Google's OpenSocial initiative to simplify the creation and adaptation of applications for social-networking sites pursues a valuable goal, but its technology platform needs further improvement.
That's the consensus from several developers who have been testing the OpenSocial APIs (application programming interfaces) and the OpenSocial implementations, or "containers," of participating Web sites.
However, the technical bumps they have encountered, while annoying and frustrating, haven't prompted them to give up on OpenSocial. Instead, the developers remain hopeful that the project, announced almost four months ago, will continue to mature.
Chris McCormick, a games industry contractor based in Australia, has encountered "a few rough edges" when working with OpenSocial, especially bugs in the partner sites' containers, but is "pretty satisfied" with the project.
"The API is intelligently designed and seems to cover all bases quite comprehensively. It should be possible to do some really fun stuff with it," McCormick said via e-mail.
Meanwhile, Aakash Bapna, an information sciences student in Bangalore, has also run into technical issues. "Bugs, bugs and lots of bugs. There are lots of issues with OpenSocial specs as they are launched. You can't tell when your smoothly working application can break," he said via e-mail.
For Bapna, a big hole is the unavailability of the server-side REST (Representational State Transfer) API, which will allow applications to tap servers, something that Thiago Santos, a Brazilian developer of an upcoming application called Partyeah, also misses.
Like McCormick, Santos also has encountered many bugs in partner site containers. Santos would also like Google to do a better job of communicating changes and updates to OpenSocial components. Still, he's confident OpenSocial will get over its growing pains eventually. "I have no doubt that [OpenSocial's promise] will be fulfilled," Santos said.
That promise is to establish a standard application-development platform for social applications so developers don't have to remake an application for each social-networking site. While Facebook hasn't signed up for OpenSocial, other big social-networking sites have, like MySpace, Bebo and LinkedIn, as well as major enterprise software players like Oracle and Salesforce.com, which see emergence of social features within business applications.
With OpenSocial, developers will be able to build the core portions of social applications and then adapt them if necessary, with, they hope, minor tweaks and changes for specific sites.
"It's not 'write once, run everywhere.' It's more 'learn once and write everywhere.' You learn the OpenSocial model once. For most applications there will be a core of code that's common to all platforms," said Patrick Chanezon, developer advocate at Google.
Then it's likely that participating Web sites will make available to developers additional extensions in their OpenSocial containers, allowing developers to take advantage of specific features in their sites that aren't included in the standard, Chanezon said.
Developers don't seem worried that OpenSocial will splinter if partner sites add too many proprietary functions to their containers. "I think it should be reasonably easy to write apps that run on all social-networking sites that support OpenSocial without much modification," McCormick said. "The core of OpenSocial contains the most important parts of the social-networking experience ... Anything which does end up adding something drastically new and wonderful will more than likely become part of the standard anyway."