On 6/12/06, James M Snell <[EMAIL PROTECTED]> wrote:

PME5 is silent on the cardinality of edit-media because I didn't think
there was any reason to specify it.  We have use cases where multiple
edit-media links are valid and very useful.  Most commonly, however,
there will likely only ever be one.  Some statement that an entry MAY
contain multiple edit-media links would be appropriate, but I'm not
convinced it's necessary.

0 to 1 is fine.
0 to many is fine.

Remaining silent on the subject has raised questions
and the Pace isn't even folded into the spec yet, which is not fine.

 -joe



- James

Joe Gregorio wrote:
>
> On 6/12/06, Andreas Sewe <[EMAIL PROTECTED]> wrote:
>> OK, so far so good. But what about the case where both the "image/png"
>> and "image/gif" version are supposed to be editable? (The server might
>> e.g., synchronize both versions, after an edit, by automatic conversion
>> again.)
>>
>> In this case the entry has multiple "edit-media" links which do point to
>> different resources (at least to different URIs, namely
>> <http://example.org/media/1.png> and <http://example.org/media/1.gif>):
>
> Yeah, that's a good catch, PaceMediaEntries5 is silent on the cardinality
> of 'edit-media' links and that needs to be fixed. I really don't care if we
> say there can be at most one or if we allow more than one, but
> PaceMediaEntries5
> does need to be clarified on that point.
>
>> Here RFC 4287's definition of "alternate" (every editable media resource
>> is also an alternate version of the entry) unfortunately provides no
>> hints as to the validity of the above, since its definition uses neither
>> the term "resource" nor the term "representation"; it just talks about
>> "versions".
>
> Yes, it appears that in some bizarre attempt to avoid WebArch
> nomeclature some ambiguity is present in RFC 4287. I'm just shocked.
>
>> At any rate, is the above extended example valid with respect to
>> PaceMediaEntries5?
>
> Good question, right now it's ambiguous and that obviously needs to be
> resolved.
>
>  Thanks,
>   -joe
>




--
Joe Gregorio        http://bitworking.org

Reply via email to