Re: SDshare push

* Lars Heuer
> 
> APP provides a complete protocol to edit/publish Atom resources and it
> remains unclear why the proposal does not use APP (maybe making
> several APP features like service documents optional). AFAIK it does
> not support POSTing of Atom feeds.

It doesn't seem to be a very good match for our needs, basically. The mismatches are:

 (1) Doesn't allow posting of complete feeds (as you note).

 (2) Semantics are different, in that APP expects a new fragment
     to be created, but that's not what we are doing at all.

 (3) None of what APP provides actually seems to be useful for our
     use case. Using APP doesn't make the SDshare push spec any
     simpler; in fact, quite the contrary.

> IMO it would be nice if SDShare push would be more aligned to existing
> solutions, especially to APP instead of creating another proprietary,
> NIH solution. SDShare push could simply use APP (POST) and done.

Well, SDshare push does use POST, but we post complete feeds for the reasons outlined in earlier discussion.

> Uploading a collection of fragments would result into multiple POST
> requests. Using multiple POST requests would also provide a better
> error handling etc.

It would also mean a much greater number of requests, etc.

In short, it's not clear to me that there's much benefit in reusing APP, although I agree that it is possible. If you disagree I would be interested to hear a more detailed rationale of what the benefits actually are.

--Lars M.
http://www.garshol.priv.no/tmphoto/
http://www.garshol.priv.no/blog/