2.1. Legal/regulatory

The United Nations Convention on the Rights of Persons with Disabilities was adopted on 13 December 2006 and then opened for signature by States Parties (or governments) on 30 March 2007.³ The convention is an instrument that legally binds governments either by “signature” (“intention to take steps to be bound by the treaty at a later date”) or by “ratification” (“intention to undertake legal rights and obligations”). There are now 147 signatories to the convention – see map of signatories and ratifiers (Fig. 1).

Fig. 1

World map showing the signatories to and ratifiers of the United Nations Convention on the Rights of Persons with Disabilities, as at 2 October 2010⁴

Specifically, Article 9 of the Convention,³ entitled “Accessibility”, includes the following clause: “To promote the design, development, production and distribution of accessible information and communications technologies and systems at an early stage, so that these technologies and systems become accessible at minimum cost.” Article 21, “Freedom of expression and opinion, and access to information”, details measures that governments shall take themselves to “urge” private entities and to “encourage” the mass media.

The Convention establishes that access to information is a human right for persons with disabilities, who are identified as: “those who have long-term physical, mental, intellectual or sensory impairments which in interaction with various barriers may hinder their full and effective participation in society on an equal basis with others.”

Countries ratifying the Convention undertake to introduce national legislation to implement the human rights of persons with disabilities and avoid discrimination on the basis of disability. For Anglophone audiences the following examples may be of interest: Australia (Disability Discrimination Act, 1992), United Kingdom (Disability Discrimination Act 1995, 2005), United States of America (Section 508 of Rehabilitation Act, 1998).

One of the aims of the Australian Disability Discrimination Act 1992⁵ is to “promote recognition and acceptance within the community of the principle that persons with disabilities have the same fundamental rights as the rest of the community”. This act was the basis of one of the first court cases relating to information accessibility, Bruce Lindsay Maguire v Sydney Organising Committee for the Olympic Games,⁶ where the judicial body was required to rule on information accessibility in respect of web sites. The case centred around the web site of the Sydney 2000 Olympic Games and the inability of people with visual impairments to efficiently utilize the web site in comparison to an able-bodied person.

In the United Kingdom, the Equality Act 2010⁷ replaces major parts of the Disability Discrimination Act 1995, 2005.⁸ From 1 October 1999, the Disability Discrimination Act has required that service providers make “reasonable adjustments” for disabled people, such as providing them with extra help or making changes to the way they provide their services. Provision of information falls within the scope of this requirement and provision of accessible information can be considered one form of reasonable adjustment.

In the United States, Section 508 of the Rehabilitation Act as amended in 1998⁹ requires Federal agencies to make their electronic and information technology accessible to people with disabilities. Section 508 of the Act applies to all Federal agencies when they develop, procure, maintain or use electronic and information technology. Under Section 508 (29 U.S.C. ‘ 794d), agencies must give disabled employees and members of the public access to information that is comparable to the access available to others.

2.2. Commercial

The World Health Organization estimates that 10% of the world’s population – approximately 650 million people, of which 200 million are children – experience some form of disability.¹⁰ There are millions of people with congenital or acquired visual impairment or other communication impairments, in addition to which almost all countries are experiencing an ageing of their population. Therefore, more people will be experiencing refraction error and visual problems. For example, the Australian Human Rights Commission estimates that businesses not offering access are missing out on 20% of consumers.^{11 , 12} It should also be remembered that over 774 million people in the world have limited literacy skills and will require information in other formats.¹³

This represents a significant audience that not only has a right to access information, but that is also a previously untapped market offering a significant commercial opportunity. Clearly, market analysis is required to establish the size and needs of the market in particular locations, but publishers are already anticipating the ageing of the baby-boomers with failing eye sight who appreciate eye-catching books with larger fonts.¹⁴

The Vision IP initiative of the World Intellectual Property Organization (WIPO) is seeking to facilitate access to information for visually impaired persons based on limitations and exceptions in national copyright laws for the blind and visually impaired.¹⁵ This is a quote from the Vision IP web site¹⁶: “Frequently quoted figures show that less than 5 percent of books become available in formats accessible to [visually impaired persons] (such as Braille, large print or audio-formats) within the first year after publication. Likewise, in the field of higher education 47 percent of blind and partially-sighted students cannot obtain textbooks in preferred formats, and 33 percent of visually impaired children have problems accessing school books in accessible formats.”

Thinking about accessibility is the right thing to do. Again, we can learn from the experience of the web industry.¹⁷

2.3. Convenience

Making publications accessible makes sense not only in terms of promoting human rights and benefiting the commercial bottom line. It also makes life easier for everyone. An accessible publication is often easier for people with unimpaired vision to use as well. If a document is clearly laid out and logically structured, it makes for better and faster comprehension. If a document is available in audio formats, it can also be useful to those who prefer to access information on the move or while they are occupied with other tasks.

