I’ve just now started to look into the details of OpenSocial, so my knowledge on the subject is rather limited. An initial though is whether the standard API will evolve quick enough for the web. I’ve been quite a heavy user of standards within the Java and J2EE technology space and standards such as the ones developed for the interaction between portlets and portlet containers (API solutions that are on quite the same level as the OpenSocial API), such as WSRP and especially JSR 168, have had the enormous disadvantage of evolving very slowly. It’s great to have standard APIs, but if they do not cover what the developers need to do, it will be difficult for them to achieve wide deployment. Sure, OpenSocial will get lots of attention, but unless it’s able to evolve with the requirements of the users (aka developers), it just wont work.
When I hear statements such as “OpenSocial is great because it has support from lots of companies”, I can’t help but worry that the fact that there are many companies with different interests could result in rigidness. The fact that OpenSocial is not in the “enterprise space” such as its the case for standards such as JSR 168, might mean that there is more flexibility. I sure hope so.
Update: Amazing to read statements such as this one from Hi5 CEO Ramu Yalamanchi: “At the moment we’re still in the integration phase. Essentially we’re waiting for the green-light from Google on when we can turn things on and launch them to the general public…” (source). On one hand it’s cool that the companies joining the standard seem to be so coordinated, but on the other hand it seems a bit scary that the companies have to get Google’s approval for launching features. Maybe I’m just being paranoid here 
Update2: According to Russell Beattie (link), OpenSocial is neither Open nor Social, at least not yet. Google wants its partners to get a head start and therefore haven’t released all specs yet.
Update3: As mentioned, Russell informs us that the container (SPI) specs aren’t public yet. That basically means that you can start building apps for the announced OpenSocial partner platforms, but you cannot compete with those partners by also implemented the OpenSocial SPI, since the specs aren’t public yet.
Update4: Tim O’Reilly points to something that makes a lot of sense: it’s not that interesting that OpenSocial will allow users to have access to the same functionality on all social platforms, what users need is access to all their data from any platform.
I share Tim O’Reilly’s understanding of OpenSocial, which is that it’s basically a contract between apps and the platform. Yes, there is a standard way for apps to get data from the platform, but it doesn’t seem to cover how a remote app can get access to a platform’s (container’s) data. We might just have overlooked something?
XML feed
About the author
No Comments Yet