Open Community Working Meeting 2023-03-27 - 14:00 PT

[benjagm]

Open Community Working Meeting 2023-03-27 - 14:00 PT

📺 See Recording

⏪ Go To Previous Meeting

Agenda:

Topic	Owner	Decision/NextStep
Review last call's action items	[facilitator]
Reminder of PRs that require review	@benjagm	Test a github action to get notifications
Status of and next steps for https://github.com/orgs/json-schema-org/discussions/329 (Still Kinda Supporting Unknown Keywords - A call for proposals)	@gregsdennis	Accept the survey results and use x-. Close the discussion and communicate properly with blog post and social media campaign
GSoC update	@benjagm
Apply to Sovereign Tech Fund Related issue: #360	@benjagm	Identify the best projects to include in the proposal
<!--	[TOPIC] [IssuePRDiscussion]	[owner]	-->

Action items:

• [x] @benjagm to test the Github action to send reminders to review PRs.
• [ ] @gregsdennis to prepare the ADR once the discussion is closed.
• [ ] @benjagm to reach out to the academic and research contacts to check their interest on Sovereign Tech Fund. Giorgio Ghelli and Michael Mior.

Notes:

• @Relequestual , @gregsdennis , and @benjagm reviewed past action items and discussed pending pull requests that require feedback.
• The group discussed the possibility of setting up a GitHub Slack integration for reminders on pull requests waiting to be reviewed. @benjagm offered to test the GitHub action.
• @gregsdennis provided an update on #329 issue. The leading contender for a prefix, which was x-. The group discussed different ways to specify ignored keywords, including @sorinGFS proposals in the discussion #329.
• @jdesrosiers wanted to incorporate @gregsdennis recommendation into the spec in a different way, so @gregsdennis left the discussion open for him to review.
• Following the survey results, the group agreed on using the x- prefix, close the survey this weekend and mark the decision in the GitHub discussion. They also talked about the possibility of creating an ADR and blog post to communicate the decision internally and externally.
• @Julian , @jdesrosiers , and @benjagm discussed the progress of their Summer of Code proposals but we are conservative until getting the confirmation from Google regarding the acceptance of all Postman's slots.
• The group discussed the status of revamping the website.
• The group agreed on applying to Sovereing Tech Fund Program and discussed some potential projects to be considered like: Bowtie, the development of vocabularies, support academic research initiatives and C++ implementation.
• We agreed on continuing the discussion on the issue #360.
• We discussed about meeting in person during the upcoming Open Source Summit North America.

Attendees

Account
@gregsdennis
@jdesrosiers
@Relequestual
@benjagm
@Julian

[jviotti]

I'm traveling and sadly won't be able to join, but looking forward to the outcome of the Sovereign Tech Fund discussion :)

[awwright]

Accept the survey results and use x-. Close the discussion and communicate properly with blog post and social media campaign

I saw this from https://github.com/json-schema-org/json-schema-spec/pull/1387

Can you please post the poll results? The results may be interesting, but last meeting I saw there wasn't a clear consensus; and most of the respondents might be comfortable with any outcome. Some might not be aware of the meaning of x- or fully understand why this is different, but could be persuaded with some explanation, it's difficult to tell.

I'm not terribly picky about prefix, but I do object to x- because usually means "experimental" and/or "user/private use" but that's not what's going on here, it's not a naming scheme to manage how we allocate keyword names, it's supposed to describe the behavior of the keyword to validators that might not know anything about its validation behavior. Specifically that it doesn't perform validation (or doesn't reject).

[benjagm]

Can you please post the poll results?

So far the results are:

[gregsdennis]

I do object to x- because usually means "experimental" and/or "user/private use" but that's not what's going on here

We actually talked a bit about this. x- historically for HTTP has been problematic because of the experimental-to-production issue, but

1. people had come to take it as, "I can add whatever I want as long as it stays with x-," which is what we're going for
2. Our use is explicitly defined.

I didn't like it at first, either, but thinking of it this way makes sense to me.

The current tally is posted in the collaborators channel in slack. @benjagm is going to leave the survey open for one more week, then post the results publicly. At that point, I'll be updating the discussion with the decision and locking it, and writing an ADR (for us) and a blog post (for the public).

[Relequestual]

In terms of the prefix for annotation only values, critically, what we're signalling to schema authors here is, "this isn't going to be interoperable unless you have specific agreement". Now, any implementation may specifically say it will support any specific x- prefix keyword, and may, based on configuration (if it wants to be compliant), make additional validation. You could argue, "it's not supposed to do that". Well, no. Legally, it would need to do a second pass using those annotations... but it may simply "cut out the middle man" on that.

We signal, this is an annotation only [as far as compliant implementations should be concerned], and let others handle whatever additional things they want. Ideally, using the vocabulary system for usecases which require n above 1.

There is no question that we need better docs for how to use vocabularies. But, we also need to better quantify what interfaces we would like to see from implementers, and provide a way for implementers to signal such interfaces exist. (I'm working on that.)