Important note: This Wiki page is edited by participants of the RDWG. It does not necessarily represent consensus and it may have incorrect information or information that is not supported by other Working Group participants, WAI, or W3C. It may also have some very useful information.

TC4R Draft Report

From Research and Development Working Group Wiki
Jump to: navigation, search
NOTE: This is an internal page that is not yet ready for public review.
Please do not distribute the URI yet.

Research Report on Text Customization for Readability

NOTE: This is a rough, in-progress draft.


[status: after more of the rest of report is done, come back to this section]

Users with low vision, dyslexia, and other conditions that impact reading need to be able to customize text display in order to read effectively. While some users can read small amounts of commonly-formatted text, trying to read more text (for example, articles) causes pain, nausea, and other discomfort significant enough that the text is essentially unreadable.

... Aspects of text formatting that users need to customize include: text size, text color and background color, font face, leading/line spacing, linearization/reflow, kerning, letter spacing, word spacing, line length, text style, justification, and more — including changes to all text and changes at the element level (e.g., headings different from body text). ...

Web user agents currently do not provide sufficient text display customization functionality, and what is provided has poor usability. Most browsers provide some text customization (font face, font size, color); but are missing other important customization, such as leading (space between lines)[@@ add more]. Additionally, even the most basic text customization is difficult to find in menus, as opposed to being available through default toolbars. Browsers do allow essentially all aspects of text to be customized through user style sheets; however, this currently requires knowledge and skill beyond what most users possess or can be expected to acquire. PDF readers currently do not provide even the most basic text customization that users need.

Some websites are designed in such a way that customizing text display renders some information unavailable. For example, reformatted text disappears because the container does not resize appropriately.

The ability to customize text display is primarily the responsibility of user agents (browsers, etc.). The current draft of UAAG 2.0 includes some requirements for text customization (under "Provide text configuration"); [additional needed?].

Web content also has some responsibility for text customization. It is not clear that WCAG 2.0 includes sufficient requirements for text customization.

...@@ what to do today to compensate for missing functionality in browsers...

...@@ what to do moving forward...

Status of the document

[mostly W3C boilerplate] @@ invitation for public review & comments -- specific questions for consideration


[status: OK for now? after more of the rest of report is done, come back to this section]

Communications and interactions are increasingly provided through the Web and electronic media. This offers the opportunity for unprecedented access to information and interaction for people with print disabilities because the accessibility barriers to print can be more easily overcome through technology. Print disabilities encompass people who cannot effectively read “normal” print because of a visual, physical, perceptual, developmental, cognitive, or learning disability [1].

Much of the effort to date in this area has been on providing access for people who are blind [2].... However, less effort has been invested to meet the needs of other people with print disabilities, including people with low vision and people with dyslexia. <from> @@ Thus there is little research and few resources that provide clear guidance on text customization. Additionally, text customization functionality has not been well integrated in mainstream user agents (web browsers, etc.), nor is it sufficiently included in some accessibility standards and support material (such as the Section 508 standards).

This research report explores the needs of people with low vision, dyslexia, and other conditions and situations that impact reading. It focuses specifically on text customization requirements and functionality, that is, providing users the ability to change (or personalize) various aspects of text display to improve readability for their particular needs. It investigates:

  • Understanding users' text customization needs and requirements - What aspects of text customization improve readability? Which aspects of text customization are necessary requirements for people to be able to read effectively, and which are optional suggestions to improve readability?
  • Integrating text customization functionality and requirements - How well do existing text customization functionality and interfaces support users' needs? How effectively do user agents (web browsers, etc.) and web content share the responsibility for text customization?
  • Moving forward - How might we increase awareness of the need for text customization, and the benefits? What text customization functionality should be included in products in order to meet users' needs? How should text customization requirements be better addressed in accessibility guidelines, web standards, and other best practice guides?

The ultimate goal of this work is to encourage user agent developers, standards developers, policy makers, web designers, and others to provide specific functionality in mainstream web products by helping them better understand and implement text customization.

WAI-RDWG Symposium

[status: section done for now ?]

One reason for lack of sufficient text customization functionality may be a lack of awareness of and research on users' needs, and how they can best be met. WAI-RDWG organised a symposium on 19 November 2012, with the intention of helping to address that gap. This research report is the outcome of that symposium.

The symposium focused specifically on text customization requirements and functionality, that is, providing users the ability to change (or personalize) various aspects of text formatting to improve readability for their particular needs.

Users: The primary focus was on people with disabilities. However, relevant studies in related areas — such as older web users, people with low literacy, and situational issues (e.g., reading in low light) — were in scope if the information is also applicable to people with disabilities.

Tools: Research on assistive technologies and specialized tools for allowing users to customize text was in scope, to inform potential inclusion of text customization functionality in mainstream products.

Technologies: The primary technology focus was on web browsers, media players, and plug-ins such as PDF Reader, Flash, Silverlight, QuickTime. Papers related to other technologies, such as eBook readers, were in scope if the information also is applicable to web technologies.

