Meeting minutes
https://
Prague
Norm: interesting talks.
Steven: 4 talks on ixml. Which was great.
… I liked how Ari used ixml to deconstruct VINs.
Review of agenda
Steven: No changes
Previous Actions
Michael: ACTION 2023-01-10-fhas some movement, but nothing to report yet
Steven: ACTION 2024-03-05-c continues
Steven: ACTION 2024-03-19-a Norm sent a message about this, so this action is now active, and Steven has to get W3C to do something
Michael: ACTION 2024-04-16-a Nothing to report
Norm: ACTION 2024-04-16-b, closed
Norm: ACTION 2024-04-16-c closed
Michael: I think the idea was to make the URL return something.
Norm: Cannot do.
Norm: ACTION 2024-04-30-a is completed, and just posted, let's discuss next time.
Bethan: ACTION 2024-04-30-b continues
Status reports
Steven: My paper at Prague provides plenty of work on my implementation, much of which is already working.
Bethan: My implementation is still on my to be done list.
Michael: I've been doing work on the algorithm I have for enumerating all the trees in a parse forest.
… Every tree has a numeral, and you can order them all, gives a total order, and you can get the next tree, even if infinite forest.
… I have figured out how to work out which numerals denote trees.
Norm: We need to talk. I have something similar.
… and this might speed it up.
Issue 250 In spec grammar, alias delimiter ">" should be hidden
Steven: You are right.
ACTION: Steven to correct
Norm: Bad we didn't notice it
ACTION: Norm to add tests for new spec grammar features
Issue #246 (linebreak tests needed) requires a volunteer
ACTION: Norm to write linebreak tests
Issue #234 Alternative syntaxes for prolog
Michael: Is this still an issue for you Steven
Steven: Yes
Norm: Since I've just posted something about this maybe we should delay.
Steven: My main problem is that the prolog now looks too much like a rule.
… If it started with *any* other character than a letter, I would be happy
… such as (
Michael: Let us pause this for two weeks.
Norm: Agree
Issue #252 Er, where are the authoritative versions of ixml.ixml and ixml.xml?
Michael: I found them by following the links from the spec
… they should be better advertised.
… and they should be in the repo
Norm: There is a reason that they aren't because they are generated from the repo
John: XPath has a way of dealing with this
<norm> https://
Norm: This is a different issue
Steven: We have a link, but we don't expose the URL of that link.
Michael: If at the front matter we provide links to resources, including the grammars, that would be useful
… I would also suggest that we add those pointers to the readme file in the repo
ACTION: MSM: Create a pull request for links to the grammar
Norm: Make it more like a W3C spec
ACTION: Norm make the spec more like a W3C report
Issue #249 A subtraction operator
John: I need to express this.
… Michael has a technique to do this, but it is expensive
… ... I have an implementation
… it needs more exploration.
Steven: I would like more detail about its generality wrt to Early and other parsers.
… a paper might be a good way
Michael: We have a choice operator for alternatives
… the meaning of a;b is a set of strings, the union of the two sets
… this is just a set subtraction operator
… in terms of an (Earley) parser, it requires a little more work. So a -| b parses against a and b, so I believe there is modest extra work in constructing the tree
… after parsing a;b you can create a tree that contains a or b
… with a -| b you'll look for a and b, there are 4 cases: a and no b gives an a tree, an a and a b, the rule doesn't match, no a and a b, not valid, similarly no a and no b.
Steven: So you need an a and no b.
Michael: Correct.
Michael: A topic of concern, if a and b are regular then a-|b is regular.
… but if either is context-free, it does affect decideability I fear.
AOB
Steven: Next call 25th
[ADJOURN]