abdera-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From James M Snell <jasn...@gmail.com>
Subject Re: Feedback on an APP implementation...
Date Wed, 11 Apr 2007 16:08:46 GMT
Correct me if I'm wrong, but you're essentially putting together a
content store, right?  To this point, Abdera has focused primarily on
providing a framework for building APP servers as opposed to providing
any specific server implementation.  I'm not sure Abdera is the right
place for the server impl but it could definitely do most of the heavy
lifting you need as far as the Atom protocol is concerned.  What might
be an interesting way forward would be to implement the server backend
using Abdera, then explore the creation of a new apache incubator project.

- James

J Aaron Farr wrote:
> Hello all.
> So it turns out I spent most of March reinventing the Atom Publishing
> Protocol.  My goal was to create a simple RESTful web service for RDF
> but the design I ended up with mirrors APP almost too well.  So after
> some debate, we (my 2-man startup) decided to adopt Atom.
> We intended to open source the server as an independent project but
> now that we've committed to Atom it may make more sense to collaborate
> with Abdera or a similar project.  Below I've included a simple
> description for our web service.  I've spent the last week and a half
> getting up to speed with Atom and I'm certainly still unaware of some
> developments.  Thus I'd appreciate any feedback, though
> specifically...
>  Is such a service within the scope of Abdera?  (I'm guessing yes)
>  If not, anyone interested in seeing something like this @ Apache?
>  Or, is anyone working on something similar?
> Thanks!
> -- 
>  jaaron
> The service defines extensions (more like conventions) to the Atom
> Publishing Protocol to support collection creation, entry versioning,
> user profiles and access control.
> The service follows the URL scheme:
>  http://example.org/service/{collection}/{resource}/v/{n}
> Which gives us URLs such as:
>  http://example.org/service/china-trip
>     A collection for data about some travel itinerary
>  http://example.org/service/china-trip/grand-hotel
>     An entry containing information about a hotel
>  http://example.org/service/china-trip/grand-hotel/v
>     Prior versions of the entry are available via a feed
>  http://example.org/service/china-trip/grand-hotel/v/3
>    The third version of the 'grand-hotel' entry
> The service defines a special collection 'root' whose entries are also
> collections.  To create a new collection, the client would post to the
> root collection.
> User profiles are Atom entries in a 'user' collection.  Each user
> account has an Atom service document which associates users with
> collections.
> Collections and entries are initially private.  Only to listed authors
> can get or update an entry.  We're working on an atom extension that
> would allow more fine grained access control.
> Our goal is to create a simple, easily accessible RESTful web service
> to be used as the backend for (web) applications or as the basis for a
> "REST application server" similar to what Dave Thomas recently
> described in his RADAR article [1] or as an open source alternative to
> services like Amazon S3.
> [1] http://pragdave.pragprog.com/pragdave/2007/03/the_radar_archi.html

View raw message