The following topics were out of scope for the symposium and are therefore not addressed in this report: optimum font and format for text, text legibility, the impact of animated content (e.g., ads) on reading, and reading level.

Related topics were addressed in a companion Easy-to-Read on the Web Symposium <held on December 2012. [@@link to report when it's ready]

Understanding the Problem

[@@ summarize section]

The User Experience

Some people with low vision, dyslexia, and related conditions cannot read normally-formatted text in [books, newspapers, manuals,] web pages, PDF documents, etc. (even with reading glasses) because the text is too small, the colors are too bright or too pale, the letters are hard to distinguish, etc. However many can read text that is formatted differently, for example, with larger text, different font, more spacing, etc. [tader ref?] When some people try to read text that is not well-formatted for them, it causes pain, strain, fatigue, frustration, dizziness, nausea, difficulty seeing, and difficulty understanding, as explained below.

Evidence of users difficulty reading normally-formatted text include:

  • User Experiences – provides @@ [<>]
  • "it is apparent that a number of other people struggle with ClearType and complain of blurring" (Lee 2012) [symposium paper], which lists examples where the issue is discussed
  • @@ others
  • {input requested: additional reports, especially in peer-reviewed literature}

When people cannot read text in websites, one approach is to copy the text into another program where they can more easily customize the text. For example, anecdotes of older adults and others copying text from a browser into a word processing application in order to make changes to make the text easier to read include:

  • "In a 2006 investigation into the usability of online accessibility advice [2], we heard stories from older web users on (sometimes unexpected) approaches to improving readability of online content. In one example, a male participant described how, when he encountered text on a Web page that he found difficult to read, he would copy the text into Word and enlarge it appropriately, since he knew that Word allowed him to customise text appearance."  [<>]
  • "Before I knew how to apply stylesheets to any web page, I used to copy the text from a web page and paste it into Word or other text editor so I could format it to read it." [<>]
  • @@ others {input requested: additional reports, especially in peer-reviewed literature}

Some people use assistive technology such as screen magnification software. However, screen magnification  provides only a small percent of users' needs; for example, it does not [@@leading, colors?]. Some people do not use screen magnification because the functionality does not meet their needs, poor usability, complexity, cost, availability, or other factors. For example, some people do not use screen magnification because while they need to increase text size to read, they do not want to increase images or other screen elements; and some need text to wrap to avoid horizontal scrolling — functionality that most screen magnification software does not provide. "I absolutely hate using the leading adaptive screen enlargement programs like Zoomtext and Magic because they tend to enlarge every image equally and waste a lot of space on the screen. Usually, they do not wrap the information into the visible field and one must scroll back and forth. This is very jumpy and tiresome on the eyes." [<>]

Of particular note is what happens when some people try to read text that is not well-formatted for them.

  • "light backgrounds on webpages hurt my eyes and make reading impossible." ... "completely fatigued" (Rivera Ley 2012)[symposium paper]
  • "This causes visual disturbance and, typically, a bad headache as well." (Lee 2012) [symposium paper]
  • "many authors choose backgrounds that are painful to my eyes" & "I use my style sheet to avoid serious headaches, eye pain and nausea."
    eye strain and pain; neck and back pain; nausea <>
  • "When Text is Not Displayed Well" provides data from over 150 people who experience pain, strain, fatigue, frustration, dizziness, nausea, difficulty seeing, or difficulty understanding. [<>]
  • @@ others {input requested: additional reports, especially in peer-reviewed literature}

Some people report that they quit reading information from the computer (such as an article on the web or instructions in PDF) and not finish it, just because the text is too hard to read comfortably. [<>]

@@ conclusion

User Group

Millions of people cannot read “normal” text, and millions more will not be able to in the coming years as their vision declines due to aging [6-7]. An estimated 15–20% of the population has symptoms of dyslexia [8] and 246 million have low vision (compared to 39 million who are blind) [6].

This report focuses on the largest groups of people with print disabilities: those who can see and can read, but have difficulty reading text in common formats and need to specify different text format in order to read effectively [6-7]. get referenes from <from> It includes: [@@DS: formal definition of user groups who do, or may, benefit from text customisation - this section should define the scope of the note in terms of user groups based on what the literature identifies as those who benefit]

  • people with low vision, including people with declining eyesight due to aging,
  • people with dyslexia and related disabilities,
  • people with other difficulties reading,
  • anyone in difficult situations that impact reading, such as high stress situations, low light conditions, reading on a moving mobile phone, reading a non-native language, and readers with low literacy.

The primary focus is on people who use mainstream technologies and do not regularly use assistive technologies (AT), such as screen magnification.

