Re: change proposal for issue-86, was: ISSUE-86 - atom-id-stability - Chairs Solicit Proposals

On 04/14/2010 05:25 PM, Ian Hickson wrote:
> On Wed, 14 Apr 2010, Sam Ruby wrote:
>> If there is a consensus to fix these and other bugs, then I would
>> support an Atom mapping remaining in the W3C HTML5 spec.
> I'm happy to fix real bugs, if they are reported. Bug 7806, however, has
> already been fixed to the extent possible in the HTML5 spec. What Julian
> escalated was not the original reported bug, which was in fact fixed; what
> he escalated was a request to say that if an implementation didn't conform
> to the Atom specification in one very specific case that is arguably not
> always possible to achieve, that implementation should _also_ be
> considered not conforming to the HTML specification. This seems to me to
> be idealistic language lawyering with no value.

Speaking again as only a member of the Atom community: given the 
importance that entry ids have in enabling user agents that process Atom 
feeds to match entries in the feed against what they have seen before, I 
personally believe that it is actively harmful for feeds to be produced 
by any software that can't guarantee that they produce the same ID given 
the same input.

Reading through the entire bug, and in particular comment #23, I think 
the spec is making the wrong tradeoff, and I suggest you reconsider the 
presumption.  I suggest that you actually test out how common feed 
aggregators react when they are presented with the same feed differing 
only in the entry ids.

FWIW, I don't see dereferenceable as mandatory.

- Sam Ruby

Received on Wednesday, 14 April 2010 21:59:52 UTC