[Standards] Entity Capabilities Woes

Ralph Meijer jabber.org at ralphm.ik.nu
Sun Mar 18 19:13:08 UTC 2007


On Sat, 2007-03-17 at 23:03 +1100, Daniel Noll wrote:
> On Saturday 17 March 2007 22:19, Olivier Goffart wrote:
> > Really, i think that XEP-0115 is good, I think that you just fail to
> > understand the exact meaning of the ext attribute.
> > let's quote the XEP (§4.1)
> >
> > | The names of the feature bundles MUST NOT be used for semantic purposes:
> > | they are merely identifiers that will be used in other use cases.
> 
> Nope, actually I understood that.  I got tripped up by the fact they they DO 
> have semantics, being that they don't change between versions of the same 
> client.  Originally I would have expected something like a "ftrans" extension 
> to add extra features over time.

This seems one of the root causes of all misunderstandings in several
XEPs we have. Somehow mentioning the word 'opaque' doesn't seem to ring
any bell with some people. So, again, the space separated sequences of
characters DO NOT have semantics by themselves. They are basically the
key to match up with a set of features.

Maybe we should have the examples in XEP-0115 use short numbers instead.
Or other strings that immediately raise the question: "Huh, what does
that particular ext string mean?'", only to be followed up with "Oh!
Nothing.".

-- 
Groetjes,

ralphm




More information about the Standards mailing list