?

Log in

Previous Entry | Next Entry

OpenSocial at Barcamp Berlin

Yesterday at BarCamp Berlin, Lukas Rosenstock, Christian Scholz, and I hosted a session on the social graph and OpenSocial. It was a very full room (~60 people) where we mainly led a discussion around OpenSocial. We started out by giving background on why opening the social graph is becoming so important (ala our Web 2.0 Summit presentation) and then Christian presented on OpenSocial itself.

By the end of the hour session, we had generated a list of questions about OpenSocial which we all were able to answer to varying degrees. I volunteered to blog these questions as the entire group in the hope that Google and others will answer them. If nothing else, I'll track down Patrick Chanezon when he speaks about OpenSocial at Web 2.0 Expo Berlin in a few days.

All the questions asked:

  • What is Google's role in OpenSocial? Is OpenSocial actually open, or does the community need to build "OpenOpenSocial"?

  • What does it take to make my site OpenSocial friendly? How do things like Microformats and FOAF relate?

  • How can I support OpenSocial APIs (e.g. get friends) without allowing the applications to be run directly on my site?

  • How does or will OpenSocial address social network portability?

  • How will OpenSocial application management evolve. Will it really be "write once, run everywhere" or will different containers impose policies that cause an application to still need container specific programming. For example, container A gives 5mb of data storage whereas container B only gives 100kb.

  • How will OpenSocial impact spam and abuse? Will it become more prevalent in a distributed fashion?

  • How does data storage work? Can an application share data between containers?

  • How does OpenSocial address internationalization issues?

  • What Google products is OpenSocial going to create? Why did they do it beyond just disrupting Facebook?

  • How does OpenSocial work with mobile?

  • What business model changes will adoption of OpenSocial lead to?

Comments

( 1 comment — Leave a comment )
lindner
Nov. 6th, 2007 03:23 pm (UTC)
Answers
Some answers, from my somewhat insider role..

# What is Google's role in OpenSocial? Is OpenSocial actually open, or does the community need to build "OpenOpenSocial"?

Google is in the process of giving back to the internet. Google Gears was the first example of a give-back. This is another. Looking at the various disclaimers and copyright it appears that google will retain some kind of controlling role in how the system is implemented. However given their work on AtomPub etc it would be safe to say that they can move this into a standards based direction outside of google.

# What does it take to make my site OpenSocial friendly? How do things like Microformats and FOAF relate?

The OpenSocial data api, based on gdata, is yet another way to retrieve a list of friends. If you're exposing data via FOAF you should be able to support that format too. If you're editing your foaf by hand, well, I feel sorry for you.

Myself, I'd advocate doing both Foaf and OpenSocial. Foaf elements are bound to end up inside of OpenSocial's Person object. Right now container specific extension fields are not well defined. The community can take care of much of this..

# How can I support OpenSocial APIs (e.g. get friends) without allowing the applications to be run directly on my site?

Good question. If your site implements a container the code will run inside an iframe. At some point I suppose there will be containers that can link a OpenSocial data feed (the RESTful way of accessing data, not JS) to another site.

# How does or will OpenSocial address social network portability?

It doesn't directly. However with almost all sites supporting the protocol it will make life infinitely easier to access and sync data between sites. With OAuth that process becomes very standardized.

# How will OpenSocial application management evolve. Will it really be "write once, run everywhere" or will different containers impose policies that cause an application to still need container specific programming. For example, container A gives 5mb of data storage whereas container B only gives 100kb.

Each container chooses their own policy. Some may be wide open, others, less so.

# How will OpenSocial impact spam and abuse? Will it become more prevalent in a distributed fashion?

"App Spam" will likely be a problem. Facebook apps push the bar for spamminess and I can see that happen here too. Individual containers will likely implement throttling or other mechanisms.

# How does data storage work? Can an application share data between containers?

This is doubtful. If an app wants shared data it should implement it's own storage.

# How does OpenSocial address internationalization issues?

The Google Gadgets spec contains full support for i18n. It's one of the key reasons we got excited about it at hi5. Very few applications get translated, which is bad -- plus the fragmented market meant that many small developers fluent in our popular languages didn't develop. That's all changed now.

# What Google products is OpenSocial going to create? Why did they do it beyond just disrupting Facebook?

I'm sure sticking it to FB was one of the goals. However the larger goal is to make life easier to create social applications.

# How does OpenSocial work with mobile?

Mobile clients like the iPhone with a working JavaScript browser should be able to access existing apps. It remains to be seen if people develop native apps that use the Data API.

# What business model changes will adoption of OpenSocial lead to?

Hard to say. My crystal ball only says that this will have a big impact. It certainly opens up the possibilities for what a container can do, and who they can partner with.

For the general good it seems that if we can get apps running inside open source web products we'll all stand to benefit.
( 1 comment — Leave a comment )