[Cosmo-dev] Re: cosmo atom publishing design doc outline

[Elliot Lee]

On 6/30/06, Brian Moseley <bcm at osafoundation.org> wrote:
> (taking this discussion onto cosmo-dev)
>
> On 6/21/06, Elliot Lee <elliot.lee at gmail.com> wrote:
> > I hope the formatting on this email comes through OK.
>
> i've added the content of your original message to the wiki,
> reorganizing the content a bit:
>
>  * <http://wiki.osafoundation.org/bin/view/Projects/CosmoFeedService>
> - entry point for pages discussing cosmo's feed support
>
>  * <http://wiki.osafoundation.org/bin/view/Projects/CosmoAtomProposal>
> - your original proposal
>
>  * <http://wiki.osafoundation.org/bin/view/Projects/CosmoFeedRequirements>
> - requirements for the feed service, including discussions of gdata
> and calatom, and a list of use cases for live clients and for
> import/export
>
>  * <http://wiki.osafoundation.org/bin/view/Projects/CosmoFeedDesign> -
> the stuff from the email i'm replying to, which we'll evolve through
> discussion into a real design doc

looks good, thanks.

>
> > GData support
> >
> > Exporting GData means creating a feed in the GData format, based on Atom.
> > Would we use Google's Java Client Library?
> >  http://code.google.com/apis/gdata/client-java.html
>
> we could potentially use the com.google.gdata.data package, which
> appears to be classes that describe the gdata model (for example,
> EventEntry). i don't think we'd want to, though, since that api is
> dependent on the client api. it should be sufficient to continue to
> use rome (or the apache library that is meant to be its replacement)
> and perhaps add wrappers to deal with the gdata xml extensions.

rome? https://rome.dev.java.net/

>
> > APIs: existing code to change
> >
> >  I think this might be implemented in the FeedServlet class, so a new class
> > isn't needed. These methods correspond with ones currently used for Atom.
>
> the existing feed code was essentially a proof of concept. now that we
> have a much larger set of requirements, we need to put more thought
> into the design of the feed service.
>
> let's concentrate for now on understanding the requirements and
> figuring out which ones we want you to address during your summer
> project. over the next few days i'll draw some pictures and write out
> some thoughts on how i think we need to approach this.

great. what kind of pictures do you mean?