On 20.11.18 17:10, Georg Lukas wrote: > * Florian Schmaus <[email protected]> [2018-11-20 16:44]: >>> -1, but we should add the Note about examples not being normative. >> >> The primary purpose of the (unwritten) rule that examples are considered >> non-normative is that broken examples cannot become or seen as correct >> behaviour as of the specification. Your suggestion would add another use >> case to the rule and explicitly encourage broken examples. I fail to see >> why we would want to do that. > > I'm not sure what your point is. I'm against changing the wire protocol > of a Final XEP to make examples in other XEPs less wrong.
It is not only about making examples less wrong. The PR is merely about documenting the current state. Which shows that the requirement is unnecessary. And also weakens the "changing the wire protocol of a Final XEP" argument (I would even say it invalidates it). The fact that even experienced XEP authors get it wrong in their examples shows that we should simply allow it. Requiring the 'disco#info' feature in a 'disco#info' query is just bad design. If we were to write the XEP again from scratch then we obviously change it, but since it is Final it appears sensible to make this simple and minor adjustment to XEP-0030. > I think that > we need to add a note into the examples about their non-normativeness, Which would create a precedent that intentionally broken examples are fine. That can not be what we want. > but I'm equally okay with fixing all examples instead. Even if you would fix all examples, the mistake would surely repeat again and again in future (Proto)XEPs. Then why not make this small non-intrusive, documenting what is done is the wild already, change to XEP-0030 instead? - Florian
signature.asc
Description: OpenPGP digital signature
_______________________________________________ Standards mailing list Info: https://mail.jabber.org/mailman/listinfo/standards Unsubscribe: [email protected] _______________________________________________
