12.148.209.198 - - [21/Feb/2003:05:45:13 +0000] "GET /robots.txt HTTP/1.1" 404 300

12.148.209.198 - - [21/Feb/2003:06:08:01 +0000] "GET / HTTP/1.1" 200 7651

12.148.209.198 - - [21/Feb/2003:06:43:01 +0000] "GET /gate/archive.html HTTP/1.1" 401 493

12.148.209.198 - - [21/Feb/2003:07:21:54 +0000] "GET /gate/video.html HTTP/1.1" 401 493

12.148.209.198 - - [25/Feb/2003:11:29:35 +0000] "GET /robots.txt HTTP/1.1" 404 300

12.148.209.198 - - [25/Feb/2003:12:00:44 +0000] "GET /gate/album/index.html HTTP/1.1" 401 493

12.148.209.198 - - [25/Feb/2003:12:31:26 +0000] "GET /gate/archive.html HTTP/1.1" 401 493

12.148.209.198 - - [25/Feb/2003:13:04:24 +0000] "GET /gate/video.html HTTP/1.1" 401 493

12.148.209.198 - - [09/May/2003:20:18:32 +0100] "GET /robots.txt HTTP/1.1" 404 300

12.148.209.198 - - [09/May/2003:20:18:32 +0100] "GET /gate/album/index.html HTTP/1.1" 401 493

12.148.209.198 doesn't resolve on a reverse dns lookup. but further investigation turns up: this link, which leads to http://www.nameprotect.com

Interesting to see the words they are looking for: album, archive, video. at least they first look for robots.txt.

I'm getting hooked on access_log!

I should note that Roger's 'blog and mine are hosted on the same server, run by a guy who lives up to his motto "World's worst ISP". We have to grep the Apache logs to find the hits just for our sites, 2 among the six hosted by this guy. He's lazy, and can't be bothered to configure Apache to spit out separate logs... a least this 'bot is relatively polite...
Comments [ ] 10:48:50 PM    

63.148.99.232 - - [09/May/2003:08:31:47 -0700] "GET / HTTP/1.1" 200 81802 "http://www.gavinsblog.com/2003/03/31.html" "Mozilla/4.0 (compatible; MSIE 5.05; Windows NT 3.51)"

63.148.99.232 - - [09/May/2003:09:00:38 -0700] "GET /2003/05/05.html HTTP/1.1" 200 27517 "http://scriptingnews.userland.com/2003/05/05" "Mozilla/4.0 (compatible; MSIE 5.05; Windows NT 5.0)"

63.148.99.232 - - [09/May/2003:09:52:31 -0700] "GET /2003/05/01.html HTTP/1.1" 200 31896 "http://www.smalla.net/infofeed/index-full.rdf" "Mozilla/4.0 (compatible; MSIE 5.05; Windows NT 3.51)"

63.148.99.233 - - [09/May/2003:10:35:04 -0700] "GET /blog/ HTTP/1.1" 200 36281 "http://www.prandial.com/archives/004099.html" "Mozilla/4.0 (compatible; MSIE 5.05; Windows NT 3.51)"

63.148.99.232 - - [09/May/2003:11:38:33 -0700] "GET /rutland/ HTTP/1.1" 404 285 "http://www.n-net.com/vt.htm" "Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.0)"

Note the changing IDs: I wonder if this is how they know which crawler is working. But it still doesn't seem to be aware of the existence of a veritable HTML shrine to all things Cyveillancebot. Googlebot knows, which lends credence to the theory that Cyveillancebot doesn't really have the horsepower to traverse the Web and keep up with new pages. Most of the crawls above are from and to old material the 'bot has already crawled.

As seen above, Cyveillancebot has, finally, crawled the page you're reading now, which has lots of links to the C-bot materials, so it will be interesting to see how long it takes before it crawls those directories. I'm not looking forward to the 'thrashing' or mode 2 behavior that it exhibited in the past - it will consume all our bandwidth while it's at it, but I guess it's the only way to discover how it works.

Speaking of Googlebot, I wonder why they just don't use Google or Yahoo search? That way they wouldn't have to tie up the resources for a crawler farm, their searches would be invisible, and their results would be much more up to date. If I was going to start a service, call it CyberSnitch, I'd cut a deal with Google to do the crawling, and then point the AI to the Google results...

Comments [ ] 1:41:50 PM    

Some ten years ago, I helped a fellow named Neil Chase and the Hearst Corporation launch a new weekly magazine called "We", a bilingual publication that circulated in the U.S. and Russia. It was, at the time, the fastest and cheapest launch of a Hearst publication ever. Though wildly popular in Russia We succumbed to market forces: Izvestia, the Russian publishing partner, couldn't afford the paper to print it on.

And though We didn't make it, the Hearst Corporation adopted the methods that Neil and I developed - based on desktop publishing - for all of its launches since. That 'miracle' publication went from idea to newsstand in something like 3 months. Yesterday, I did what amounts to the same thing, by myself, in 5 minutes.

It helped that I had excellent software - Userland Radio and NetNewsWire - as well as Apache, but it is still an awesome thing. Five minutes from idea to a publication with global possibilities... and I can afford the 'paper' to 'print' it on...
Comments [ ] 9:23:05 AM    

Chris Gulker's view from Silicon Valley - in words and pictures