Re: QT4CG meeting 136 draft agenda, 30 September 2025

Dimitre Novatchev <dnovatchev@gmail.com> writes:
> How many times did the syntax (operator) and meaning of method execution change? Is the `%method` annotation existent at present? What is the latest signature for methods - is the map passed as a parameter, or accessed as `.` ? Is all this stable now?

I think most of the relevant PRs have been dealt with. Which is not an assertion on my part that some other change won’t come up next week.

> Also, the comments from members of the group (MHK) that "I would prefer to see it in a separate specification that is not part of the core language specs", are, to say the least, discouraging. If this group is dictated by the opinion of a single person, as it is most of the time, we need strong guarantees that a PR would be judged objectively, regardless of the pre-existing opinion of anyone.

This comes across as accusatory, Dimitre, and it disturbs me. Everyone is entitled to their opinions and expressing them during meetings is a way of establishing where consensus lies. On any given opinion, some people agree, some people disagree, and some people are unsure. Some agreements are “yeah, well, I guess so” and some are “that’s absolutely true!”. Some disagreements are “uh, personally, I think that’s a bad idea” and some are “never, not on my watch!” MK’s opinions are no different from anyone else’s in that regard: I recorded my own view that his proposal to allow user-defined functions without a namespace was a bad idea. I still think it’s a bad idea, but we did it anyway because that’s where consensus lay.

The goal is always to find a consensus position that everyone can live with. That is not always the best solution, and the result of a consensus-based process is not likely to produce the cleanest, most consistent design. But it’s the only process we have.

Ultimately, PRs are accepted or rejected by the group as a whole, not a single person. Every substantive PR that’s submitted gets reviewed by the group. Its champion has an opportunity to attempt to persuade the group that the change (to the spec, to implementations, to the test suite, etc.; at this point in the process no substantive PR is simple) is in the best interest of our design and our users.

> I am willing to invest time and effort in writing this PR, but we need clear answers to these questions, so that this effort is meaningful and not destined for rejection even before it is started.

I don’t have a crystal ball. How well a PR is received is a combination of technical and social factors. I can only promise you that I will give you every opportunity to fairly present it to the group and to try to gain consensus to accept the PR.

My advice, if you’re interested, is to consider carefully the questions raised in our recent discussion of #1965. Especially, how much of what you envision for generator records can already be implemented in XPath? If all of it can already be implemented, then you (with the help of interested collaborators, I hope) will need to tell a very compelling story about why the change is warranted. If some part or parts of it can’t be implemented and need to be in the spec, how clearly can you define and motivate adding those specific things?

                                        Be seeing you,
                                          norm

--
Norm Tovey-Walsh
Saxonica

Received on Wednesday, 1 October 2025 16:55:28 UTC