When can I do some XML with you? Adapting and reusing this Beatles lyric for our theme works best in the UK, where Rita rhymes with Meta and DITA; It's a stretch in the US, where Meta sounds like better and DITA sounds bitter.
But since adaptation and reuse are core ideas of DITA (Darwin Information Typing Architecture), perhaps we'll be forgiven if we adapt and reuse old Beatles standards to explain the newest XML standards (hey, maybe it's the only way to make XML sound catchy). DITA is an IBM gift to the technical documentation community that was approved as a standard this spring by OASIS (Organization for the Advancement of Structured Information Standards), the hosts for many XML interchange standards such as ebXML. Ever since, tech writers have been buzzing about an easier way to get into structured topic-based writing with DITA XML and asking XML Editor vendors to add support for DITA.
Leading XML authoring tool vendor Arbortext (owned by PTC) has included DITA in its Epic Editor for some time. Adobe's September release of Framemaker 7.2 adds a bit of DITA support (along with rumors that Adobe's internal documentation team has used DITA for their latest manuals--print, Web, and online help all derived from the same original files). And Blast Radius last week announced its XMetaL Author DITA Edition, purportedly in response to customer demands. XMetaL integrates the DITA Open Toolkit, a complete DITA implementation available for free download from SourceForge. Altova XMLSpy, Ektron eWebEditPro+XML, Ephox Edit Live!, Syntext Serna, and Stylus Studio all promise DITA enhancements.
So why the buzz? Can DITA do anything XML could not? The answer is probably no. XML can do almost anything needed for structured authoring but its arbitrarily extensibility has been a stumbling block to XML adoption. The promise of DITA is ready-made metadata structures that may fit the needs for many documentation projects, initially for software, which was IBM's principal use.
While it is doubtful that out of the box DITA will find widespread use without customization (called specialization in DITA speak), the ready-made generic topic, and three "information-typed" specializations called concept, task, and reference, will get documentation teams producing very quickly. These documents will also be easily exchangeable with others. Because specializations inherit (thus the Darwinian name) properties from the general topics, their default behaviors--like printing, conversion to PDF, or XHTML Web pages--will produce decent results when transformed by default DITA XSLT style sheets.
Nothing in technical documentation is simple, and with several hundred terms to learn DITA is not trivial, but it is relatively very simple compared to arbitrary XML, and the XML editing tools with their sophisticated visual IDE's mask much of the complexity. To their credit, these editors do not hide the structure from authors who appreciate structured topic-based writing.
So may we compare DITA to RSS (Really Simple Syndication), today's most-used XML application? When Dave Winer offered his special XML tags for wrapping news feeds a few years ago, the explosive adoption rate was because this was XML that everyone could do the same way (pace the syndication wars with the RDF folks and the new Atom standard). If XML extensibility had invited programmers to "do-it-yourself (DIY)," RSS produced a network effect where we could all "do-it-together (DIT)." I can hear another Beatles tune, "All Together Now" with new lyrics: Do It Together All! D-I-T-A, can I XML this way?
The really big return on DITA investment will be teams with a usable XML Schema agreed on in days rather than months, plus a few standard style sheet transforms they can cut their teeth on. They will learn very quickly that they can't just stick their old content into DITA topics without learning how to structure their writing. This will force them to learn about structured content models. They will group topics with DITA maps (much too easily confused with Topic Maps) to help them practice good information architecture.
By then they will be ready to specialize their topics. This will be almost the same amount of work as creating new XML tags, but with the advantage that new topics will immediately work with default DITA behaviors. And once new DTD or XML Schema modules have been written to extend (or specialize) the default schemas, and XSLT processes extended to handle the new information architecture, they will be doing single-source publishing and reusing content as well as any XML guru.
Turning again to our Liverpool lads, can't you hear DITA singing a little ditty: "I want to hold your hand" to every technical writer and wannabe XML developer?