While the focus of this report is on people who have difficulty reading normally-formatted text, there is evidence that text customization benefits others as well. For example, [<>] reports "One surprising finding in this study: teenagers don't like tiny font sizes... small type often caused problems or provoked negative comments from the teen users in our study. Even though most teens are sufficiently sharp-eyed, they move too quickly and are too easily distracted to attend to small text." and Quesenbery 2012 reports that guidelines for older users, teens, and lower-literacy users "urged a 'larger' default font size".

Readability Beyond Legibility

In order to understand the needs of people with low vision, dyslexia, and related conditions that impact reading, it is important understand the distinction between legibility and readability. Legibility is related to perceiving text by distinguishing letters. Readability is related to reading and comprehending textual information. Thus text could be somewhat legible to a user, yet not functionally readable; that is, with effort the user could distinguish one letter from another, but could not effectively read sentences because of the text formatting. Henry 2012 discusses readability and legibility research and considerations in the context of text customization.

[move to guidelines/stds gap section] [The authors assume] that many of the accessibility guidelines for electronic media are also focused on small amounts of text, such as website navigation, forms, and short descriptions. Thus the guidelines may be sufficient for legibility, but not for readability. ... To read large amounts of text, users need to be able to customize more aspects of text formatting.

Footnote 1: In his seminal research, Tinker [4] used only the term legibility to avoid confusion with readability formulas for the level of difficulty of the language; however, most literature distinguishes between legibility and readability as used in this paper.<from>

Lack of Awareness

[@@ make more clear two points: 1. lack of awareness of designers, ax experts, etc. 2. users not aware of what could be (and thus not advocating]

[Additionally, some people seems not to realize the impact of text display on many users (might consider text customization optional).]

Much of the effort to date in providing access to electronic text has been on providing access for people who are blind [2 <from], with However, less effort has been invested to meet the needs of other people with print disabilities, including people with low vision and people with dyslexia.

Now there are cases where information is accessible to people who are blind and use screen readers, yet not accessible to people who can see.

  • "my totally blind husband is able to successfully use his screen reader on web pages that are completely unreadable by me" (Rivera Ley 2012)
  • "Just because it is accessible through voice output does not mean it is accessible for people who are not blind but have serious visual impairments." <>

@@ "accessible PDF" is not —"with the technology currently available, PDF is not sufficiently accessible to many people with low vision, dyslexia, and related conditions that impact reading" [ref:]

[There appears to be a general lack of awareness of the importance of text customization.] ["this problem stems from a lack of understanding of the needs of low vision users" (Rivera Ley 2012)][For example, the Product Manager for Accessibility at a major software developer said that he has only heard of the need for text customization from two people [personal communication, March 2012].] ...

Possible reasons for the lack of awareness include:

  • Lack of a well-organized advocacy efforts (for example, that exist for users who are blind) ["insufficient self-advocacy by users with low vision who lack the resources and forum to express their web text customization needs" (Rivera Ley 2012)]
  • Lack of awareness among users of the potential for text customization in user agents ["After interviewing professionals with low vision, I found that many do not use the current workarounds like the Zoom slider bar in Internet Explorer" (Rivera Ley 2012)]
  • Lack of clear requirements in WCAG 2.0 at Level AA or A
  • Lack of awareness of UAAG
  • ...

Awareness of text customization functionality amongst users who could benefit from it has been a persistent challenge over many years. Nielsen (2002) pointed out that text resizing had become more difficult to implement in Internet Explorer version 4, compared to version 3.

need some commentary on changes since then...?

If customization functionality is not made obvious to browser users, then awareness of functionality is unlikely to be high: ["After interviewing professionals with low vision, I found that many do not use the current workarounds like the Zoom slider bar in Internet Explorer" (Rivera Ley 2012)]

Anecdotes of older adults copying text from a browser into a word processing application in order to make changes to make the text easier to read could indicate a lack of awareness amongst users of a Web browser's facility to customise the appearance of content, compared to the familiar options to customise text appearance offered by an application intended for text authoring. [ref Sloan's email comment]

An addition to lack of awareness, attitude to customization as an accessibility solution can be a challenge. Sayago et al (2011) presented evidence to indicate that older adults, in particular, may be more willing to accept increasing difficulty in reading text, or use physical adjustments, like putting on reading glasses or moving closer to the screen, than to implement an accessibility change even when made aware of it. Sayago et al suggest this could illustrate an issue of social acceptability of accessibility solutions - reading glasses being more socially acceptable than enlarging text on a computer screen.

However, as mobile devices with smaller screens require greater numbers of people to make changes to text display in order to make it easier to read. So, just as the increased motivation of content authors to deal with the constraints of mobile devices could also lead to accessibility improvements [ref WAI Mobile Accessibility Overlap], so the this normalization of customization may lead to greater acceptance, expectation and adoption by users of text resizing functionality across different platforms.

[IRC log comments that mobile helping to bringing awareness...] [Eileen's comments in the symposium on using Youtube to encourage sharing of personal experiences]

Refs to add for this section:

Sayago, S., Sloan, D. and Blat, J. (2011) Everyday use of computer-mediated communication tools and its evolution over time: An ethnographical study with older people. Interacting with Computers 23 (2011). Elsevier: 543–554.

