I first heard about Michael when I started to work on my O'Reilly book "XML Schema." I was already convinced that XML Schema was doomed to be technically horrible, and my motivation was to write a book as honest as possible that would help people work around its weaknesses.

Michael was chairing the W3C XML Schema Working Group. The WG appeared to be growing out of control, in size and in political divergence between big companies. I had no idea who Michael was; his initials "CMSMCQ" looked like a conglomerate by themselves, and I had no doubt he was some kind of incompetent manager from a major company.

I was involved in competing schema languages, and my heroes were James Clark, Murata Makoto, and Rick Jelliffe, whom I considered (and still consider) geniuses way above my own technical skills and abilities.

Things were pretty binary for me: on one side, free, independent geniuses making no compromises and striving for technical excellence (the Good), and on the other side, business-driven mercenaries ready to accept any technically ugly compromise for their personal or corporate interest (the Bad). And their leader was Michael.

I started questioning this vision when he answered some of my questions on the XML Schema mailing list. This manager had read the latest drafts, and he seemed to understand what it was about. He was even honest enough to acknowledge that he might be wrong. That was a time when the specification was so contradictory and complex that WG members would start their answers with "my understanding is that…"

And of course, I saw that I was completely wrong when I met him and saw some of his presentations.

After that, Michael has been, and still is, something of a mystery to me.

Why would someone so intelligent and erudite chair a WG editing a specification as technically insane as XML Schema? He was on the same level as my heroes; alone, he could have drafted a masterpiece. I cannot believe he might not have seen the weaknesses of the work he was chairing.

To me, comparing XML Schema and Relax NG or Schematron was like comparing elevator music and Beethoven's Ninth Symphony or Mozart's Magic Flute. Why would a major composer write elevator music if not for money? And Michael was not chairing the W3C XML Schema WG for money.

I have always felt intimidated by Michael. He was kind and encouraging to me, but his stature, his intelligence, and his knowledge made me feel like a child in front of him, and I never dared ask him the question.

My guess is that he knew XML Schema was a terrible specification and that he was probably hurt to be associated with this work, but that he did it because he thought it was the right thing to do for the community — that his goal was not to strive for excellence but to avoid a complete disaster, that he sacrificed his personal inclinations for the interest of the community.

I will never know if I am right or wrong, but that's the image I keep of CM Sperberg-McQueen.