Microsoft grows DAISY for blind computer users; Adobe wilts

Accessibility format gets boost from maker of world's biggest text-authoring tool
  • Eric Lai (Computerworld)
  • 12 May, 2008 10:16

The release of an esoteric plug-in for a twenty-year-old piece of software normally doesn't merit much attention... except when the software is the ubiquitous Microsoft Word, and the add-on could have a major positive effect on the 1.5 million blind or visually impaired Americans who use computers, millions more like them around the globe, and, potentially, tens or hundreds of millions of people worldwide with developmental disabilities or reading problems.

Earlier this week, Microsoft announced the availability of a plug-in (downloadable from openxmlcommunity.org) that lets users of Word 2007, 2003 and XP easily save documents in the DAISY (Digital Accessible Information SYstem) XML format.

DAISY XML is the latest iteration of a decade-old standard developed by the DAISY Consortium, a leading nonprofit group serving the vision-impaired, to be the most accessible format for blind computer users.

Hidden structures

Why DAISY, when screen readers and text-to-speech tools already let blind computer users hear HTML Web pages and Word or PDF documents recited aloud?

For one, the experience, as illustrated in an April feature on computing for the blind, remains intensely frustrating Narrator, the screen reader built into Windows XP and Vista, is so crude that even Microsoft admits that it is not suitable for daily use.

Meanwhile, popular third-party readers, such as JAWS, are expensive. The standard version of JAWS, for instance, costs US$895; another package, EasyConverter from Dolphin Computer Access Ltd., weighs in at US$5,200. And the experience with JAWS and others remains, according to Curtis Chong, president of the National Federation of the Blind in Computer Science, uneven or poor. "If something is coded up wrong, your screenreader sees nothing," Chong said.

Why so bad? The problem is that for the blind, the most important parts of a page are the parts even the sighted can't perceive -- invisible metadata embedded in the document. What's missed isn't the stylistic metadata that sighted users usually think about, such as font, size, or color, but attributes such as paragraph marks, table structures, and headings, which determine a document's actual structure.

Good structural metadata lets a blind user nimbly navigate, browse and search a document. Word and PDF weren't built from the ground up to support that. DAISY was.

"DAISY is a fantastic format due to its flexibility," said Sam Ogami, an assistive-technology expert for the California State University system's chancellor's office. "From DAISY, you can easily move to other accessible formats, such as Braille or large print, in addition to audio, with little to no extra work."

Page Break

Letting users of Microsoft Word -- the most popular text-authoring tool on the planet -- save documents in DAISY format with one click is a "great step in the right direction of creating accessible content," said Ogami, who has tested the plug-in.

George Kerscher, secretary-general of the DAISY Consortium, was even more effusive.

"We would like all publishers to make their content painlessly accessible," he said in an interview earlier this spring. "Microsoft is the first one to step forward to do this."

Growing a DAISY plug-in

The plug-in was developed by Microsoft, the DAISY Consortium and an Indian software vendor called Sonata Software. It is also being hosted on SourceForge as an open-source project.

Andrew Savikas, a publishing software guru at O'Reilly Media and head of its annual Tools of Change for Publishing conference, said the Word plug-in was "long overdue."

When Microsoft started switching from binary to XML document formats in Word 2003, "this kind of conversion / transformation became much more transparent to implement," he said. He conceded, though, that "you could argue that a plug-in like this should have come from a third party, rather than from Microsoft, who I'd assume aren't interested in developing and supporting a bunch of plug-ins for formats they don't control."

Chong called the plug-in "wonderful," but cautioned that there remains a gap between theory and practice.

"If the initial Word document wasn't marked up properly by the author [with metadata] in the first place, then it's as bad as not having the document at all," he said.

Jutta Treviranus, a professor with the Adaptive Technology Resource Center at the University of Toronto, argues that this dearth of "consistent guidelines" for authors interested in creating potentially accessible documents in Word is only one of several problems.

In a paper she co-published earlier this year, Treviranus argued that Word 2007's native document format, Office Open XML (OOXML), violated other fundamental tenets such as not conflating stylistic metadata with structural metadata.

