Meeting minutes
Accept the minutes of the previous meeting
Accepted.
Review of open actions
Steven: I've completed 2023-10-01-e.
Steven: I've completed 2025-04-29.
ACTION: 2025-08-19-a: Steven to review pull request #296.
Status reports
Nothing reported.
New open issues
John opened #303, Using multiple iXML grammars on ‘pre-processed’ input, invisiblexml/
John: It's a mechanism for describing grammars that you would include together by burying grammars inside an XML structure
… You can then parse by doing one pass that identifies sections that are then parsed by the appropriate grammar.
… I've done nothing on this since I created the issue.
John: I suggest that everyone read this and I'll see if I can make a version that runs on CoffeePot. I have one that runs on my machine.
… There's some follow-up from Nico.
Bethan: Perhaps we should wait until Nico is able to be here.
Steven: I did post an email about an issue.
John: If your parse knows about the grammar, I don't see why the processor shouldn't be able to give you a more complete error.
John: If you're treating this as the grammar that's going to be run, then you get this possibility of a more detailed error code.
… But if it's just any general sentence, I'm not sure.
John: There is a dual nature to our grammars, they're a sentence in our grammar and they have other constraints.
Bethan: This is a static errors.
Steven: We have three categories of errors: syntactic, static, and dynamic.
Some discussion of other errors in the same neighborhood.
We are discussing: https://
Steven: The grammar excludes this, S06 can never happen.
Norm: I think Steven is right, S06 can't actually occur.
ACTION: 2025-08-19-b Steven to remove S06 from the specification.
Pragma requirements
Bethan: We're at the last two points, #19 and #20.
19. A pragma’s attachment to a specific syntactic construct must be unambiguous
to software for parsing iXML grammars.
Steven: I still don't know what attachment means
David: Attachment is defined at the very top of the requirements document.
Requirements document: https://
Norm: I think this boils down to saying that if two processors parse the same grammar, they have to put the pragmas in the same place.
Bethan: We need different processors to minimally produce the same XML structure from the grammar.
Norm: Attempting to paraphrase #19, any processor that parses an iXML grammar must place all of the pragmas in the same place in the XML version of that grammar.
ACTION: 2025-08-19-c: Bethan to attempt to rewrite #19 so that it's clearer.
David: One possible source of ambiguity is that two processors could unambiguously believe that the pragmas went in different places.
John: If we take the XML version of the grammar as definitive, it's very clear where things are. We're getting a little more abstract than the XML structure of the grammar is the definition.
… The relations between every piece can the be based on the structure of that XML.
Bethan: I don't think there's one that's primary.
John: Making declarations within the XML grammar is very clear.
Bethan: The XML version that a processor produces has to be the same.
Steven: That's the essense of iXML
Bethan: The idea with this requirements document is to nail down the essential things, even the obvious ones.
David: Pragmas seem different than other aspects in that they necessarily have a processor-specific component. In any other construct, it would be self-evident.
Consensus: in principle, this makes sense, but we'd need to see the rewrite.
Steven: I still don't know what exactly we've agreed on.
Bethan agrees to redrafting along the lines discussed today.
Moving on to #20
20. The relationship between a pragma and the construct it attaches to should be
clear and unambiguous to human readers, in both iXML and XML notation.
Consensus: that makes sense
Bethan: Let's try to sort out #19 in email!
Modularization
Steven: Any progress, John?
John: Some but there's still a ways to go.
Steven: Since I gave the talk, I haven't done anything.
Norm: Steven, can you add the paper to our repo.
ACTION: 2025-08-19-d: Steven to add the modularization paper to the Invisible XML repository
An unconference...
Steven: I was really impressed by the talks at Balisage
… I don't think I have the time and energy to organize another whole conference, but something really light weight seems like it would be doable.
… I think that would be fun and exciting.
Bethan: I think it would be some work, but less than a full conference.
Steven: We need to write the call, and have some sort of registration system
Bethan: We'd want a web page, a code of conduct, and that kind of thing.
… There's a little more behind the scenes work.
John: A half day thing?
Steven: A half day would allow overlap between America and Europe.
Norm: That seems good to me.
David: How long would the sessions be?
Steven: I think we should let the speakers ask, up to maybe a maximum of 30 minutes.
John: It's about cross-pollination.
… We could also let speakers show their grammars and have folks looking over their shoulder.
Bethan: I think a maximum of 15 minutes makes sense.
Norm: I think 30 minutes slots would be good.
Steven: I think we might get a lot more than 8 talks, based on what we've seen so far.
Bethan: Maybe two half days?
Steven: That also sounds good.
John: What about January/February?
Some regrets about the fact that XML Prague has moved to the summer.
Steven: Yes, that might work.
… I'll try to assemble a discussion document for more folks to comment on.
ACTION: 2025-08-19-e Steven to assemble a discussion document for the conference.
Norm: What teleconference service?
Steven: I think I can use CWI's business account.
Next meeting
Tuesday, 2 September 2025.
Any other business
None heard