(In the weeks leading up to the XML 2006 conference in Boston (5-7 December), I’ll be featuring some of our presentations here on my weblog from our four specialized tracks.)

Even before we read the abstract and realized what a great presentation this was, the planning committee was kidding me that they knew the paper would make it — after all, I keep a flying blog and had flown my own small airplane (not a Cessna) down to Boston for the planning meeting.

In fact, what attracted me most to this paper was not my own flying, but my work on large, SGML-based documentation systems during the 1990s. These systems, which often cost many millions of dollars to create, are still running and doing good service for organizations in sectors like automotive, aerospace, and the military, but it has been many years since vendors released new products for SGML, and support for old products from the 1990s is dwindling fast, as is the pool of people with high-end SGML skills. At what point is it worth tearing apart and rebuilding a working system to upgrade to new technologies? What are the pitfalls? This presentation could just as easily have ended up in our hands-on track, and it’s a strong incentive to make sure you arrive Monday night, so that you’re ready to drop by and hear Michael on Tuesday morning.