[Scooby-dev] Re: [Cosmo-dev] VOTE: cosmo/scooby merge
mikeal at osafoundation.org
Tue Jul 25 10:10:32 PDT 2006
On Jul 25, 2006, at 9:49 AM, Matthew Eernisse wrote:
> Where does this idea come from that Web client won't be able to
> pull data from a remote Cosmo (or other server)? The ability to
> pull data from multiple sources is to me an absolutely fundamental
> requirement for the Web client -- merged or not.
> Even after a merge, the UI will still make HTTP requests to the
> backend and get the results back as a string of JSON. An HTTP
> client library on the server can easily proxy the client requests
> through to a remote Cosmo server, and responses back. Any server
> that can produce a valid string of JSON can serve as a data source
> for the Web client.
> We would still have the issue of authing the user on the remote
> Cosmo, and managing the list of remote servers for a user -- but
> that will be the same whether we merge the two projects or not.
This brings up another point I was thinking of.
Are we going to treat the new integrated JSON server interface as a
general interface for anyone to use and pull data from or are we just
going to build it as the backend service for the UI?
Since the interface is public I know the first response will be "of
course, it's public", but what I'm getting at is, are we going to
take the time to write some basic conventions and document the
service as an open interface for anyone to easily use? If we're just
thinking of it as the backend UI service then it doesn't sound like
we would prioritize that.
More information about the cosmo-dev