QT4CG, the commuity group developing the 4.0 versions of XPath, XQuery, XSLT, and related specifications, has been hard at work now for more than three years. We just had our 151st meeting.

It’s a large body of work. Lots of changes have been requested: there are 2,437 issues today, with 70 remaining open. Lots of changes have been proposed, drafted, reviewed and accepted (or rejected, as the case may be): there are 2,441 pull requests today, with 12 currently open. We’ve added tests to the test suite and implementations are underway.

I’m pleased, in my role as chair of the group, to announce that we’ve published our first set of “stable drafts”. These are not final specifications, but they do reflect the consensus of the group near the end of January, 2026. The current drafts are still the place to look for the most recent consensus positions.

Publishing dated, stable drafts serves a couple of purposes. First and foremost, it gives users and implementors a stable set of documents to refer to. As Saxonica works towards a Saxon 13 release, we’ll be able to make assertions about what features we do (or do not) support, confident that those features won’t change in the draft we’re referencing. Those features may still change, of course, but we can say we implemented against a stable draft.

It’s also a strong signal to the community that we think we’re making progress towards the finish line.

Yes, but when will you be done, I hear you ask?

In practical terms, when the group thinks the specifications are complete and coherent, and when it agrees to stop changing them. That’s a social engineering issue, not a technical one. Specifications, perhaps even more than code, are governed by the 90/90 rule: the first 90% of drafting takes 90% of the time, the remaining 10% of drafting takes the other 90% of the time.

Personally, I’ll be very surprised if we finish in 2026 (that remaining 10% really is a lot of work), and surprised [if somewhat less very], if we don’t finish in 2027.