3.1. File formats

Disability can range from low vision to blindness or it can involve limited mobility or cognitive disabilities such as dyslexia. Vision-related conditions include age-related macular degeneration (no vision in the centre), glaucoma or tunnel vision (peripheral vision damaged), diabetic retinopathy (patchy vision) and cataracts (blurry vision). Readers with these conditions may use their remaining vision and/or assistive devices, such as magnifiers, hand scanners, screen readers, Braille and audio players, or a combination of these, to read different types of information.

There are several different formats available for producing accessible publications. They range from the more traditional Braille,¹⁸ large print¹⁹ and “talking books” to more recent formats such as accessible PDF^{20 , 21} and DAISY.²² These formats are optimized for different types of media, applications and content. People have different requirements and different preferences, and therefore they may require and/or prefer different file formats. Rather than considering the particular format that is best particular people, an alternative is to provide a choice of formats.

This paper focuses on using JATS to create accessible journals and books in print, as well as accessible PDF and DAISY formats.

3.2. Production

A well-structured publication, with clear relationships defining the logical reading order and the navigation through the content, is more accessible to a broader audience than a poorly written and edited publication. Coherent and consistent writing and editing improve accessibility.

Visual design has a significant impact on the accessibility of printed content. When designing for accessibility, one should rely on best practice design principles to create simple and easy-to-follow page designs.

The following is a summary of the “10 golden rules of clear print” as presented by the British Royal National Institute of Blind People²³:

1. use 12-point font or ideally 14-point;

2. use clear typeface;

3. avoid italics and underlining (bold is OK);

4. avoid blocks of caps (indent or use a box);

5. left align text (for right-to-left text) with images to the right;

6. keep text horizontal;

7. aim for good contrast between text and background; avoid combination of red and green, or blue and yellow; and if you reverse white type on a coloured background make it bigger;

8. do not set text over images;

9. use a consistent and logical layout (with navigational aids, clear levels of headings, page numbers);

10. print on matt or matt coated paper (not glossy) thick enough that light does not shine through (more than 90 gsm).

Alternatively, a dedicated format can be produced, namely “large print”. The British Royal National Institute of Blind People defines large print as being 16-point type and above.¹⁹

It is interesting to note that ePub readers, which enable the reader to adjust the size of the font, may increase the range of accessible publications available to some persons with disabilities.

Whatever the format, the British Royal National Institute of Blind People stipulates that no information should be conveyed solely through the use of images, diagrams or colour. Again, this is good practice in general, as is writing comprehensive captions and headings for images and diagrams. A further way to satisfy this requirement is by using “alternative text” – or alt text – text that conveys the same information as the image.

4.1. Change in workflow

During 2010, the Bulletin of the World Health Organization (peer-reviewed public health journal) changed its workflow from generating XML at the back of the process to a single source XML-based workflow as summarized in Fig. 2. Journal content is authored, edited and translated (into 5 languages) using a word processor (Microsoft Word). The editors prepare the structure for conversion into XML (using Inera’s eXtyles). Artwork is prepared in separate files. XML and artwork files are then used (run through Typefi Publish) to create the page layout in InDesign. Corrections from author, 5 translators and proofreader are then collated by an editor and the process repeated. A designer reviews the InDesign files before going to press. The final XML is then ready to be used to produce multiple formats through a series of XML transforms.

Fig. 2

Flowchart of the single source XML-based workflow adopted at the World Health Organization

The driver for this change was the opportunity to use the Bulletin’s XML to create page layout in InDesign. Therefore, the change was planned in two steps: April 2010 for the migration of print version and html, and January 2011 for the addition of accessible PDF, DAISY and ePub. The same workflow has been implemented for the production of two books (using the book DTD): the WHO/World Bank World report on disability and WHO’s The world health report - health systems financing: the path to universal coverage.

4.2. Discussion about approach and extent

An alternative to creating accessible publications as derivative products from an XML-based workflow is to create accessible PDFs at the end of the production process. This can be done by marking up the PDF used for printing (levels of heading, order of reading, etc.). We did not choose this workflow as the markup would not have benefited from the multiple checks (editor, proofreader, designer) from which the XML-based workflow gains strength. Furthermore, it is probably not what the Convention intended; it specifies “... at an early stage, so that these technologies and systems become accessible at minimum cost”.

The large print format increases page count and therefore paper, printing and distribution costs. It therefore cannot be financially justified for general production and may need to remain a derivative product or print-on-demand option.

However, the Clear print guidelines ²³ have provided us with some interesting opportunities. The World Health Organization has adapted these guidelines to its particular context and is looking to develop supplementary clear print guidance for other official languages, particularly those using non-roman characters (i.e. Arabic, Chinese and Russian) and those that read from right to left (i.e. Arabic).

