Meeting minutes
Accept the minutes of the previous meeting
John: We had a whole section on epsilon etc; it says "John didn't know", but I was quoting Manuel from Fawlty Towers, and said "I know nothing".
Review of open actions
2024-03-05-c: SP to prepare pull request to resolve issue #139 (new grammars, new README)
Augmented by 2026-07-07, SP to replace his PR with three separate PRs.
Continues
Status reports
Norm: My implementation runs in C# in context of Saxon 13
… transpiled from Java
… so now invisible XML function works in Saxon
Steven: I have been writing a C implementation over the summer. I added negation to it yesterday. Not yet tested on the suite, but seems fast.
Norm: We should encourage implementors to publish results against the test suite
John: I'm having trouble getting the test suite from codeberg
… it thinks I'm a bot, and so won't do anything for me.
Norm: You can get the test suite from invisiblexml.org as well
… same problem with W3C if you try to get the HTML DTD too often.
Steven: Does codeberg host invisiblexml.org?
Norm: No. There is some hosting in the eu.
<norm> My recollection is that it's statichost.eu, but I'm not sure off the top of my head.
ACTION: Norm to discover where ixml.org is hosted
Progress report on symposium output?
Bethan: All done
Norm: I sent everyone a mail, I will upload to archive.org and announce
ACTION: Norm to upload and announce videos
Next symposium
Steven: Last time we had a deadline of 15th December for submission, so we need to think about it.
Steven: Last Thursday and Friday of February would be 25/26 Feb
All: Looks OK
[We look at last year's call]
ACTION: Norm to upload new call to website
ACTION: Steven to dig out last year's announcement and share
Pull requests
+ A sample grammar for Gregorian dates
See https://
ACTION: Steven submit example grammars
Detecting ambiguity
https://
[Steven reads proposed new wording]
<norm> If we're word-smithing: "Parsing algorithms may differ slightly in the details of how ambiguity is detected, however, if there is more than one interpretation of the parse possible according to the algorithm used, this *must* be reported."
Bethan: I think that all parsers will spot ambiguity in the same way.
Steven: Fine. Then we delete this sentence.
Steven: By the way, long ago we agreed to change the should to a must for adding the state="ambiguous" attribute
Bethan: the attribute should be "ambiguous"
Steven: But you can have state="prefix ambiguous"
David: Machine or human readable?
Steven: Both
John: The attribute space could be expanded
David: Maybe it should be several boolean attributes
John: There are other possible values an implementor may wish to report
Steven: State is almost always single valued
John: Spotting a failure is hard when the result is an XML instance. We should rethink failure modes
Steven: I need XML in cases where it's in a pipeline
Norm: I wouldn't want to prevent that
ACTION: Steven update the ambiguity section
Next meeting
October 13th
[ADJOURN]