[Standards] On making "Compliance Suite 20xx" a Non-XEP

Sam Whited sam at samwhited.com
Tue Feb 7 16:33:03 UTC 2017


On Tue, Feb 7, 2017 at 10:25 AM, Georg Lukas <georg at op-co.de> wrote:
> Can we make the "Compliance Suite" a stand-alone document that is not an
> XEP?

There is a lot of process around publishing and updating an XEP. It
requires discussion, approval from the council or board (once draft
status is reached), etc. I think this is good for the compliance
suites. It reduces the agility, but also means changes are well vetted
by the community (a wiki page or some other document may not do that).

I'd be all for a BCP style document that acts as a pointer to the
compliance suites, but I don't personally want to try and set up
infrastructure for maintaining those documents, or try to figure out
what the procedures look like for publishing them, etc. it sounds like
a lot of work for very little benefit to me.


> - the Compliance Suite should help new developers navigating the zoo of
>   XEPs, having it hidden inside that zoo is counterproductive.
> - there is only ever one Compliance Suite that matters at any given
>   point in time

We should be better about deprecating or obsoleting old compliance
suites, but once they're in that state they don't show up in the list.
I don't think this is a problem.


> - it would be great to have a stable link/identifier to spread to
>   developers and reference in documentation

The link is probably stable for at least a year (and it's "stable"
afterwards too, it just might also say "actually, go look at this one"
at the top).


> - developers couldn't just stamp "implements XEP-0375" on their release
>   to indicate a given level of compliance, they would rather need
>   to write "Implements Compliance Suite 2017"

This seems like a blocker to me; I was undecided about which way was
better before (and didn't really care either way), but this makes me
pretty convinced that a new number is better.

> - having this as a non-XEP might increase the maintenance burden or
>   reduce the "credibility" of the document

I agree.

I don't want to completely shut down the discussion, but I'm not sure
it's useful. While I'm writing these I'm just going to submit new
documents. If someone else wants to take over and figure out a
different way, I'm happy to let them work on it instead.

—Sam


More information about the Standards mailing list