Re: [Myexperiment-discuss] Facebook application platform

[David De Roure]

Steve

One of the things we discussed in the scoping workshops was the extreme
case that myExperiment is "just" some some workflow-specific services
(executing, matching, provenance, ...) which can then be assembled with
all the other services out there in order to bring "myExperimentness"
through to whatever interface the users are already using - in fact one 
of our test scenarios is executing workflows from a wiki.  This is the
broader myExperiment vision (there's a great picture of this in our
powerpoints which I'll put on the front page of the Wiki later).

So we want people to be able to work that way - it's very much the web2
approach to use what's there as required. But for the JISC project we
needed to take a focused cut through the myExperiment space, and we
decided that building myExperiment.org was the right thing for a number of
reasons - such as giving us a VRE "product", having a space we can control
for privacy purposes and experimentation (the myExperiment experiment!),
providing something for people who haven't already got an established
solution, giving a focus for community building and raising awareness 
(and because we'd already said this is what we were going to do!)

The way myExperiment is constructed draws heavily on existing code and
services, so we have followed the model - we haven't built everything from
scratch - but nor have we used facebook yet either (other than to create
the myExperiment group on it!)  We could do facebook the same way we have
looked at twiki etc, but we're not resourced to do every interface - as
Matt says this will happen in response to demand and hopefully by
community contribution. Facebook is a good example but there are many
other sites that people might want to integrate in some way, such as
linkedin, and for other pieces of the puzzle look at freebase etc.

PS You can be my friend :-)

Thanks

-- Dave

On Sun, 22 Jul 2007, Steve Pettifer wrote:

> One thing I suppose we should be considering, or at least have a good  
> answer somewhere obvious as to why-not, is the facebook application  
> developer platform.
> 
> http://developers.facebook.com/get_started.php
> 
> Just in case someone hasn't come across it, facebook is yet-another  
> social networking site. However, unlike myspace, it seems reasonably  
> well controlled, and is a bit more 'serious/professional' (i.e. it's  
> not covered in insane textures, and doesn't play Your Favourite Mp3  
> as soon as you look at someone's page -- though i've just noticed  
> someone has added an application that allows you to throw pies at  
> your friends, so maybe not that serious...). It seems slightly more  
> aimed at adults rather than just teenagers, but isn't quite as dull  
> and specific as something like linkedin.com.  It also has a _lot_ of  
> good functionality on the social side (and there are a surprising  
> number of manchester CS people on there already!), and in particular  
> what seems like a very sane API for adding in one's own applications  
> for sharing stuff. The API is there to allow you to embed your own  
> services in the facebook pages; storing and retrieving data can be  
> from any server / system you like I believe.
> 
> So the question is... Why are we going to the effort of building all  
> the social networking side for myExperiment, rather than exploiting  
> what's already there in facebook (or other) and adding in the  
> scientific part. The obvious advantage of the facebook approach would  
> be that there are already a large number of users, and we'd get a  
> whole bunch of stuff for free. Obvious disadvantages are it's not as  
> well branded as being 'us' (we'd become an application within an  
> existing social networking system), and there may well be practical/ 
> technological reasons why we can't do it (though at a first few- 
> minute scan of the architecture, i'm not sure there's anything too  
> scary in there to stop us).
> 
> Apologies if this is a well rehearsed argument that i've missed...
> 
> Steve
> 
> 
> 
> 
> _______________________________________________
> Myexperiment-discuss mailing list
> address@hidden
> http://lists.nongnu.org/mailman/listinfo/myexperiment-discuss
>