Ask HN: Has the GMail graph been published?

9 points by albertcardona 16y ago ↗ HN

When GMail launched, it was invitation-only, and there was a limited amount of invites to give away. My immediate thought was that Google was building a graph of all users, and that the nodes were being kept artificially small in purpose for some reason.

The other immediate thought was the prevention of spam agents using GMail for their purposes.

Sampling users a few at a time, starting with trusted users, also enabled the creation of a profile of a "good user", to further prevent spammers from using GMail--their profile would be very different.

Is there any publicly available graph of GMail or other email systems, quantifying the connectivity of the users (I expect a power law)? And any analysis of GMail "good" and "bad" users?

2 comments

[ 3.2 ms ] story [ 14.1 ms ] thread

mcherm 16y ago ↗

I don't know the answer to your question, but I do know that one of your assumptions is incorrect. The reason for making gmail invitation-only was NOT to build a graph and keep nodes artificially small. Rather, the goal was to provide a way of ramping up usage in a controlled manner. By throttling the number of invites given out, Google could control the rate at which new people signed up for the service and make sure that there were always enough servers available to handle the load so the experience for existing users wasn't degraded.

They are currently doing the exact same thing with accounts for Google Wave. In this case, there is an added benefit: unless you know someone else who has a wave account, having one is of absolutely no use because you have no one to communicate with.

albertcardona 16y ago ↗

I didn't think about the scaling-up issues. Makes a lot of sense.