Re: Consensus checkin

Confirmation of my views inline below:

Thanks,
Tom
On 21 Apr 2022, 09:45 +0100, Norm Tovey-Walsh <norm@saxonica.com>, wrote:
> Hi folks,
>
> I would not be surprised if some of you are starting to regard me as
> something of a nag. I can live with that. By my count, there are seven
> (7) more meetings scheduled for the ixml CG before XML Prague. I expect
> we will miss at least one of those, so it makes more sense to count it
> as six. And, realistically, the only thing we could hope to do on the
> last meeting before the conference is agree that the spec is finished or
> propose some small editorial corrections. So five (5) *at most*.
>
> If we don’t make progress in email, we won’t be finished in time for
> Prague and I think that would be very unfortunate. I happened to mention
> ixml to a couple of other XML folks recently. One asked “is that still
> going on?” and the other said, explicitly, that they’d given up on it
> because it was taking so long. We need a finished spec in Prague.
>
> To that end, here’s a summary of where I think we are with respect to
> the open issues. It’s a bit of an exercise in reading the tea leaves of
> email threads, so please don’t take offense if you think I’ve got some
> of them wrong. Alternative readings invited.
>
> * Error codes
>
> I think we have consensus forming around option 3, (two digit?) numeric
> error codes inlined in the prose, with the text of the error messages
> exclusively in an appendix. I think there’s one outstanding disagreement
> record.
Option 3 good; I would request bi-directional linking between appendix and reference(s) in the text.
>
> * Change “~” to “!”
>
> Consensus forming? Most folks commenting seem to agree that “!” is
> likely to be more familiar to our users than “~”. I think there’s one
> outstanding disagreement on record.
Happy with either; agree that users are likely to be more familiar with '!'.
> * Use “=” and “|” exclusively in rules
>
> I think there’s consensus forming around this change. There’s clearly
> some outstanding affection for “:” and “;”, but I think everyone who has
> replied has expressed support for the proposal and agreed that it would
> clarify grammars for users. There are no disagreements on record, as far
> as I can see.
Not a necessity IMO, but an improvement to readability to use '|', and familiar to users of RNG/DTDs.  I prefer ':=` to either character on its own, to be honest, but I know that there is an appetite for single character syntax 'furniture'.
> * Version declaration
>
> Unclear. I count one explicit endorsement and no objections. There’s
> only one comment on the proposed wording for the spec:
> https://github.com/invisibleXML/ixml/issues/63#issuecomment-1100917298
> and it’s positive. More discussion would be good, but I think we might
> have the beginnings of consensus here as well.
I think an optional version declaration is an invaluable addition to the spec.
> * Change the insertion mark
>
> I’m tempted to describe consensus forming around “+”, but I think that
> might be overstating the facts. Two members of the CG have expressed
> the strong opinion that “^” is unsatisfying, perhaps even
> unsatisfactory. At least one member has agreed that “+” would be
> better. More discussion would be good.
I am strongly in favour of '+'.
> * Namespaces
>
> This appears to be the most hotly contested of the new features. By my
> reading, the majority is in favor, although I suspect there are a few
> more details that need to be worked out. I think there’s one
> outstanding disagreement.
I don't think I need to spell out my position any further.
> Personal proposal:
>
> In the interest of reducing the amount of work we have to do in
> meetings, I’d like to encourage the chair to declare that we have
> consensus on, at least, error codes, the syntax changes, and the version
> declaration and see if anyone raises an objection. I’m happy to
> volunteer to help get a new spec drafted with those changes in time for
> next Tuesday’s meeting.

> If there are no objections to changing the insertion mark, then I’d add
> that to the list as well.
>
> That would leave five weeks to come to consensus on namespaces and deal
> with the hundred other small issues that are bound to arise as we try to
> cross all the t’s and dot all the i’s. We might just make it.
>
> Be seeing you,
> norm
>
> --
> Norm Tovey-Walsh
> Saxonica