Nielsen, J. (2002) Let Users Control Font Size. Alertbox: August 19, 2002.

Sloan, D. (2012) Email comment to WAI-RDWG public...

The Gap Between User Needs and Current Support

There is no single optimal text display that will work well for all users, so some people need to be able to customize text display to suit their own needs. Evidence to support understanding of the exact nature of customization, and on the impact on users of being unable to make display changes, is relatively scarce. However, the evidence that does exist indicates a substantial diversity in the type of customization that is needed by users.

There is also evidence that users experience difficulties in being able to make necessary display changes. This may be due to a web page’s design, or that a specific customization is not supported by a user agent or operating system.

This section summarizes current knowledge of the nature of text customisation needs, the methods that are available to support customisation, and the current limitations of these methods.

User Text Customization Needs

User needs for text customization are diverse across different users: "Text customization is not a one size fits all proposition. Text customized for one person may be unreadable for another. Even for an individual, optimal customization may vary as vision fluctuates. Web guidelines and practices must provide highly flexible solutions that meet the "one size fits one" requirements of users with low vision." (Rivera Ley 2012)

An individual’s text customization needs may also fluctuate, as visual capabilities change over time (Rivera Ley 2012). Needs may also change depending on the type of reading activity. For example, Dick [reference to add] reports using different text size settings for composing text and for reading text.

So, the aspects of text display that can affect readability, and therefore which need customization, are wide-ranging. Henry 2012a surveyed users who need to be able to customize specific aspects of text display, and also inspected examples of user-defined style sheets. These aspects are presented in Table xxx.

[Table of text customisation properties reported, and CSS properties adjusted]

Changing font from the default fonts specified by web content authors may help improve readability for some users. The relative merits of using sans-serif over serif fonts to aid on-screen readability have been documented over many years [@@ reference to current thinking…] Additionally, a number of efforts have been made to develop fonts intended to make reading easier for people with dyslexia, including [@@ dyslexia fonts – and level of evidence to support their use?].

Lee (2012) describes problems with reading text presented using sub-pixel rendered fonts, and the need for customization support to allow changing of font to one that does not use sub-pixel rendering.

Another aspect of customisation to consider is control over what is customised on a page.

[@@need to word this carefully. Something to do with id or class level customisation] Dick(2012) argues that text customization control is needed at a finer-grained level than this.] argument for customisation at individual element level - where different instances of the same HTML element in the same document may be styled differently, in order to give distinct semantic meaning. [@@ to reword...Applying customisation to that element would then remove this visual 0- and hence semantic - distinction 
Silas Brown - there are limitations in control over specificity of user CSS (relating to Wayne Dick's observation above), of impact on script-generated content, on content not subject to CSS (e.g. title attribute. Argument that this *is* web content, rather than extension of the user agent UI)]

Henry [@@ ref] found evidence that some people who need to make text customisation will inspect a site's style sheet, identifying element id and class attributes, in order to make specific changes to display. This is understandably a time-consuming process that must be applied on a site-by-site basis, so may be adopted only for sites that are used regularly.

Text customization needs also extend to when content is printed out. Customization applied to an online version of text may still be required when that content is printed. For example, the ability to print content is important for some people with dyslexia [27 – ref to find].

There is also an issue of customisation as a user preference versus customisation as a diagnosed improvement to reading. In a study, Rello and Baeza-Yates (2012) reviewed the reading performance of people with dyslexia when reading text presented in a variety of colour combinations, and also asked participants to rate their preference for the different colour combinations. They found no correlation between reading performance and preference. Dick (2012) has argued for the need for a more empirical approach to diagnosing an individual's reading needs in terms of text customisation, and making adjustments accordingly.

Quality of Support for Text Customization

Given the diverse text customization needs that exist, it is important to establish the current level of support for these needs in user agents and at operating system level. This includes considering:

  • the extent to which it is possible to apply a specific customization;
  • the effectiveness of applying a customization on the text;
  • the usability of the process of applying a customization;

Extent of Customization Support

The wide range of text customization requirements identified as being necessary for some users is not fully supported in all user agents.

Operating Systems

At operating system level, a range of system-wide customization features are offered. These should be respected by user agents and inherited in the display of web content. Text customisation features include:

  • Text size changes.
  • Text and background color changes, for example via Windows High Contrast Mode.
  •  ?
Web Browsers

Most mainstream web browsers provide functionality for users to customize font face, text size, text color and background color. Most provide overall page zoom that causes horizontal scrolling, and many provide text-only zoom that rewraps to avoid horizontal scrolling in many cases.

Most browsers provide functionality for users to set their own style sheets and customize all aspects of text display - leading/line spacing (line-height), letter spacing, word spacing, width, text style, justification, and more. However, this is an advanced feature - currently, creating and using user style sheets requires knowledge of CSS (cascading style sheets) and browser functionality that is beyond most users' knowledge.

