i always thought that going from the last IETF draft to an RFC means there are no changes whatsoever, but i was wrong. the Atom Publishing Protocol has now been published as RFC 5032 and changed names from APP
to AtomPub
. so now there finally is an official document with an official name and an official abbreviation.
apart from the naming issue, the much more important thing is that AtomPub has now been officially published, which means i have no longer to apologize for championing some obscure draft-status protocol. i honestly believe that AtomPub will change the way how a lot of things will be done on the web. It is the first general-purpose RESTful protocol that is flexible enough to be usable in a large number of application scenarios, but still is simple enough so that people can understand it rather easily. many things on the web will be atomized sooner or later.
for example, all the social networking apps could (if they really were about social networking and not just amassing users to better serve
them with highly targeted ads) implement Atom and AtomPub interfaces, and then people could seamlessly mix and match stuff from different sites. this of course violates the business model of many of these sites, but maybe they have to become more open and thus have to support Atom at some point in time.
maybe in the same way as people came up with ways to integrate advertising into videos (distracting as trailers or less distracting as overlays), there will be AdAtom
, where the distracting way of advertising in Atom feeds is one ad entry for every 5 useful entries and the less distracting way would be a standardized way of embedding ad info into all entries whatever happens, there will be a lot of buzz around Atom and AtomPub.
I hope this means there's going to be stable, robust library support forthcoming in most programming languages?
And hopefully not one of these "some code in a servlet, sitting inside a ponderous Java application server" type of deal.
Posted by: yiming | Sunday, October 14, 2007 at 03:16
with more stuff on the way (https://datatracker.ietf.org/drafts/?filename=atom), built-in support for basic atom and atompub and some extensions definitely would make sense. i hope it will happen.
my plan is to start working on an XSLT implementation of all of that pretty soon. it will be limited (because XSLT can only do HTTP GET), but should be useful anyway as long as read-only access is required.
Posted by: dret | Sunday, October 14, 2007 at 04:01
Thank you for your post dude. This article is very helpful. I am waiting for new ones..
Posted by: gelinlik | Tuesday, October 16, 2007 at 14:45