HAVE ALREADY RATED AN ITEM

To ensure that users don’t waste time rerating an item, let users know if they have already rated it ( Figure 9.15 ) Although users should be allowed to change their mind and remove their ratings, to prevent abuse, they should not be allowed to rate the same item more than once

SHOW USERS A BREAKDOWN OF THE RATINGS

Average ratings are typically used to show an item’s rating However, when an item has received fewer ratings, the average rating is not very reliable For exam-ple, if an item has received a rating of fi ve stars, four stars, and one star, when showing the rating, it will display the average rating of three stars — not a true refl ection of user preference To ensure that the item’s quality is judged more accurately, show users a breakdown of the ratings ( Figure 9.16 )

ALLOW USERS TO RATE AN ITEM USING MULTIPLE CRITERIA

When an item may be judged on several criteria (e.g., ambience, quality of service), offer users additional evaluation criteria to encourage more precise evaluations ( Figure 9.17 ) Providing additional rating criteria may also identify specifi c aspects of an item that users favored (or did not favor)

FIGURE 9.14

Hulu, an application for watching TV episodes and movies, anchors each star rating with a verbal descriptor

FIGURE 9.15 Netfl ix shows the user’s star ratings in yellow and aggregated average ratings (that users have not rated) in red