My personal views about Facebook & OpenSocial platform
I was responding to a clarification on the Bangalore OCC group, and decided to make that a blog post.
Now that the Facebook platform is available, a network would consider both options.
While OpenSocial was quick to get off the ground ( its almost hacked up together ), it still has serious issues.
Will list out some of them:
* Functionality
# Virality - except Hi5, no container has implemented virality(inviting friends, notifications, myspace does not even have news feeds) - which is key to application growth
# Profile - This is a serious flaw in the design. In OpenSocial, profile views are fetched from the application servers. Not only does it burden the app provider considerably, it also puts at risk the performance & usability of the network itself. 1 slow application (possibly because of a bug) , can make all profiles that have it become very slow. I can execute javascript on profile pages, and can continue to show alert boxes ... etc
# Application management - This does not exist, except for Hi5. In Orkut, an application is tied to a URL, so I cant move to another URL. Moving to new servers, is painful. Taking downtime on a server is painful. Yes you can do DNS changes, but they take days to propagate. The basic idea of starting with small infrastructure costs & growing as the traffic grows, becomes impossible.
* Checks & Balances
- This is where Facebook has been very good, they have built powerful systems, that automatically regulate quotas based on the spaminess of your applications. OpenSocial networks have only relied on manual monitoring & correction, as of now, which is difficult to scale.
- Facebook's application metrics (and they have evolved) are an effort to promote more useful applications. Other networks will have to go through that entire exercise figuring this out themselves.
* Javascript Apis only
- Just because this is supported by Myspace & Google, it has actually taken off - but an api thats available only in javascript, would have been laughed at & ridiculed to no end in the dev. community. Whatever happened to web-services! You cant do any batch processing like group actions of a user together & send a concise & more meaningful message out. Now, its 1 message - 1 notification.
- If this continues to be the case, most developers will learn & pick up Javascript to get the distribution that G & Myspace provide, but in the longer term, thats not going to take them far. Javascript, as a language will probably gain a lot.
Eventually though, it wont really be a simple logical decision, there are much larger & longer term issues which will play a role in deciding which network goes which way.
p.s: These are my personal views, and I hope I get corrected by networks, platforms & people.
Now that the Facebook platform is available, a network would consider both options.
While OpenSocial was quick to get off the ground ( its almost hacked up together ), it still has serious issues.
Will list out some of them:
* Functionality
# Virality - except Hi5, no container has implemented virality(inviting friends, notifications, myspace does not even have news feeds) - which is key to application growth
# Profile - This is a serious flaw in the design. In OpenSocial, profile views are fetched from the application servers. Not only does it burden the app provider considerably, it also puts at risk the performance & usability of the network itself. 1 slow application (possibly because of a bug) , can make all profiles that have it become very slow. I can execute javascript on profile pages, and can continue to show alert boxes ... etc
# Application management - This does not exist, except for Hi5. In Orkut, an application is tied to a URL, so I cant move to another URL. Moving to new servers, is painful. Taking downtime on a server is painful. Yes you can do DNS changes, but they take days to propagate. The basic idea of starting with small infrastructure costs & growing as the traffic grows, becomes impossible.
* Checks & Balances
- This is where Facebook has been very good, they have built powerful systems, that automatically regulate quotas based on the spaminess of your applications. OpenSocial networks have only relied on manual monitoring & correction, as of now, which is difficult to scale.
- Facebook's application metrics (and they have evolved) are an effort to promote more useful applications. Other networks will have to go through that entire exercise figuring this out themselves.
* Javascript Apis only
- Just because this is supported by Myspace & Google, it has actually taken off - but an api thats available only in javascript, would have been laughed at & ridiculed to no end in the dev. community. Whatever happened to web-services! You cant do any batch processing like group actions of a user together & send a concise & more meaningful message out. Now, its 1 message - 1 notification.
- If this continues to be the case, most developers will learn & pick up Javascript to get the distribution that G & Myspace provide, but in the longer term, thats not going to take them far. Javascript, as a language will probably gain a lot.
Eventually though, it wont really be a simple logical decision, there are much larger & longer term issues which will play a role in deciding which network goes which way.
p.s: These are my personal views, and I hope I get corrected by networks, platforms & people.