The conference just finished, and I'm writing down some impressions while they're still fresh.

Day 1 kicked off with my keynote on "DITA evolves" - starting with an introduction to DITA, then talking about the ways in which DITA can adapt to different environments or authoring requirements, can scale from simple pilot projects to complex cross-enterprise applications, and the resulting network effects that can emerge from the combination of pervasive usage and robust reuse.

Then the two-track conference really started, with strong sessions in both management and technical tracks. I concentrated mainly on case studies on both days - I'm always interested in seeing how well DITA is delivering on its promises, and where the seams are stretching in a way that suggests new requirements on the standard.

The first case study I attended was really a survey of 5 different reference projects presented by Comet Communication GmbH, with domains including software documentation, product information (including manuals, online help, and leaflets), and training. It was strongly FrameMaker based, with authoring in a book context and generated topics and maps feeding into the DITA Open Toolkit. The result was not what I would call strongly topic-oriented, but it got them into DITA with a close fit to existing authoring patterns, with most content contributed by part-time authors with little topic-oriented or XML experience. The main lesson seemed to be that it was possible to start with DITA with relatively few elements and simple training, and grow over time as user needs dictate.

The second presentation, from XyEnterprise, looked at an existing Docbook-variant project implemented five years ago, and compared that experience with a DITA thought experiment: would that same project done today have chosen DITA? The answer, given the conference, was unsurprisingly affirmative: again, DITA's ability to play outside the stereotypical topic-based realm of Web and help was showcased: in this case, for complex/sophisticated print publishing scenarios. In this case, the company's eventual adoption of more topic-based approaches in any case affirmed that DITA, while certainly providing an adequate initial fit, would have also been a good strategic match with the company's direction towards more frequent and granular updates and more single-sourced delivery.

We then had a case study from Schlumberger and Mekon, on DITA's adoption for engineering documentation: in this case, DITA's advantages were quick implementation plus the ability to quickly respond to changing requirements: instead of a large initial design stage followed by a lengthy implementation, it sounded more like a quick minimal implementation that then evolved as the customer played with the real system towards a more complete design.

In contrast, Tanner AG's presentation on Siemens Medical's adoption of DITA as an enterprise content strategy was a full-on DITA adoption story making use of almost all of DITA's features in the initial design, even up to sharing specializations across corporate boundaries and using generalization to enable compatibility across different specializations: capabilities that DITA is designed for, but it's nice to see them being used:-)

Another nice piece of news out of Siemens Medical was the potential creation of a Medical Devices subcomittee - which, along with a potential subcomittee for semiconductors launched later that night, broadens DITA's appeal to specific industries through the use of specialization.

Day 2 kicked off with an end-to-end demo from Trisoft, which emphasized (for me) two things: the need for a more complete picture on taxonomy-based classification and retrieval down to the paragraph level (ie kicking up a notch our existing simple-value multiple-attribute profiling), and the need for more vendor support on DITA's draft-comment or review information capabilities.

Frankly, the fact that reviewing is now the weakest point in DITA's end-to-end tool chain is fantastic. I wouldn't have dreamed a year ago that planning, authoring, reuse management, and publishing would be considered mostly solved problems. That said, the ball is in the vendor's court to place the last piece of the puzzle.

Then we had a number of good case studies, including ATI with Ixiasoft, Kone with Seicodyne, and Boeing and the Canadian Competition Bureau with Stilo.

So: graphics chips, elevators, airplanes, and training for the legal domain. Again, a common thread was DITA's ability to start small and simple, then evolve quickly with the customer's requirements: lowering the cost of entry, and allowing for earlier proof points to establish ROI and refine requirements based on real usage.

Since adaptability, scalability, and evolution were the themes of my presentation, needless to say (but I'll say it anyway) I was unbearably smug: it's always nice to assert a generality that is proven right by other people's hard work :-)

We closed off with a presentation from Mary McRae from OASIS on the general value of driving DITA's design through an open standards process, and then Miel De Schepper of Trisoft on how DITA is crossing the chasm from early adopters to widespread adoption, with beachhead reference applications in the mechanical equipment and telecom industries.

 The conference was significantly larger than last year, and with many more case studies across a variety of industries: we're well beyond IBM's original case study now, and I can't wait to see what the next application will be.

Greetings and thanks to my fellow presenters and attendees, and I will really truly honestly try to be more regular in future blogging, and perhaps as a result present my content in smaller, more topic-oriented chunks.

Thanks especially to JoAnn and co. for running a great conference. 

PS:I should note that I attended less than half the presentations. That is the eternal torment of a two-track conference, and means that my report gives considerably less than half-justice to the conference program.