bzero v0.03 out - fixes a calendar drawing bug that was screwing up my calendar when I posted today ;-)
... more like this: [2002-9-3
Sri Chimnoy spam report
A while back I noticed a whole heap of URLs with 'srichinmoy' in the name showing up in my ecosystem's daily URL-suck from weblogs.com. I didn't do anything about them, but coming back to the ecosystem a couple of days ago I see that they have risen quite high in the rankings. That's not good.
Then, today, I noticed that andersja has seen this, and Mo Morgan has a page tracking the spam domains. Interesting.
The way to fix this is to implement some sort of spam-proofing system like the one Google uses to cut link farms out of the equation (the PageRank equation, that is). Kind of like garbage collection, the technique that languages like Lisp, Python and Java (among many others) use to get rid of unused data in running software: look at all the links inside the ecosystem and look for groups of pages that only refer to each other, then prune them out. Real weblogs connect to each other: if you can't find a path from scripting.com to a blog, it's probably spam.
Hey, I wonder if nearest.py (in the apps area) will do this. Hmm.
Update: Here's the ecosystem stats page for srichinmoylibrary.com. No links from 'real blogs', although weblogs.com and myelin.co.nz/ecosystem will form a path that will stop my garbage collector idea from working. I'm going to have to hack together something like PageRank or Raph Levien's technique.
Then, today, I noticed that andersja has seen this, and Mo Morgan has a page tracking the spam domains. Interesting.
The way to fix this is to implement some sort of spam-proofing system like the one Google uses to cut link farms out of the equation (the PageRank equation, that is). Kind of like garbage collection, the technique that languages like Lisp, Python and Java (among many others) use to get rid of unused data in running software: look at all the links inside the ecosystem and look for groups of pages that only refer to each other, then prune them out. Real weblogs connect to each other: if you can't find a path from scripting.com to a blog, it's probably spam.
Hey, I wonder if nearest.py (in the apps area) will do this. Hmm.
Update: Here's the ecosystem stats page for srichinmoylibrary.com. No links from 'real blogs', although weblogs.com and myelin.co.nz/ecosystem will form a path that will stop my garbage collector idea from working. I'm going to have to hack together something like PageRank or Raph Levien's technique.