Chrome is withdrawing support for user style sheets [@@more details, including reason]

PDF Viewers

In contrast to functionality provided through web browsers, several aspects of text customisation are not provided by PDF viewers.

The most commonly used user agent to view PDF content is Adobe Reader. It does provide some customization functionality, including text and background color customisation, zoom, and reflow that temporarily puts text in a single column. However, there are limitations to the latter two,. Additionally, PDF documents cannot be printed when zoomed or reflowed.

Adobe says of Reader’s reflow limitations:

"Text that does not reflow includes forms, comments, digital signature fields, and page artifacts, such as page numbers, headers, and footers. Pages that contain both readable text and form or digital signature fields do not reflow." [28]

In reflow mode, the search/find-in-document feature does not work at all.

Documents with some layouts are not functionally readable to some users when zoomed, such as research papers formatted in two columns. When users get to the bottom of a column, they have to scroll up to find the top of the next column and the physical and cognitive effort required can break the flow of reading and understanding substantially.

Adobe Reader does not provide functionality for users to specify font face, text size for specific elements, leading/line spacing, and most other aspects of text formatting.

[@@ add Support in other PDF readers.]

  • VIP PDF-Reader (Swiss National Federation of and for the Blind) provides a substantial set of text-customiztion features.

Supplementary Approaches

Limitations in text customization support in user agents are marked by the existence of supplementary approaches that users may need to take in order to successfully apply text customisation. IN some cases, these tools are proofs-of-concept, results of academic research Tools to aid text customization exist in several formats. Browser extensions can be added to a browser toolbar to provide enhanced support for customisation, for example

  • Readability
  • Stylish
  • Techdis toolbar
  •  ?

Dick has worked on developing tools to aid the process of creating optimised user style sheets. ZoneClipper is a Firefox extension that simplifies web pages so that element level style can be applied (Dick & Mhatre, 2010). Subsequently, Typometric Rx has been developed to enable users to create element level user style sheets (Dick, 2012).

Server-side transcoding of content to apply customization to HTML content has also been investigated. This involves a server-based approach, where the system scrapes the HTML of a web page chosen by the user for customization, applies a custom style sheet to change aspects of text appearance and may make additional changes to display such as linearizing column-based layouts or replacing images with text alternatives.

  • An early attempt was BETSIE, from the BBC, which redisplayed web pages
  • IBM developed a transcoding tool…
  • Silas Brown
  • Topac (2012) describes development of Text4all, a transcoding tool to aid text customization of HTML content.

Drümmer (2012) describes an approach to applying customization of PDF content

Some web authors provide widgets for customization, embedded within web content, for example text resizing options or style sheets with alternative color schemes.

Effectiveness of customization

Some customization functionality is limited in its effectiveness. For example, customization may not be applied to all text successfully, or may not be applied to printed versions of content. Some of this limitation can be explained to web content that does not meet relevant WCAG success criteria, specifically those relating to Guideline 1.3 Adaptable. For example:

  • list of relevant WCAG SCs

In other situations, the OS may have limitations in terms of the quality of native customisation support. For example, in iOS (version 7), customisation is offered in the form of inverse colour schemes, which changes text and background colors. However, it also inverts colors of video content, making video effectively unwatchable.

Limitations also exist in browser extensions, standalone tools, transcoding services and text customisation widgets. Limitations of transcoding tools include an inability to deal effectively with changes to the DOM, inability to reach secure content. Additionally, users need to be aware of such solutions in order to be able to take advantage of them. For text customization widgets embedded in web content, the main advantage is that these options are highly visible within the content of a web page that requires customization. However, they are only effective on the single site on which they are implemented, so cannot be relied upon by users who benefit from the changes they apply.

Usability limitations

Some customization functionality is supported, but there are usability challenges relating to the complexity of the method of customization, which mean that some users will find it unduly difficult or even impossible to make their desired customisation changes. For example, it can take 20 steps to change font color in a common browser [@@elaborate].

[@@Suzette Keith - line spacing control as an example of customisation that is currently difficult to implement] [@@Anthony Lee - personal experience of difficulty with sub-pixel rendering of fonts, and difficulty in making necessary customisation]

There is anecdotal evidence of people adopting workarounds to implement text customisation needs, indicating either a lack of awareness of the ability to do this at browser level, or undue difficulty in using the browser for customisation. For example, there is anecdotal evidence of users coping text from web user agents and pasting into other programs in order to customize it:

  • 'Reading some web text requires me to copy it into Word, increase font size, adjust colors, and increase margins to insure that text reflows properly. These text gymnastics are tedious and time consuming." (Rivera Ley 2012)
  • "a participant described how, when he encountered text on a Web page that he found difficult to read, he would copy the text into Word and enlarge it appropriately, since he knew that Word allowed him to customise text appearance." (Sloan 2012)
  • "Before I knew how to apply stylesheets to any webpage, I used to copy the text from a web page and paste it into Word or other text editor so I could format it to read it." (Henry ??<>)

