Minutes ixml call 2020-11-26

[ACTION: Steven to start a github org and admins]
[ACTION: Tom to set up github pages]
[ACTION: Tom to point the DNS to the github]



Agenda 2020-11-26


Previous Actions
Namespace
Specification
Implementations
Test Suite
AOB
Next call

Present: J, M, S, T.

Previous Actions

ACTION: Steven to specify what happens when a name isn't an XML name
[Continues]
ACTION: Tom to look for an IXML domain.
[Done: invisiblexml.org]
ACTION: Steven to upload the spec to github
[Done]
ACTION: Steven to research where to put S for attributes.
[Continues]

Namespace
S: So now we use invisiblexml.org
M: We should have a document at the location; so /NS and a year
S: I would prefer no date in the initial version
T: Wherever...
S: It's nice to have something you can remember

Specification
T: https://github.com/spemberton/ixml
T: Maybe we should start a github org, so that we are all admins
S: Easy?
T: Yes
[ACTION: Steven to start a github org and admins]
[ACTION: Tom to set up github pages]
[ACTION: Tom to point the DNS to the github]
[Discussion of what to point to]
S: The root should be an intro to ixml, not the spec

Implementations
S: I plan to convert my ABC implementation to C.
T: Since Declarative Amsterdam I haven't done anything more on mine. But 
since we last spoke I have. Still no complete implementation. Also very 
slow. Needs thought and work. Rather than returning nodes, and using 
recursive descent, we now return a map, so that everything only needs to be 
traversed once. I think that's an improvement.
J: Happy to help. I can run it and give feedback.
T: I'd love feedback.
T: I don't want to do several parses [In the context of expanding 
'parameter entities']
S: Earley does all parses in parallel. It may be not be until the last 
character you know which parse is the right one.
M: Even for non-ambiguous grammars.
S: Right.
T: I'm not trying to do macro expansion, but make a parser than can be 
extended to do macro expansion.
M: I'm trying to represent Earley declaratively, without a reference to 
"this moment".
S: What is your actual problem?
T: Using ixml to parse DTDs, which have an additional processing step of 
entitiy expansion. Like macros in C.
S: You can have an initial processing step in C that expands the macros 
before it gets to the parser.
T: You can have nested entities in DTDs.
T: My implementation pre-expands repetition rules

Test Suite
S: Michael you sent us a long screed on generating tests; it looked like 
the underlying issue was how to guarantee coverage.
M: The problem is also generating negative tests that aren't positive in a 
different way. I was hoping to have a demo ready. Alas.

AOB
[None]