The design of XML was a truly creative effort, and my contribution as co-editor was far from the largest. The idea of well-formedness was originally, however, mine. It never got voted on; I just wrote the terms "validity" and "well-formedness" into the first draft; Michael Sperberg-McQueen demurred briefly, as the term had, for him, some other semantic baggage, but it survived in the absence of other suggestions, and when the draft became public, nobody ever challenged the idea.
It should be noted that the term "well-formed" does have another meeting in formal mathematical logic; an assertion in formal logic is well-formed if it meets certain grammatical rules; whether it is "true" or not in some framework is another question entirely. I found the analogy very compelling, and nobody else ever thought up a better name.
In SGML, they have adopted the terms "tag-valid" and "type-valid" corresponding respectively to XML's "well-formed" and "valid".
Back-link to specCopyright © 1998, Tim Bray. All rights reserved.