| 10:32 pm on Jan 12, 2004 (gmt 0)|
> it doesn't seem to matter
I think it has more to do with Google being a search engine, rather than a human edited index of over three billion URLs.
Even a fairly rudimentary attempt to interpret the meaning of sentences with software would be a rather costly excercise on such a massive scale.
| 12:54 am on Jan 13, 2004 (gmt 0)|
Think like a bot and you'll be able to write a great site for google :) Although if it reads like canned googlebot food when the human googlers come along they may take action...
For information on becoming a bot, dial 1-800-555-BOT.
| 1:39 am on Jan 14, 2004 (gmt 0)|
Some blatent spam sites do really well in searches. Keyword stuffed dynamically generated filenames, with pages filled full of total junk.
I can recognise such stuff just by the URL, without looking at the description part; but the description is the giveaway.
Why do people bother creating such useless cr*p? Clicking the link tries to set multiple cookies, and triggers a large amount of popups.
| 2:52 am on Jan 14, 2004 (gmt 0)|
I would think it would be really easy for google to use some grammer filters. Nothing too harsh, but at least something that could tell the difference between useful information and blatant spam.
Sorry if my original post sounded bitter.
| 3:28 am on Jan 14, 2004 (gmt 0)|
>I would think it would be really easy for google to use some grammer filters. Nothing too harsh,
>but at least something that could tell the difference between useful information and blatant spam.
Would never work. I don't believe artificial intelligence is advanced enough these days to recognize, say, poetry as English grammar, yet spam paragraphs as not. A filter that would block Shakespearean sonnets wouldn't be doing students a great service. And that's not even getting into modern poetry issues.
The differences between auto-generated spam and dadaist literature... there's a master's thesis out there for some enterprising soul... :-D
| 3:45 am on Jan 14, 2004 (gmt 0)|
Yeah I guess youre right. Better not to penalize than to penalize and mistake a few innocents for guilty.
One day in the future hopefully there will be AI smart enough not to make mistakes.
| 4:23 am on Jan 14, 2004 (gmt 0)|
Even High IQ Humans will have trouble in realising that this postmodernism article [elsewhere.org] is machine generated let alone google :)
| 5:26 am on Jan 14, 2004 (gmt 0)|
regardless of whether or not the technology exists. The costs of using such a technology, in cpu processing and thus $$, would out weigh the benifits.