Sociablecode

First, read this.

This admirable result is the product of a lot of conversation, both on IRC and certainly at the hi5 headquarters. Developers have been begging for some visibility into relative performance of applications, specifically requesting install numbers to be made public. Many containers take the easy way out and make their install numbers public. Some (sorry MySpace, we still love you) even rely on this questionable metric as the sole factor for ranking within the application directory.

Suhail brought up a very good point which is that public metrics provide developers a proxy for monitoring the app ecosystem. Spamming and other nefarious tactics become transparent when a community of developers has their finger on the pulse of install rates, relative pickup etc.

PL makes a strong case against simultaneously exposing the ranking metric(s) values and relative importance: developers will inevitably jump to optimize that/those metric(s), effectively throwing user experience to the side in order to arbitrage their position. This is a strong argument, and leads happily to the conversation we see started above by Zack in his blog post.

Developers need something to watch, and as Paul points out, will inevitably optimize their apps pursuant to the reward mechanism of the container. For MySpace developers, it is easy to watch installs all day, and live and die on how many people are using the app. But today, two days into the hi5 launch and with a sea of opportunity before us, what is the thought process invoked by the proposed ranking algorithm? Answer: focus on quality. I can have 10 users, but if I as a developer can optimize the experience of my - few - users, I have a shot at the system rewarding me handsomely.

The past two months have been hectic as well as wrought with unrealistic promises, high expectations by all parties, skepticism from outside the OpenSocial community. The whole Orkut fiasco was in part explained away by a (sudden) emphasis on fairness, and we’ve seen both MySpace and hi5 totally surpass a certain larger organization in their ability to implement this new technology as well as support an enthusiastic community of developers.

The emphasis should be squarely on building quality apps that add value to the host network. Without this principal, there is no reason for meganetworks to even ‘go there.’ OpenSocial excites big networks because it represents an opportunity for third-party development to affect their own bottom line positively. It’s the - now old - F8 philosophy: your apps generate pageviews, eyeballs, loyalty and more, so bring them on and keep all the money you can make with them. Quality is key here. Frankly I use Facebook less now after being barraged by notifications.

“Laura has sent you a banana! Click here to peel it!

Yeah… I’m all set.

Momentary sycophancy notwithstanding, another hi5 decision bears mentioning. As the network rolls out the brand-new app gallery to its users, hi5 is randomizing the gallery order until they have collected enough meaningful data to rank the apps intelligently. Fantastic.

Whatever hi5’s algorithm is, it will certainly change over time as the team receives feedback and more data with which to work. The direction is the right one, and only good can come of it.

-Christopher (markitecht)