Driving adotpion of activitystrea.ms spec

[Monica Keller]

Hi im back home from TechCrunch's Stream CrunchUp Event.
http://www.techcrunch.com/2009/07/10/crunchup-live-real-time-business/
I was expecting it to cover activity streams more but instead it was
very Twitter/Twitter Client focused

The best thing I got from it was the opportunity to meet some of the
people who read this list and discuss their adoption of the spec

Eric Marcoullier from Gnip and Bret Taylor from FriendFeed seemed
interested. I hope so ! The main thing discussed was the need to have
the spec be non draft so that companies can feel confident writing
some code against it.

Martin has added the remaining items here
http://code.google.com/p/activitystreams/issues/list

How do we feel about completing these items for a v1 of the spec ?

Please take a look at the specs on
http://www.activitystrea.ms/
If you have not done so in a while.

Thanks

[Martin Atkins]

The problem with taking off the "Draft" moniker is that we won't be sure
it's right until we get implementer feedback. If implementers aren't
willing to implement until it's final -- a few companies excepted, of
course -- then that would reduce my confidence that we've actually
written a good spec.

I wonder if there's some intermediate state we could put parts of the
spec in, such as "Implementer's Draft", whose purpose is to encourage
experimental implementations and solicit feedback.

Of course not all companies are interested in participating in the
project and just want to implement it when it's done. That's fine, but
right now we only have two producer implementations and (I think?) one
consumer implementation that can only currently consume Facebook's feed,
so I think we're a little way off where we'd ideally be going into a
final specification.

[Ben Metcalfe]

On behalf of a potential consumer implementor who hasn't implemented yet (Seesmic, for Seesmic Desktop, Seesmic Web, Twhirl and other projects in the pipeline) the main feedback I can give is that it's not the "DRAFT" status of the spec that is causing us to hang back but that it's not implemented much on the Producer side - just MySpace and Facebook (and for the latter we heavily use FQL instead because of the amount of transactions we have to make, as we can batch them into FQL to reduce the number of requests being made over HTTP).

I know we'd really see the value in finding out the roadmaps of other producer sites to indicate when we can prioritize the implementation.  I think it would be valuable to see that information disclosed "in the community" but I'd also be interested in talking to producers directly - hit me up ben [at] seesmic [dot] com.

Otherwise there are many other competing tasks for developer time - same old story I guess.

Ben

on behalf of Seesmic

--
Ben Metcalfe (aka @dotBen)

Follow me:
/ Twitter - http://twitter.com/dotben
/ Facebook - http://facebook.com/benmetcalfe
/ Blog - http://benmetcalfe.com/blog/

ma...@benmetcalfe.com | San Francisco, CA

[David Cramer]

Over at Lifestrm we plan to include it within our initial release. However, I can't quite say when that's going to be. We're both supporting exporting to the Activity Streams feeds (via API), as well as importing your activities via a feed.

I do agree though, that the constant evolving state is an issue with us as well, and we're a small group of people with an ever changing project. I also believe that there's a lot of inconsistency in what people expect from the format, and what's offered as well.

----
David Cramer

[Rob Dolin]

Windows Live consumes the Activity Streams data from Facebook for the Facebook Web Activity (http://profile.live.com/WebActivities/) that users can add to their Windows Live.

 

I’ve been optimistic that Activity Streams would be consistently implemented by many partners so that we could write one set of code/config for enabling partner services that implement Activity Streams as Web Activities. 

 

It would be great if there was a locked early version of the list of verbs and object-types.  (Maybe call it v0.8 or something similar to indicate it’s not final.)

 

Thanks—

--Rob

[Martin Atkins]

Rob Dolin wrote:
>
> It would be great if there was a locked early version of the list of
> verbs and object-types. (Maybe call it v0.8 or something similar to
> indicate it’s not final.)
>

I think we're pretty close to being able to provisionally mark as final
the syntax spec and the core section of the schema spec. I think there's
still lots of research and design to be done for many of the other
schema sections, though.

[Monica Keller]

Yes if we can focus on finalizing the syntax spec and the core section
of the schema spec then implementors will not have to worry about
changing. I have had to change our implementation 3 times so far. So
my question is if anyone thinks we are missing anything really
important that should be in the syntax or core schema spec.
And include into your considerations the list of issues I linked to
above.

Thanks