And yet, if we don't seize opportunities to put the technologies we are working on into practice, how shall we escape the charge that what we are saying to others is, in effect, "Do as I say, not as I do"?

When we began work on the Text Encoding Initiative, a large project to develop an SGML (now XML) vocabulary for use by scholars in the humanities, I suggested that all internal documents be written in SGML, preferably in an approximation of the tag set we were defining. Lou Burnard, my co-editor, argued that this was crazy, since (a) the TEI Guidelines were primarily intended for representing pre-existing documents, not for document production, and (b) we didn't have any software to produce documents using SGML. He wanted to use LaTeX instead. From a project-management point of view, perhaps he was right: we might have been finished faster if we had used LaTeX instead of struggling to refine our tag set and develop our tools at the same time as writing the Guidelines. But I think the Guidelines are a lot better for having been edited with the understanding of markup in practice that we got by insisting on using SGML ourselves.

It's not clear whether I am disagreeing with you or not. In practice, I don't always do what I'm preaching here: I continued to use SGML for most of my documents until well after XML became a recommendation, since I had a well-working SGML tool chain, and did not switch to pure XML until ... well, a lot later than some of my colleagues.

Then there was the proposal that the chairs make it a requirement for membership in the [name redacted] Working Group that each member submit, every six months, a working example they had written of a [name redacted] using the current draft spec, to show that they knew how to use the technology they were specifying. I can think of Working Groups which would have been much less crowded if that test had been applied.

Using our own technologies as we develop them would generally make them better, I agree wholeheartedly. Instituting requirements along these lines in standardization bodies would be an intriguing experiment, but in such a political setting, it might be dropped on a technicality.

But criticizing others for not using their own technologies should be done after careful consideration. But then, I'd like all criticism to be done after careful consideration, so I guess there's nothing new here...