"I have grave concerns with the DAISY XML that will be produced" from a Word 2007 document, Treviranus said.

Meanwhile, Marino Marcich, managing director of the ODF Alliance, argues that Microsoft's format remains inferior to the OpenDocument Format he champions.

"The accessibility of ODF was reviewed and subsequent changes were incorporated in ODF v1.1, establishing ODF as the benchmark, exceeding the accessibility features of any other document format," he said.

Reed Shaffner, a product manager for Microsoft Office, admits the Save-as-DAISY translator is far from perfect today. Highly structured documents -- an IRS 1040 tax form, for instance, with its multiple fill-in boxes -- still pose difficulties for the plug-in.

Shaffner is quicker to defend the OOXML format's accessibility features.

Page Break

"It's already pretty strong. We're making animated graphics accessible, for instance," he said. Some improvements were made during OOXML's ISO ratification process two months ago, and more are coming.

Treviranus, he argues, "may be confusing [a document author's] bad behavior with document standard behavior. The information is there in the standard."

As for criticism from the ODF camp, Shaffner said that ODF may support accessibility "in theory, but we've put out a translator."

What about Adobe?

Microsoft's proactive embrace of DAISY begs the question: what is Adobe Systems, the other leading document creation software vendor, doing to support DAISY?

Nothing directly for now, admits Andrew Kirkpatrick, senior manager for accessibility at Adobe, though he claims "it is under serious consideration. DAISY may be a useful format to export to, particularly in the case of longer documents such as those created by FrameMaker."

We've heard that before, says Kerscher, who says he has repeatedly asked Adobe for a commitment to support DAISY and failed each time.

Agrees Chong: "Adobe has done a lot of work to make reading a document accessible, but it has done far less work on the composition side."

What Adobe has done, instead, is enable the latest version of its page layout program, Creative Suite 3 InDesign, to export into the ePub format. ePub supports the NIMAS standard, which Kirkpatrick characterized as a "subset of DAISY."

O'Reilly Media's Savikas thinks Adobe has made a smart choice by backing the ePub format, which he says is fast becoming the "mp3 of eBooks." But he said his initial experiments with Adobe's ePub feature left him "disappointed."

"We ran tests on a few of our 'Head First' books, and the resulting output was essentially useless," he said. After some "back-and-forth with Adobe," it became clear that the books had to be redesigned with extra formatting metadata for ePub, said Savikas. "Most people who hear that InDesign can export to ePub assume it's as easy as "Save as..." and it's not."

Besides creating a formal program to help third-party screen-reader vendors make their software work better with Adobe software, Adobe is trying to improve the accessibility of its own formats, such as PDF, SWF (Flash) and others, said Kirkpatrick.

"DAISY is popular and may well be a valued format for blind users, but Adobe is not prepared to determine what is the format of choice for blind users," Kirkpatrick said.

Chong agrees, saying that most blind computer users could care less about the war of words over DAISY versus ePub.

"It's not the format that is of concern to the blind, we just want to make sure there are programs that we can use," he said.

Page Break

Let a thousand flowers bloom

In that vein, a plug-in to convert PowerPoint slides into DAISY XML is in the works, Shaffner said, with the DAISY Consortium just finishing gathering requirements for such a tool.

An Excel plug-in is unlikely, he said, due to technical difficulty and lack of user interest.

Meanwhile, the existing Word plug-in is built in such a way that it could someday be ported to versions of OpenOffice.org that support OOXML, such as Novell's version, Shaffner said.

Besides accelerating the conversion of electronic documents and books into DAISY for visually impaired computer users, the plug-in(s) could aid in the DAISY Consortium's new goal of helping to make documents and books accessible to the illiterate, dyslexic or developmentally disabled.

For instance, the DAISY Consortium is helping to convert HIV and AIDS prevention manuals into the DAISY format so that poor, illiterate South Africans can access them. DAISY might also be used to "help synchronize text and audio and help kids who are learning to read," Shaffner said.