Meeting minutes
Bethan: I do not consent to having my voice or likeness recorded for use on cloud-based LLMs.
Accept the minutes of the previous meeting
Accepted.
Review of open actions
2024-03-05c Continues
2026-07-11-a Completed
Status reports
<Aljo> Norman: Some updates with nineml and adding options. Working on the C# transpilation of nineml and good news just today.
Norm: I've done a bunch of things, working on getting C# transpiling working
Steven: I presented something about designing a language so that you can use iXML on it.
… Creating a textual version of a doctype so that you can use iXML to create the documents.
… Declarative Amsterdam has a couple of iXML talks in the works.
… I'm doing a mini-tutorial. I'm working out a way to design a grammar so that you don't get tripped up by the problems of spacing causing ambiguity.
… Will submit the paper to XML.com as well.
Steven: One of the papers accepted was from the students at Delpht.
… They've moved to Codeberg as well. They are preparing a release.
… They may announce it formally at Declarative Amsterdam.
Bethan: Do we know how they're doing against the test suite?
Steven: 80% was the last number I saw.
Symposium on iXML
Steven: All we really need to settle on are the dates.
It was 26/27 February last time.
Steven: So shall we go with the last Thursday and Friday, 25/26 February?
Steven: There's no hurry, we're a little earlier than last year.
John: Are we going to change the title?
Steven: That's my proposal, but we don't have to decide that now.
ACTION: Steven to prepare an initial call for presentations.
Steven: Announce it at the beginning of October.
… Schedule as before.
Bethan: Would it make sense to reuse the CfP from last time?
Steven: Yep.
ACTION: Norm to confirm that submit@invisiblexml.org still works.
Bethan: I also think we should spread the load for presenting speakers and such.
Steven: Definitely.
Steven: We should also try to spread the load of making the video transcriptions.
Progress report on symposium output?
Bethan: Two to go and I have the intro sequence worked out.
Some discussion of the work involved.
Bethan: It's okay for English voices, but it has real trouble with speakers with stronger accents.
Some discussion of the accessibility aspects.
David: I may or not be the only person in this company who has impaired hearing; but I wonder if folks like me already have tools that can do it.
Bethan: Maybe, but it's not going to be prepared for a Dutch person speaking English and using technical terms.
… That may be good enough for a meeting, but it's not going to be good for presentations.
Some discussion of the distinction between auto transcriptioning and proper subtitling.
Bethan: We want to avoid the 'disability dongle' which you hear discussed in the disability community. It's not a great attitude.
David: I understand.
… The question was only about the value that is added by our doing the subtitling but I concur that value comes from our correcting things.
Bethan: I think the other thing that it adds is that it is a sign that we are willing to do that work. I do believe there's significant value add, but that's not the only value.
Norm: I don't think generating a transcript would be significantly less work anyway
New issues
+ Update the test-catalog to include the source files
See https://
See https://
John: I'm happy with that.
Accepted.
+ Replace github links with equivalent codeberg links
See https://
Accepted.
John: Do we have a plan to remove the github.
Norm: I've made it say "we've moved"
+ A simple grammar that matches UK postcodes
See https://
Accepted.
ACTION: Norm to update the example grammars section of the homepage
John: Can we have a disclaimer?
Norm: I'd have thought that was obvious from "sample" but I can take a look.
+ A sample grammar for Gregorian dates
See https://
Norm: I'd like to leave this until next time.
Steven: Could we make a grammar that identifies the day of the week? Something to ponder.
Open issues
+ #369 Namespaces in test catalogue still point to GitHub
https://
See https://
Norm: I've made a PR to change the namespaces; I think we should accept it. But it will be require implementors to update their drivers.
John: You only have to change a couple of places.
ACTION: Norm to announce this change on the mailing list.
+ #368 Add a symbol for "epsilon", don't allow absence of a symbol to be epsilon
https://
ACTION: Norm to start an email discussion.
Bethan: I think "" would be thing.
David: If I've understood, () already works. The issue is avoiding "nothing" to match, | |.
… So using "" would be two things.
Bethan: My concern about () is that semantically, we're still saying nothing matches epsilon, you just have to put parens around it.
Steven: It just makes the emptiness explicit.
Bethan: Depending on what you put in parens, you either have a sequence of things that is an alt (that is the sequence)
… Or you can have a bunch of things separated by ; or | (those are in fact alternatives)
… Or it can be empty.
John: Can I give an example? You can use () but if you have (()), is it the same thing?
Norm: I think works out to be the same.
Bethan: So parens can contain epsilon or a sequence or an alternative. But you can put () in side to put epsilon.
Steven: I'm not grokking the problem. I think then that this is something we should take to email.
Bethan: I agree, but I think if we don't hash out what the technical possibilities are, before we have that discussion more generally, we're going to spend a lot of tome explaining what is technically impossible.
… It would be better to work out what would be technically possible.
Norm: I think () or "" could work. What else could work?
Bethan: I think my concern is that an alts must contain at least one alt, but an alt contains 0 or more terms. You can have an empty alt, but you can't have an empty alts.
Steven: You'd make that term++ so that you had to have at least one.
… Or you could say that an alt is either a term++ or an epsilon.
Bethan: But if epsilon is going to be (), that's an alts.
Steven: So alt would be term++ or epsilon.
Bethan: Right. But what is epsilon? If we say it's () then that's an alts
… Are we saying that epsilon is an alts?
Bethan: So what is the alt that's inside the ()?
Steven: I think it would only match for an alt?
ACTION: Steven to send the rules he envisions for this solution to the list.
+ #334 Move iXML off GitHub?
https://
Status: primary movement
Finished. Close the issue?
Accepted.
Next meeting
Tentatively: 1 September. Steven may or may not have access.
Any other business
Adjourned