XSF Discussion - 2022-09-09

Does anyone have thoughts on whether the protocol specification for enabling XEP-0198 as part of SASL2 negotiation should be part of XEP-0198, part of XEP-0388 (SASL2) or a new XEP?

Definitely not part of 388.

Also worth considering is that this same question applies to SASL2 w/ binding

In fact all three are pretty closely linked. If we decided to put the 198 stuff in a new XEP, there is also the option of definining binding in the same new XEP

and the existing Bind 2 XEP factors into this somewhere

e.g. we could "upgrade" XEP-0386 (Bind 2) to "Instant Session Establishment" or whatever

Imho the only reason for not putting it into 198 is that 198 is stable. So a new xep for the sasl extension. Bind 2 is not stable so it can define the sasl extension directly ✏

Though it would probably help discovery if 198 could simply define the sasl extension itself

At a minimum we could make a small "editorial" update to XEP-0198 to link to the SASL2 extension, wherever that goes

The joys of modular specifications!

Doesn't "stable" mean it can still be changed anyway? Not that it probably matters a lot in this case

Yes I think technically we would be allowed to add it directly to 198

It's just a question if we would consider this good practice

Sure, needs a council vote or somesuch to change.

Should anyone want to implement the xep as written now without being aware of this change? Is that a reasonable thing for them to do 5 years from now? If yes probably don't edit, if no my gut says edit

In the absence of any strong feelings, I think my current plan would be: 1) update XEP-0386 Bind 2 with details of the SASL2 integration, 2) a new XEP detailing the SASL2+SM+Bind2 integration, 3) Add a link from XEP-0198 to the new XEP

The new XEP will in turn link back to XEP-0198 for the details of acking and such (only the initial negotiation is replaced)