Meeting minutes
minutes review
<kaz> Nov-17
Fady: transfer of chairs to Fady
… also met with McCool to discuss flow, tools, etc.
… did that offline
… also looked at PR197, merged even with error, which seems to be in the tool, not the data
Fady: no objections to publishing minutes?
… none, to be published
repo, wiki, and github reorg
Fady: there is a lot of old information hanging around
… up till now, TDs put in org's folder; but better if organize by implementation instead
… this aligns better with the report
McCool: generally speaking good to align plugfest contributions with test data reorg
McCool: still a few fixes needed to test data; also we need to define what an implementation is
… I can copy the current text from the IR into a README, make a PR, and we can discuss as part of PR review
Fady: as another point, we are discussing plugfest results using issues; but seems like it is a problem to solve
… and they never get closed
… was thinking we could migrate to the github wiki instead
… so instead of issues, we can have wiki pages
McCool: think about requirements; we would like to link to results for one project and for all projects for one plugfest
… also, everyone needs access to wiki
… or rather, just the group
… another option is just have a directory under the plugfest with an md file for each project
Fady: going to create an issue then to discuss
Kaz: can discuss permissions for wiki pages and should be able to generate a group-only wiki page, but only three levels, editor, member, public by default
McCool: concerned that non-members can't document projects using wiki; md files would allow PRs from non-members
… also suggest that once we have a new system we can copy all the content from the old issues and close them
Fady: next topic and I'd like to clean up the wiki
McCool: suggest we archive rather delete content
… should create a sub-page like in main wiki
… but definitely agree wiki needs to be cleaned up
Plugfest organization
McCool: need a date first
… but it seems we generally agree on Feb
… but need a doodle to nail it down
… then can create basic structure
Fady: right, will do that, also set up readme, etc.
profile testing
Lagally: would like to create profile spec that aligns with plugfest
McCool: in general I would say profiles are more restrictive than what is done in plugfest
… also we need to distinguish experimental vs. testing examples
Lagally: also there is a lot of out-of-band communication to resolve how TDs should be interpreted
Fady: how do we intend to test profiles?
Lagally: first, need a JSON schema or validator
… that would extend TD testing
… don't think we need a big machine; maybe some additional assertions
McCool: do think the existing tools will work, but we should extend the assertion tester to use different JSONSchemas, which we need anyway for TMs
… and then we can use manual assertions for corner cases that can't be automatically tested
Lagally: doing review of TD spec; current information model is not correct for TMs, 5.2
… for example, forms are mandatory in TDs but not TMs
McCool: summary of actions: PR on data directory to add impl desc; doodle for plugfest (kaz); start organizing event dir for Feb (fs); look at requirements for assertion tester to support profiles (fs)
Kaz: when should we aim to finalize?
McCool: probably two weeks from now, otherwise we get into xmas
… post doodle now, decide in Dec 14 main call
Fady: adjourn