google to separate blogs from the web -- not


Evan posted a comment to my previous entry google to separate blogs from the web, squashing the article from The Guardian with a simple-but-effective: "It's B.S.". He also has a great entry on his weblog titled "Register to fix Orlowski noise problem". LOL. He notes the claim in the article (that also sounded strange to me) that the number of readers of weblogs is "statistically insignificant". I could only make sense of that if they consider the whole population of the Internet as the basis for the measure, so, several million against half a billion or something... anyway, Evan's statements are quite definitive (e.g., "as far as I know, Orlowski is full of crap"), here and on his blog, so I don't think there's much room for doubt. Hopefully the rumour-squashing will move as fast as the rumour itself!

Also in the comments to my entry Don was pointing out that one could use Bayesian filters and some simple rules to figure out "suspected blogs" (Ashcroft would like that phrase I think), which is a good point. However, a filter like that is never going to be 100%, and even if the error is less than 1%, when you're indexing 3 billion pages, that's too much. Google wouldn't be able to cope with the requests from people that have been incorrectly marked, while those that haven't been marked (and should have been) would be riding it out, something that would probably make others in the blogging community a little "unhappy". In the end it would just start an "arms race" of sorts, with bloggers looking for ways to fool the rules... totally unproductive, and neither side would really win. Bloggers would stop using Google. Google would get pissed off customers. Not a good deal for either.

Regardless: given Evan's comments, I think we can put the question to rest. At least until the next rumour :-)

Later (about 6 hours later :-)): Evan's entry in which he squashed the google-not-for-weblogs-rumour, has disappeared! In its place it says "</snip>". The conspiracy theory wheels begin to turn.... but wait, not so fast. If you look at the page source, you get the deal. The entry is still there, just commented, with the following note: "you know, in order to spread more 'Google censors Evhead' suspicions". Heh.

Categories: technology
Posted by diego on May 9 2003 at 9:45 PM
Comments (please see the comments & trackback policy).

Three points:

1. I million out of one billion is statistically significant if I remember my math correctly..

2. Blogger which google just bought has 1 miillion users by itself..the correct underlying total of bloggers is probably at 5 million..

3. If you assume 5 unqiue readers per blog that is 25 million ..but that is a very conservative number..

Posted by: Fred Grott at May 10, 2003 2:23 AM


I told you via IM, I suspect Evan is completely wrong. That's a nice joke about the conspiracy theories on his site but if he wacked an entry, it's for a reason.

-Russ

Posted by: Russ at May 10, 2003 11:59 AM

Fred, in statistics of this kind, where you are extrapolating population behavior based on sampling, errors of 1% to 5% are the norm. Polls and such usually hover around plus or minus 3%. So 1% is not statistically significant. Also, blogger has 1 million users but around 200,000 that post actively (I read that somewhere a few weeks ago). The blogging community might be between ` and 5 million, IMO, but who knows.

Russ: as I said, Evan's posting wasn't whacked, he just commented it, as a joke. In fact, I just checked and the text is back in.

Posted by: Diego at May 10, 2003 2:04 PM

Copyright © Diego Doval 2002-2007.
Powered by
Movable Type 3.35