while i cannot dispute the utility of this proposal, i'd strongly
urge that the TC work towards a proposal which does not alter schema,
merely processing rules (as i think that is what i would imagine is
the challenge for some implementations). Proposing a lite version
which breaks compatibility with the present 2.0-cd would be a huge
mistake.
=peterd
On Dec 20, 2007, at 6:22 PM, Drummond Reed wrote:
> Gabe explained that the idea of "XRDS lite" was to do a short spec
> about how
> to use a reduced set of the full XRDS element/attribute set that
> does not
> require processors to understand Redirect or Ref processing and
> full-scale
> SEP selection. Such a profile would including marking the outer
> XRDS element
> and/or each XRD element within it it with an attribute that
> indicated to
> processors that this XRDS or XRD uses only this limited set of
> semantics.
>
> John asked if a simple profile that could only be processed by simple
> processors would invite problems in compatability with full XRDS
> processing.
> The feeling was that while that was true, it would not be much of a
> problem
> as long as the XRDS lite spec was a proper subset of XRDS, because
> then
> inverse would not be a problem, i.e., all full-scale XRDS
> processors will be
> able to handle any simple XRDS document or mixed simple/full XRDS
> document.
