The DITA 2006 conference has started its regular program. Dave Schell, the keynote speasker, presented some IBM experiences with publishing, outlining the costs of some of the previous ways of publishing product documentation, how those experiences informed on the design of DITA, and how the DITA model has brought about specific savings on several projects that have had one or more publishing cycles by now.

Cost avoidance is one of the key issues that Dave discussed. With traditional publishing, the cut/copy/paste penchant of writers was identified as one of the main causes for increasing the amount of review and update as various information sets "forked" from some original version.  In moving to a topic-oriented architecture, one team was able to relegate up to 80% of its information as "common", now completely out of the picture for ongoing review and translation and other handling. Another team working on a completely different product got up to 77% common reuse at the topic level (including some conref reuse as well--phrase-level reuse by reference).

Dave discussed some other values, particularly the general value of DITA as a foundation for collaboration (enterprises sharing content with their OEMs, for example) and as a foot in the door for consultants and vendors looking for document architectures that might meet particular customer requirements.

I'll attach some pix throughout the day. End of Dave's session.

11:00 am--I spent some time with exhibitors.  Last Fall, JoAnn Hackos wrote an article, Is DITA Going to Tip?, which discusses the level of usability that enables new technologies to really resonate with adopters.  I have the impression that the pieces are starting to fall into place... many vendors "get DITA"--the understanding of how the architecture differs from traditional XML applications, so that users can exploit the tools, rather than fight with them.

11:30 I presented on the activities of the OASIS DITA Technical Committee, represented mainly by the outline in the Roadmap for DITA development.  Robert Anderson followed with a demo of some of the translation capabilities of DITA, based on an actual DITA specialization based on a music collection cataloging application--very interesting!  Can you imagine exchanging playlists in valid DITA?  Well, its coming soon to a DITA Open Toolkit near you, as soon as he gets the last of his legal approvals to release that code.

1:30 Alex Povzner commented on the role of information architectures in an enterprise environment, suggesting some uses of the semantics that can be encoded in the markup.  His premise of content vs knowledge reminds me of the quote that I use in my email sig file:

"Where is the wisdom we have lost in knowledge?
 Where is the knowledge we have lost in information?"
   --T.S. Eliot

DITA's reuse model helps eliminate redundancy, one of the most common issues that prevents information for being reliable.  Alex invoked the phrase, "Semantic Web," and discussed the OWL and RDF standards for ontologies (facts are triples: the subject, the object and the predicate of the fact).  He gave an example of how to build a semantic model using an example of a bicycle.  Far too many good ideas to record them all--but I expect that there will be many discussions in days to come about how to embrace DITA for better use of the knowledge embedded in the doc assets of companies.  Even before SGML was formalized, it was recognized that proper markup is actually a corporate asset--done correctly, a structured document can be analyzed to give a lot of information depending on how you can query the relationships in the document. With missing or wrong markup, those relationships don't exist; with properly used markup, especially well-defined specializations, the document becomes an exhibit for interrogation by SiberLogic's tools. Very cool capabilities, indeed!

A final reflection on Alex's presentation--if you expect to use techniques like these to derive value from your DITA documentation, it becomes more important than ever that your writers invest their deep knowledge about the product and systems by using appropriate markup at every opportunity, and above all avoid using markup wrongly.  A key phrase marked up using <ui-control> just because this element gives bold output in an editor will be almost impossible to catch as a markup mistake, and that phrase will be lost to any ontology-aware searches that expected a different semantic markup usage.  Don't abuse your DITA data!

I later attended Michael Miller's session on Antenna House, a formatting tool that is based on the W3C XSL-FO specification.  Because DITA's processing is fully based on the W3C processing model, FO tools are very friendly for printing DITA content. The group had a short side discussion about outputting to other formatting tools including FrameMaker, and recognized that FO is always available as a fairly easy and reliable alternative for getting PDF output.

Attachment	Size
IMAGE0009.JPG	91.84 KB
IMAGE0011.JPG	72.26 KB
IMAGE0003.JPG	47.76 KB
IMAGE0007.JPG	167.1 KB