Gaps in Standards and Guidelines

[status: this sections needs lots of work]

[@@SLH - I think we want a whole section on how to help address this — e.g., Does WCAG provide sufficient guideance for supporting text customzation? Are additinoal techniques needed? Would it be useful to have a Best Practices for Supporting Text Customization application note? Is text customization sufficiently covered in UAAG 2? How to we raised awareness of UAAG and encourage user agents to follow it? etc. :-] [@@DS Yes, agreed - pulling out those specific guidelines from UAAG underlined just how important it is that we produce something of practical use to the UAAG working group]

[@@ Current situation is that 1.4.8 covers visual presentation, including provision of a mechanism to enable aspects of customisation. But is allowing a user stylesheet to override presentation sufficient?]

[some relevant WCAG success criteria]

UAAG 1.0 @@

UAAG 2.0 Draft [at time of Symposium, little. As of this writing, more...] from previous draft:

[Text customisation in Access For All 3.0 and ISO 24751.]

Supporting Text Customisation

Supporting Text Customisation Ideally

While user agents have a key role in providing text customization, web content authors also have a role. For example, by not putting important text in images making the text unchangeable - "Such text is therefore off limits to me because once it has been embedded/subset, it cannot be changed (other than zoomed) and is therefore blurred." (Lee 2012)

The ideal model of supporting Web accessibility in general, as articulated in the WAI document [Essential Components of Web Accessibility], is that:

  1. web content is created with accessibility in mind;
  2. user agents support accessibility features that allow a user to receive content in a format appropriate to their specific needs;
  3. users with specific accessibility needs are aware of those needs and have appropriately configured their user agent (and assisitive technology if necessary) to accommodate their needs.

For text customisation, this means that:

  1. web content is designed in a way that allows the presentation of textual content to be changed from the default presentation to meet a user's personal reading needs. This includes meeting WCAG success criteria relating to:
  2. user agents support text customisation through [@@DS - referencing draft UAAG 2.0 here]:
  3. web users who have difficulty reading textual content understand what customisation needs adequately address their readability needs and have the means to apply this customisation.

In reality, though, sufficient accessibility support is often not provided:

  1. web content has accessibility issues that prevent effective customisation;
  2. user agents do not adequately support customisation in a way that addresses all readability problems;
  3. many web users who have difficulty reading textual content are not be aware of specific customisation solutions that would improve readability.

Longer term, this situation could be resolved by:

  1. greater awareness and implementation of accessible web authoring practices by web authors and those commissioning web content;
  2. user agents with more extensive and prominent support for text customisation;
  3. education of web users on the role that text customisation can play in improving the web browsing experience.

[@@SLH - I think we want a whole section on how to help address this — e.g., Does WCAG provide sufficient guideance for supporting text customzation? Are additinoal techniques needed? Would it be useful to have a Best Practices for Supporting Text Customization application note? Is text customization sufficiently covered in UAAG 2? How to we raised awareness of UAAG and encourage user agents to follow it? etc. :-] [@@DS Yes, agreed - pulling out those specific guidelines from UAAG underlined just how important it is that we produce something of practical use to the UAAG working group]</p>

Improving the User Experience of Customizing Text

[previous heading: Improving the user experience of applying text customisation]
[I think the also applies to Today - maybe bring it up a level somehow?]

There are a number of enhancements that could be made to the process of of selecting and applying appropriate customisation. These coould include improving the usability of the user interface of text customisation tools (e.g., showing the font options in the font itself; showing the colour choices in the colours themselves; making sure text customization is accessible, e.g., text display widget is keyboard accessible; how to help users know what helps them - do this test and heres a set of display options that will help you best; providing minimal simple options, e.g., 3 color combinations choices, PLUS advanced functionality for multiple options, i.e., picking each colour from full palette)

[some suggestions for browsers & content developer <>]

Supporting Users in Customizing Text for their Needs

[previous heading: Supporting users in choosing appropriate customisation]
[doesn't see to flow well here. maybe it will fit better under Moving Forward?]

Most tools and research on text customization has left users to set text display with little guidance. [refs like Hanson & Richards?] Dick 2012 proposes a more formal approach for determining optimum text display for users. ...more here...

Of note is that users' preferences might not always correlate to the best settings as determined by other diagnosis. In a study of text customisation for people with dyslexia, Rello and Baeza-Yates 2012 found no positive correlation between participants' color preferences and reading performance as measured by eye-tracking recording fixation duration. [@@ "Text should be printed with the highest possible contrast. There is good evidence that for many readers who are older or partially sighted, light (white or light yellow) letters on a dark (black) background are more readable than dark letters on a light background. However, the traditional dark on light may be aesthetically preferable." (Arditi, Aries. (1999) Making Text Legible - Designing for People with Partial Sight ]

Automated Text Customisation

[doesn't see to flow well here. maybe it will fit better under Moving Forward?]

