yeah - that makes sense. don't worry, i'm not advocating any spec changes at all. within shindig though I think we should keep them == until we get all the basics down. no need to add more complexity before we are ready (and without anybody needing it). especially because external websites are using the existing interfaces.. we should try and keep the changes as minimal as possible.
once we get the easy apis done maybe a simpler contacts and friends interface for shindig will become clear too... - Cassie On Tue, Apr 29, 2008 at 8:14 PM, John Panzer <[EMAIL PROTECTED]> wrote: > FYI -- regarding people/john.doe/@all/jane.doe vs. people/john.doe/@self: > I > don't think the _protocol_ should assume these are always identical. That > assumes that social networks are never going to have private copies of > contact information, and I don't think that's a good assumption to make. > Even if Shindig just aliases people/john.doe/@all/jane.doe to > people/john.doe/@self (which is fine), clients should still be able to > specify which one they care about. > > Specific example: I may have the rights to change name.unstructued to > "Janey" in people/john.doe/@all/jane.doe if I am john.doe, but _not_ to > change name.unstructured in people/jane.doe/@self (Jane hates the nickname > "Janey"). I know that Shindig doesn't necessarily support this but it > could > be added behind the same interface. Shindig itself can just disallow such > games of course. > > > On Tue, Apr 29, 2008 at 4:53 AM, Cassie Doll (JIRA) <[EMAIL PROTECTED]> > wrote: > > > > > [ > > > https://issues.apache.org/jira/browse/SHINDIG-210?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12593007#action_12593007 > ] > > > > Cassie Doll commented on SHINDIG-210: > > ------------------------------------- > > > > So I am going to check this patch in almost as is. I have only made > > modifications to the service interfaces. The methods you added Vasu > didn't > > really match up with the style nor meaning of the other added methods so > I > > tried to reconcile things. The activity service is all fine, the methods > > have gadget tokens and return response items and the adapter is good. > > > > The people service on the other hand needs to be reconciled. The rest > apis > > for /uid/@all/pid don't make sense with our code. The services and sample > > container don't have any concept of "contacts". Because the interfaces > > should remain more stable I took this stuff out until we can figure out > the > > proper way to do things cleanly. For example, we should use the gadget > token > > properly. That is how we do all of our permission checking and data > > restriction. With the gadget token it almost seems like > > > > people/john.doe/@all/jane.doe == people/jane.doe/@self > > > > john.doe doesn't really come into play at all unless he happens to be the > > viewer (or owner) in the current context. > > So... we'll want to resolve that conceptual problem through email before > > we change the service apis. > > > > Also - that RestGuiceModule has a lot of bindings which should probably > > just move into the SocialApiGuiceModule. You had a todo here so you may > know > > that already. > > > > Thanks, and I look forward to more patches! > > > > - Cassie > > > > > Implement RESTful spec in Shindig > > > --------------------------------- > > > > > > Key: SHINDIG-210 > > > URL: https://issues.apache.org/jira/browse/SHINDIG-210 > > > Project: Shindig > > > Issue Type: New Feature > > > Components: RESTful API (Java) > > > Reporter: Bob Evans > > > Attachments: restful.patch > > > > > > > > > This is a placeholder for a patch to add RESTful services to access the > > opensocial data services. > > > > -- > > This message is automatically generated by JIRA. > > - > > You can reply to this email to add a comment to the issue online. > > > > >
