Mitchell Kapor wrote:
Lisa (Dusseault) and I spoke to Ray at When 2.0 about SSE. SSE is defined at the level of abstract item replication; it's not a network access protocol and is mute on those issues. That is, it addresses how to replicate an RSS feed to have two or more instances which are kept in sync with each other.
Jeffrey and I both read over the microsoft spec on SSE after I posted about it... its a pretty good idea but, as you said, it leaves a lot of work up to the client.

Surprisingly, it seems like a pretty good solution for a sort of peer-to-peer network over http, if you imagine that a running chandler with an HTTP server could publish an RSS feed of a shared calendar, poll other RSS feeds for the same calendar, and publish back the user's changes in the hope that other clients would be polling it... The only problem being that peer-to-peer over HTTP is just nasty, with firewalls, changing IP addresses, and the like.

But it got me thinking anyway and possible solutions to try and leverage the idea, but in a way that would work better for peer-to-peer.. (for chandler 2.0 maybe :))

1) Allow chandler to publish to an external HTTP server over WebDAV, sftp, ftp, etc. (e.g. in my life I would make chandler publish to flett.org. In a university setting, I might post it to a common university-run web server. As a client, chandler would still be polling peer web sites. 2) RSS-over-jabber - instead of polling an http server, just subscribe to remote jabber clients and they could push out their RSS changes.

And of course, none of these three options (these last two plus the original idea of serving up RSS feeds directly from a running chandler over http) are exclusive of each other.

Alec


It detects conflicts but leaves it entirely up to the client to decide what to do. Typically, SSE would be implemented by publishing a feed to a server and handling subscriptions to that feed but in theory there could be P2P implementations. SSE permits feeds to have multiple writers, so the terms "publish" and "subscribe" are a bit misleading. SSE adds unique ID's for items to RSS feeds.

Because SSE was meant for different Microsoft applications and services such as Outlook and MSN to share calendar items, it has been assumed that each app would deal with all of the messy implementation details not defined in SSE itself. SSE itself is incomplete as a standard and doesn't address issues of what happens if two clients are trying to write to the feed simultaneously, for instance. So for SSE to be a useful public standard, these aspects would have to use an existing standard or borrow from one.

It would certainly be possible for a WebDAV server to be used to support SSE with a modest amount of work. Since RSS support is in the queue for Cosmo, it makes sense for us to follow SSE development carefully, and consider adding support if it looks like SSE is gaining momentum outside Microsoft.

Apologies for any lack of technical precision. I'm sure Lisa can amplify and correct as needed.



On Dec 8, 2005, at 8:59 AM, Alec Flett wrote:


I ran into this yesterday, thought it even more interesting than coverage of ourselves:
http://news.com.com/1606-2-5984715.html

Anyone know anything about this RSS+SSE for publishing and syncing calendars?

http://msdn.microsoft.com/xml/rss/sse/

Yeah yeah, its microsoft, but it seems like its something we can't ignore...

Alec

Philippe Bossut wrote:
In the (very) unlikely event some of you haven't seen it, there's a quite extensive video coverage of this conference (held yesterday at Stanford) on CNEt with words from Mitch on Chandler... :)
   http://news.com.com/1606-2-5985814.html

Cheers,
- Philippe
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

Open Source Applications Foundation "Dev" mailing list
http://lists.osafoundation.org/mailman/listinfo/dev

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

Open Source Applications Foundation "Dev" mailing list
http://lists.osafoundation.org/mailman/listinfo/dev

------------------------------------------------------------------------

_______________________________________________
Cosmo mailing list
[EMAIL PROTECTED]
http://lists.osafoundation.org/mailman/listinfo/cosmo

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

Open Source Applications Foundation "Dev" mailing list
http://lists.osafoundation.org/mailman/listinfo/dev

Reply via email to