Tim Bray wrote:
> On Oct 16, 2005, at 6:43 AM, Bill de hÓra wrote:
> 
>>> Restating: I am definitely in favor of accomplishing this using
>>> something within the content of the <atom:entry> that you PUT,   because
>>> (a) I can look at the XML and know what's going on and (b) I   might
>>> want
>>> to ship self-contained updates around on non-HTTP channels.
>>
>>
>> I think this is akin to treating HTTP as a transport.
> 
> 
> Is this a problem? -Tim

There may be if the APP control part ends up smeared into the data part,
ie some control directives are in HTTP, some are in Atom entries. I'd
like to be sure it was neccessary to make things work.

If posible I would like the state machine for pushing entries around to
be reasoned about separately from what's in the content. Otherwise
there's two data dictionaries to look at (http + atom), two agents,
whose atom dictionares need to be merged or resolved in some implicit
way because we're trying to keep the number of requests down, plus
scenarios where the atom dictionary might need to be invalid at certain
stages.

I'm not coming at this from some webarch purity perspective btw; I'm
thinking operations. If others have implemented with control data in the
entry and say it's a non-problem I won't argue this any further; I don't
have any code using control data in atom entries to back up my concerns.

cheers
Bill

Reply via email to