An evaluation of the Bulletin with the Clear print guidelines was positive, although improvements could be made by increasing font size, using bold type for emphasis rather than italics and including alternative text for images. The 3-column page layout does not lend itself to 12 point font and we have not changed the scholarly editing rules of using italics for words in foreign languages, book titles, journal names and conference names; however, we will be implementing accessible PDF, DAISY and ePub in January 2011. Although the solution does not make the print version accessible, the other formats offer options to persons with disabilities. A similar strategy has been taken for The world health report - health systems financing: the path to universal coverage.

Given the priority of making the print version of the World report on disability accessible, the additional step to make the font larger was taken, as well as the decision to write alternative text for figures and tables. Although the Clear print guidelines do not mention alternative text for tables, it is more interesting than to listen to a list of data values because it helps the reader to contextualize the information – and alt text for tables is in addition to the automated reading of data values. Alt text for tables is also useful when viewing tables on small devices with screen limitations, for example, ePub files with tables that do not fit legibly on the screen, or tables presented as images.

We have had some debate about writing alternative text for scholarly publications and solicited feedback from our network of nongovernmental organizations. Following guidance from the web industry it was clear that alternative text should describe the image but focus on conveying the meaning (see example in Annex A). When writing for graphs with multiple lines it becomes less clear how one should proceed – whether to convey the main messages (e.g. a particular country having the highest rate and another country having the lowest rate) or whether to provide a complete description (this would involve the read out of data points, as for a table). Concentrating on the main messages keeps the reading interesting but may introduce some bias, whereas providing all data points makes the reading very dry but avoids introducing bias. Ideally, a publisher would do both – but that is perhaps not financially feasible. We decided to aim for the middle ground (see example in Annex B). The bottom line is that any alternative text is better than the default reading of the file name of the image.

4.3. Costs

4.4. How we apply the tag set

We use Inera’s eXtyles to generate JATS XML NLM 2.3²⁴ from Microsoft Word documents. The <disp-formula>Display Formula, <fig>Figure, <graphic>Graphic, <inline-graphic>Graphic and <table-wrap>Table Wrapper elements contain two child elements named <alt-text>²⁵ and <long-desc>²⁶ as accessibility elements. (Other elements also contain these two child elements.)

The JATS documentation for <long-desc>states: “Full description or summary of the content of a graphical object, table, or textual object such as a text box. The <long-desc> is not a visual element; rather, it is used to be spoken in circumstances where the graphical element cannot be viewed. A similar element <alt-text> is also not usually visual, but instead is short and may be displayed when the figure, caption, or other major element is not available or cannot be processed by the person or device type being addressed, for example, “behind” the graphic on a website.” Therefore, it seemed a better fit to use <long-desc>based on the accessible content and targeted delivery mediums.

5.1. Decorative but informative content

Many publications have content that is part of the “design” but is omitted in the JATS markup, either due to convention or lack of a JATS tag. Examples of such content range from the dingbats (graphic glyphs) used to separate sections, to illustrative but relevant artwork that may adorn a page. Some of these illustrative design elements re-enforce the messages in the text. An example may be picture of a girl with polio at the start of an article that discusses the disease. This may apply also more to the book than the journal tag set but is probably relevant if JATS is used for news-type free-form articles that could appear at the front of journals.

Ideally there should be a way to signify that a given image, text box or other element contains “supplemental content”. An example of this could be:

<xref ref-type="fig" rid="FA" type=”background | marginalia | decoration” alt=”World Map as background image for the text”/>

There are other images that are purely presentational and a publisher would continue to mark these as an “Artifact”.

5.2. Alternative text on text

Abbreviations are a very useful feature of written language but they are quite often misread by screen readers. Such cases include:

1. Abbreviation is the same as a normal word: WHO (World Health Organization) versus who (who ate my porridge?)

2. Abbreviations whose meaning is context-dependent: Jones Dr (PhD) versus Jones Dr (road) – or St Thomas (the saint) Thomas St (road).

3. Abbreviation whose non-abbreviated version is no longer used: Qantas (used to stand for Queensland and Northern Territory Airlines).

Therefore, it is important to keep using abbreviations in publications while making them more accessible. The Web Content Accessibility Guidelines 2.0² has recognized the value of such a solution.

JATS²⁶ includes the <abbrev>tag but notes: “Abbreviations, while common in STM books, are rarely tagged in the STM journal world, and will not be present in many journal DTDs. This element is expected to be quite rarely used.” Publishers could investigate approaches to using the <abbrev>tag more effectively, such as automation.

Pronunciation of a word can change its meaning – for example, “I read a book yesterday” and “I read books regularly”. Different pronunciations for words are quite common (e.g. Worcester, MA) and have been taken into account in implementing tools, such as DAISY readers, where users can load pronunciation files to enable a better user experience. But publishers may wish to provide an “authoritative” pronunciation for which it would be useful to include a tag within JATS that specifically provides for a reference pronunciation. This tag may be similar in nature to the <abbrev>tag.