Beyond this, text customisation could become automated to at least some extent. Access For All 3.0...and GPII...[tighten up with refs] describe a future scenario whereby an individual's accessibility needs - including text customisation needs - are stored in a portable profile, that can be applied to a digital resource which adapts itself accordingly in line with the needs expressed in the profile.

However, in the short-term, web content authors who wish to ensure that their content is as accessible as possible to the widest number of their target audience are faced with the challenge of deciding whether or not to attempt to bridge this gap in text customisation awareness and support, and, if so, what to do.

Supporting Text Customization Today

[@@DS - short piece on paradox of customisation widgets not being visible to those who need them most?]

Smith (2010) WebAIM - Web Accessibility preferences are for sissies argues that web authors should focus on authoring accessible content, and that attempting to go beyond this to provide customisation features in web content can lead to reduced usability.

Others have argued that the limitation of this approach is that for people who could benefit from customisation, but are not aware of their own needs, the visibility of customisation as a possible solution is lost. Sloan et al (2006) Evaluating the Usability of Online Accessibility Information describe a study where a group of older web users were asked to find accessibility information provided on a sample of sites from organisations in the Financial, Governmental, Education, Commerce and Disability Advocacy sectors; plus a set of sites chosen as exemplars of good accessibility practice. The researchers found that there was a low awareness amongst participants of text customisation, and that participants perceived accessibility information provided by sites as frequently hard to find, excessively technical and unhelpful.

Some authors and organisations have focused on providing educational resources intended to raise awareness of text customisation amongst users who could benefit from customisation:

[@@ fyi, some notes on this.]


Moving forward

[@@ review note - we have a lot to say here that's not here yet :-]

  • How might we increase awareness of the need for text customization, and the benefits?
  • What text customization functionality should be included in products in order to meet users' needs?
  • How can we improve discoverability and usability of text customization features in products?
  • How should text customization requirements be better addressed in accessibility guidelines, web standards, and other best practice guides?
  • What areas of research show promise to inform and evolve text customization for readability?

[@@DS - something on the impact on text customisation needs of changing monitor technology - resolutions, electronic paper etc]


  • Translating knowledge of users' customisation needs into policies, standards, user agents. Submissions did not cover this issue in any great detail, but if there are gaps in support that could be addressed by revising guidelines, standards, policies covering content authoring and user agents, we should identify them here
  • this work could also take into account evidence supporting the need to rebalance efforts between user agent support for and usability of customisation features, and user effort in creating and maintaining user CSS.
  • Supporting discoverability of customisation vs education on the existence of customisation as a user option (and helping users diagnose their needs)
  • Specific research projects that are ongoing:
    • research presented in symposium submissions
    • other research currently underway that could inform knowledge in this area
  • Research that has not been done but needs to be done




[status: after more of the rest of report is done, then work on this section]


Many user agents (i.e., the tools that people use to interact with electronic information, such as web browsers) do not sufficiently meet the needs of these users. They do not provide adequate text customisation that is necessary for some people to be able to read, understand, and interact with information.

While electronic media provides an opportunity to increase readability for large groups of people who have difficulty reading and processing text, this opportunity remains unrealized in several areas. Additional research, guidance, and education is needed to encourage user agent developers to provide the text customisation functionality required by people with conditions that impact reading. <from>




Symposium Proceedings

proceedings with BiBTex (example format)


Participants of the W3C WAI Research and Development Working Group (RDWG) involved in the development of this document include: Simon Harper (Co-Chair), Shadi Abou-Zahra (W3C WAI Staff Contact), @@....

RDWG would also like to thank:

  • Chairs and Scientific Committee members
    • Shawn Henry (W3C/WAI), Co-Chair
    • David Sloan (University of Dundee), Co-Chair
    • Shadi Abou-Zahra (W3C/WAI)
    • Vivienne L. Conway (Edith Cowan University)
    • Robyn Hunt (AccEase)
    • Caroline Jarrett (Design to Read and Effortmark Limited)
    • Clayton Lewis (University of Colorado)
    • Kerstin Matausch (KI-I)
    • Klaus Miesenberger (Johannes Kepler Universität)
    • Christopher D. Nicholas (Language Technologies, Inc.)
    • Birgit Peböck (KI-I)
    • Luz Rello (Universitat Pompeu Fabra)
    • John Richards (IBM T.J. Watson Research Center and University of Dundee)
  • Paper authors
    • Ricardo Baeza-Yates, Yahoo! Research & Universitat Pompeu Fabra
    • Wayne E. Dick, Ph.D. Knowbility, Inc.
    • Olaf Drümmer, callas software GmbH
    • Shawn Lawton Henry, W3C Web Accessibility Initiative (WAI)
    • Luz Rello, Universitat Pompeu Fabra
    • Eileen Rivera Ley, MBA. Ley & Associates, LLC
    • Anthony Lee
    • Vasile Topac, Politehnica University of Timisoara
  • Additional contributors via e-mail:

This document was developed with support from the WAI-ACT Project.

