Meeting minutes
<mjoel> good morning/afternoon/evening
<norm> Hello Joel
<Steven> Previous minutes:
<Steven> Previous minutes: https://
<mjoel> AT&T fiber has been down for 5(!) days at my location, so I'm on a shaky tethered mobile internet stream
<mjoel> is there a backup phone bridge?
<cmsmcq> No one seems to know of a backup phone bridge.
Administrivia
<norm> My attempt at a pubrules-correct community draft is: https://
<cmsmcq> ACTION 2023-01-10-b - continued
<cmsmcq> ACTION 2023-01-10-c - continued
ACTION: 2023-04-11-a: done, at least a draft.
<cmsmcq> Everyone has 48 hours to point to errors; if none are reported, NDTW will send it to the publication team
<cmsmcq> ACTION 2023-04-11-c - partially done in June, but not yet discussed.
ACTION: 2023-05-09-d: done
ACTION: 2023-06-27-a: done (and turned into a PR by NDTW)
Status of implementations
<cmsmcq> 3.0 version of NineML is imminent.
<cmsmcq> The GLL parser is now consistently a little faster than the Earley parser.
<cmsmcq> Enumeration of trees in cases of ambiguity has now been implemented.
<cmsmcq> Renaming proposal implemented (but just found a bug).
<cmsmcq> Steven: renaming is half done.
Status of testing and test suites
<cmsmcq> Norm has requested that we merge a test on output ambiguity: see https://
ACTION: MSM to add ambiguity as discussion to agenda
ACTION: MSM to add discussion of wording proposal for renaming to agenda for next meeting
Technical issues
<cmsmcq> NDTW suggested that in order to use the new syntax for renaming, the grammar will need to be marked with a version number (e.g. 1.1)
<cmsmcq> Otherwise, a 1.0 processor may just interpret the grammar as syntactically incorrect and reject it.
<cmsmcq> Examples need to have a 1.1 version string.
<cmsmcq> But it may be that nothing further is needed.
Issue 176
<cmsmcq> This was closed in the last meeting and should be dropped from the agenda.
Issue 181 Non serializable characters in the input
<cmsmcq> Raised by John Cowan
<norm> Comment in here: invisibleXML/
<mjoel> EBCEDIC...now there's a name I haven't heard in...
<cmsmcq> We discussed John Cowan's suggestion that line endings be normalized.
<cmsmcq> Some sentiment that ixml doesn't now do any normalization, so this would be a change.
<cmsmcq> Some sentiment that we do have (implicit) normalization to Unicode -- or at least, a processor could without non-conformance handle non-Unicode input by normalizing to Unicode.
<cmsmcq> JL: the question is, where is our design center? Text input? or binary input?
<cmsmcq> JL thinks the design center is clearly text.
<cmsmcq> SP: the grammar doesn't now say anything at all about input, or changing it. We have a sequence of characters, and we do stuff with it.
<cmsmcq> SP: line-end normalization could be an option.
<cmsmcq> But it shouldn't be required.
<cmsmcq> BTW: an option would work, but end users will want the default to be normalization, not non-normalization.
<cmsmcq> SP: there are three options here.
<cmsmcq> 1 Leave it as is, and you get Unicode characters for CR etc.
<cmsmcq> 2 Processors SHOULD offer line-end normalization as an option, with default of no-normalization.
<cmsmcq> 3 Processors SHOULD offer non-normalization as an option, but with normalization as the default.
<cmsmcq> JL: one wrinkle is that really it's not the processor invocation but the grammar that needs to say "this grammar assumes normalized line-ends".
<cmsmcq> In that case, a run-time option is not quite the right thing.
ACTION: NDTW to make an issue for the question of line-end normalization.
C0 and C1
<norm> invisibleXML/
<cmsmcq> In the current spec, C0 and C1 control characters are not allowed in literal strings.
<cmsmcq> Some implementations of ixml treat 7F as a member of C1, some don't.
<cmsmcq> We seem, after discussion, to have consensus that 7F should be covered by this rule (whether it's part of C1 or not).
<cmsmcq> SP notes that the Unicode Database classes U+007F as CC (control character, other)
<cmsmcq> It may be better to recast the rule in terms of the Cc character class.
<cmsmcq> The range of that class is 00-1F, 7F, and 80-9F.
ACTION: Norm to make PR for recasting the rule.
Next meeting
<cmsmcq> We will meet again on 8 August.
<cmsmcq> JL suggests that we try to spend some time on grammar combination at the next meeting.
ACTION: JL to redraft his notes on grammar combination and send something to the public list/
<cmsmcq> Adjourned at xx:52.
<cmsmcq> s/ACTION: 2023/ACTION 2023/
<cmsmcq> s/ACTION: 2023/ACTION 2023/g
<cmsmcq> s/ACTION: 2023/ACTION 2023/
<cmsmcq> s/ACTION: 2023/ACTION 2023/
<cmsmcq> s/ACTION: 2023/ACTION 2023/