Re: ATOM Pub vs. JSON vs. ...?

Kristian Rink wrote:
> 
> Well actually, despite knowing there actually always will be some 
> differences in "mapping" a custom data semantics to an existing one,
> I figured out that, in our given use case, the concepts of Atom and
> APP both fit better than any other approach I found so far, and, then
> again, seem to be backed by sufficient support (frameworks,
> documentation, tools out there) to eventually provide a reliable
> foundation being a "primary representation" in a prototype system
> that doesn't want to come up with a custom / proprietary API which
> essentially wouldn't do much more than re-inventing things like
> workspaces, feeds, entries, categories (one of the things about Atom
> that definitely got me enthusiastic about it when I started dealing
> with it) or representations accepted when POSTing things to a certain
> resource.
> 

Spoken like someone who understood Roy's thesis.  What's perfect may
not be what's best. And, what already has high cache retention provides
immediate benefit that's unlikely to diminish, which newer options
won't be able to equal until some nebulous point in the future.  The
bandwagon may have passed Atom by, but I've never heard of it being de-
allowed on intermediaries already configured for it; thus it is neither
obsolete, nor without its advantages.

-Eric