NOTE: This is a rough, in-progress draft.

OLD draft info not included in above

Outline Structure from Korea

  1. Abstract
  2. Status of Document
  3. Summary (?) (moved to Abstract)
  4. Introduction
    1. Readability beyond Legibility
    2. About the RDWG Symposium
  5. Understanding The Problem
    1. User Stories (scenarios)
  6. The gap between needs and current support
    1. The details of what users needs (requirements)
    2. The details of what user agents offer
    3. Standards and guidelines

Existing Research Overview

The needs of people who have disabilities and conditions that impact reading have been addressed in a wide range of research studies. Most of the research focuses on one user group, such as older users, and one domain, such as websites.

Most research on making text more readable for people with low vision is designed to determine optimum characteristics such as font face and size [9]. There is similar on improving readability of text for people with dyslexia [10]. Many of the results suggest characteristics for readability in print, and more recent literature addresses electronic media. There are some recommendations for online readability, both for older technology and for newer technology such as is available in e-book readers.

There is less research on what users should be able to customize in order to optimize readability for their particular impairment and situation. Work in this area tends focus on a specific user group and situation, such as older users who are new to the Web or adult students with dyslexia. Specialized software has been developed for such users, for example, [11-14]. Yet most of this customisation has not been well integrated in mainstream user agents, nor is it sufficiently included in some accessibility standards and support material (such as the Section 508 standards [15]).

Additional research is needed to understand, document, and communicate the needs of users in order to encourage inclusion of additional text customisation functionality in mainstream user agents. [16] says, “…end user customization plays a central role in accessibility considering dyslexia. Nevertheless, only two guidelines where found regarding this subject. Thus, a deeper study on end user customization is an identified gap that needs to be bridged.” <from>

Background... Specialized Text Customisation Tools

[?DS work on this section?] [include ATbar ]

Over recent years, several tools have been developed with the aim of supporting individuals in customising the presentation of web content to suit personal reading needs. These tools take various forms. In some cases, visibility of browser functionality relating to text customisation has been enhanced, for example Google Chrome's text resizing feature present in the browser address bar. In other cases, browser functionality has been extended through browser add-ons or extensions, or via widgets present in web content itself.

[@@referencing to be fixed] Browser add-ons that support specific text customisation capability include AT Bar [1], Readability, Stylish, the NoSquint Firefox extension:, the Theme Font & Size changer for Firefox:, Readable for Firefox:

Widgets provided in web content...


What do we know about the usability of these customisation tools (in terms of how easy it is for users to use the tools to apply desired customisation)?

[@@DS - need to discuss most effective presentation of these tools - text, table, etc. Classification according to technical approach e.g. transcoding a web page; browser add on to apply specific styling; widget added to page content][@@SLH list seems fine. don't think it needs the complexity of a table. [@@SLH update: although RDWG didn't like it as a list. What are the goals of this section? Just to show prior work? Or to say more specifically what's been learned from these tools? the latter would go in more of a sentence style discussion. Maybe don't need table or formatted list, but just inline list with references? or perhaps formatted list in an appendix or footnotes or something?]


  • IBM Web Adaptiation Technology (Richards and Hanson 2004; Hanson et al 2005)
  • PAN (Iaccarino et al 2006)

Adaptive interfaces

Some TC4R symposium authors described the development of customisation tools under development. [Olaf Drummer, symposium paper 6] describes exploration of text customisation for PDF... [Vasile Topac, symposium paper 8] describes text4all, a server-side transcoding tool, but at the time of writing no evaluations of the tool had been evaluated.

Internal Notes

Previous drafts:

note from Korea: put up front - the current situation for users

For RDWG review & comment on 19 Dec:

  • NOTE: This is a very rough, in-progress draft. Some sections are not drafted, there is some redundancy, we expect things to move around some, we will pull additional input from the transcript, we haven't worked on transitions yet, etc. (Do not bother with wordsmithing-level comments.) Instead, you might want to focus on the outline (contents) and summary for now.
  • General structure of the document?
  • Any topics missing?
  • Any points that people took away from the Symposium that are not covered?
  • Ideas for making it a useful report?

Goals & Key Points

  • Bring light on the issue - raise awareness
  • Make clear that text customization is not just nice to have, but REQUIREMENT for many people
  • Encourage more work in this area
  • ...

Style Guide

  • [@@open] Research Report (capitalized) or research report (lower case) -- as in "This Research Report addresses..."

Reference formats

[@@open] e-mail format:
Doe, J,, 2012. [TC4R Symposium] Subject. [E-mail] Message to RDWG Comments ( Sent 00 November 2012. Available at: [Accessed 12 December 12].



Other Symposium Reports:

[More TC4R Report Notes and Archive of previous draft info]