[calsify] Re: JSCalendar, iCalendar, and the calext charter

[Ricardo Signes <[e-mail]>]

On Tue, Apr 22, 2025, at 12:19, Michael Douglass wrote:
> I don't think there's much point in this work if you don't assume JSCalendar ultimately being used in mainstream calendaring. However, it's going to take time and maybe we can't expect the world to stand still while services get round to it.

I don't think anyone is saying the world should stand still.  We should, though, think about what it is that calext should be doing.

The charter says:
 • Any extensions to icalendar or jscalendar must include a representation in both formats, and define a robust mapping between them.
So, one option is that we agree with this, and we work this way:  at the very least we document existing incompatibilities, and preferably minimize them; and in new work, we require both formats.  We find extension drafts in flight that do not include both formats, and we require that they are updated to include robustly-mapped representations in both formats.

If we don't do that, the WG is shirking its responsibilities under the charter.  That's not "the world", it's us.

This problem is compounded by jsCalendar already having a fair bit of semantic mismatch with iCalendar, including in places that (as far as I know) there are *no* implementations.  So we're standing far down a path to a hoped-for outcome, while the world is moving along on its own trajectory, and it's not clear we're doing whatever labor is required to adjust both paths toward some happy future intersection.

It seems to me like we'd do well to draw an outline around those parts of the calendaring model that are really implemented and in use, and really compatible between formats, and say "This is the trustworthy core", and then carefully move that line outward through trial and implementation, which may sometimes result in having to change, discard, or clarify things that have not yet been really tested.

The question of "why would somebody use this" is, as you say, a really important one.  It will help us figure out how where to draw the line now and how to expand it later.  The clear case now is "JMAP for Calendars", which doesn't support the totality of jsCalendar.

I think that's a way to really engage with what the charter says.

The other option, of course, is to recharter, and to understand that we have two closely-related standards that are compatible where we have shown, with no further requirement to ensure compatibility.

-- 
rjbs