A common shock in this virtual world is the first in-person encounter with someone you've known by email or phone. First, the double vision - your mental image contradicting the tangible person before you. Then the switch, and the person before you becomes the one you knew all along.  That good rewiring has happened a lot here at the CMS Conference, along with catching up with folk from previous encounters.

The conference had one track with an official DITA focus, but as people talked about their Content Management strategies in the management, technical, and demo tracks, DITA kept coming up. (I'm tempted to try to make capping remarks about that other, East Coast DITA conference, but the people who went to both said they were glad they did and neither conference sticks to the one coast anyway.)  The conversation has shifted from curious buzz ("What's a DITA good for? Will it clean my carpets while I'm away?") to experience ("Well, I don't see that it's necessary to centralize all of my conref targets").  Lots of case studies, progress reports, and lessons learned along with an optimism that we can make more progress - the highlights including insights from people like France Baril, Frances Gambino, Anna Hartman, and the Research in Motion folks.

Besides the pragmatism, one of undercurrents seemed to be early adoption for structured information outside of the technology domain.  That makes sense -- so far as I know, base topic doesn't have anything specific to the technology domain -- and people were noting the content fit with financial reporting and medical content (for instance, as Alex Witzigmann explained in his presentation on Siemens Medical) as well as the learning architecture that John Hunt has been working on.

Another undercurrent seemed to be a growing interest in formal metadata and  taxonomy, which came up in JoAnn Hackos and Miel De Schepper's tag-team keynote, in the Siemens' presentation (where Alex referred to the need for relationships to formally defined resources), in Erik Hartman's survey of different semantic initiatives, and in Paul Arellanes's progress report on IBM STG's taxonomy work.  Shameless plug - if Paul's presentation piqued your interest in the DITA taxonomy specialization or taxonomy / ontology in general, you might want to check out:

• Subject classification with DITA and SKOS
• Managing semantics with content using DITA XML
• Ontologies Come of Age

The basic thought is to use as much formality as you need and no more. The caution is that the DITA taxonomy specialization is still a work in progress (at the "Will it clean my carpets?" stage).

My only complaint to the ComTech folk is over the frustration of wanting to be at two or three sessions at once.  Well, that, and having my brain torqued a bit by converstion with people like Andrzej Zydron and Bruce Esrig.  With a nod to San Francisco, half reverted to a sleepy waterfront in the rain, it's a wrap.