Via Jeneane Sessum I jumped onto Doc's thread about the splog problem and his supposition that its ultimate roots lie deeper inside the homogeny and monoculturalism in the tech world. He raises some interesting questions I'd love your feedback on: Do you think Google has become a monoculture in search and advertising both? How big is the problem of link devaluation, and can the splog problem be solved by rethinking the concept of page-rank? And more largely — how do we take concrete steps to move towards polyculture in search, in advertising, in the workplace, in the tech industry?
Splogs, links, attention, and tech monoculture
Reader Comments
(Page 1)2. Forgive me if I'm repeating myself, but I think that one solution could be sliders to set the bar for some sort of collaborative attention/validation standard for results from any given search. ("show me only results that have been flagged as spam by fewer than 0-5 unique IPs")
Second, I think that learning who you like in social recomendation enabled search ecosystems will also help. (I've been checking out http://wink.com)
Splogs in my search-RSS feeds really bum me out, but I think that a good place to start dealing with the problem could be in aggregated attention data of some sort (people don't spend lots of time looking at splogs, for one thing). And in as much as social recomendation would further help the problem, the more lock-in to a proprietary silo there is, the less likely I am to use or evangelize that system.
Finally, I think that building a culture that prioritizes delivery of value as the key to earning money and moves away from the IT=easy $ paradigm, the less scum bag b.s. like splogs we'll see. I hope so, anyway.
Posted at 4:52AM on Dec 19th 2005 by Marshall Kirkpatrick
3. I agree with Marshall - people are harder to foil than algorithms: we need to adopt people-powered search, or at least allow people to report spam results (I think Google has started to implement this). And obviously we need to verify that these people are indeed *real* people, and not just scripts. I think http://www.wink.com could succeed here, but I'm not so convinced by http://www.rollyo.com.
Using attention data could be key to all this, and I don't think we've yet fully understood how important attention could be.
Posted at 4:52AM on Dec 19th 2005 by PeterCashmore
4. None at the moment
doncash_ow@yahoo.com
Posted at 4:52AM on Dec 19th 2005 by Dan
5. In response to allowing people to report spam, Blogspot / Blogger has recently done so with alarming results. I know of a few people who make their living from splogs etc (which I dont agree with) and have been hit hard by the report function - while not the main solution, it may certainly help
Posted at 4:52AM on Dec 19th 2005 by Indie Music News
6. I was just putzing around on Technorati and some japanesse blog just came out of nowhere to pass BoingBoing as the most popular blog,with links from close to 36,000 sites.
Take a look for yourself http://www.technorati.com/pop/blogs/
Granted I am no blog expert and I dont know japanesse, but this looks like a case of automated linking software
7. I feel search engine companies believe that page rank is the key reason for splogs and web spam. That's why the moment the nofollow tag was introduced for comments and the like, they felt they were sort of done with comment spam, and now they're mainly concerned about indexing crappy sites than anything else. I beg to differ.
Spammers may try taking advantage of search engines, but their main goal is not to be first in search results, but to get their trash to as many people as possible. Sure, being top in the search results is a way to accomplish this, but it's not the only way, and should PageRank-like algorithms stop being used, web spam would not cease. Perhaps might even get worst, as if they (the spammers) cannot get to be in the top search results, they'd hammer even stronger - kind of like what happens with email spam: now we have better spam filters so what do they do? Send even more spam in the hope some of it might get to someone.
Posted at 4:52AM on Dec 19th 2005 by RBA
8. Forgive me if I'm repeating myself, but I think that one solution could be sliders to set the bar for some sort of collaborative attention/validation standard for results from any given search. ("show me only results that have been flagged as spam by fewer than 0-5 unique IPs")
Second, I think that learning who you like in social recomendation enabled search ecosystems will also help. (I've been checking out http://wink.com)
Splogs in my search-RSS feeds really bum me out, but I think that a good place to start dealing with the problem could be in aggregated attention data of some sort (people don't spend lots of time looking at splogs, for one thing). And in as much as social recomendation would further help the problem, the more lock-in to a proprietary silo there is, the less likely I am to use or evangelize that system.
Finally, I think that building a culture that prioritizes delivery of value as the key to earning money and moves away from the IT=easy $ paradigm, the less scum bag b.s. like splogs we'll see. I hope so, anyway.
Posted at 4:52AM on Dec 19th 2005 by Marshall Kirkpatrick
9. I agree with Marshall - people are harder to foil than algorithms: we need to adopt people-powered search, or at least allow people to report spam results (I think Google has started to implement this). And obviously we need to verify that these people are indeed *real* people, and not just scripts. I think http://www.wink.com could succeed here, but I'm not so convinced by http://www.rollyo.com.
Using attention data could be key to all this, and I don't think we've yet fully understood how important attention could be.
Posted at 4:52AM on Dec 19th 2005 by PeterCashmore
10. None at the moment
doncash_ow@yahoo.com
Posted at 4:52AM on Dec 19th 2005 by Dan
11. In response to allowing people to report spam, Blogspot / Blogger has recently done so with alarming results. I know of a few people who make their living from splogs etc (which I dont agree with) and have been hit hard by the report function - while not the main solution, it may certainly help
Posted at 4:52AM on Dec 19th 2005 by Indie Music News
12. I was just putzing around on Technorati and some japanesse blog just came out of nowhere to pass BoingBoing as the most popular blog,with links from close to 36,000 sites.
Take a look for yourself http://www.technorati.com/pop/blogs/
Granted I am no blog expert and I dont know japanesse, but this looks like a case of automated linking software
1. I feel search engine companies believe that page rank is the key reason for splogs and web spam. That's why the moment the nofollow tag was introduced for comments and the like, they felt they were sort of done with comment spam, and now they're mainly concerned about indexing crappy sites than anything else. I beg to differ.
Spammers may try taking advantage of search engines, but their main goal is not to be first in search results, but to get their trash to as many people as possible. Sure, being top in the search results is a way to accomplish this, but it's not the only way, and should PageRank-like algorithms stop being used, web spam would not cease. Perhaps might even get worst, as if they (the spammers) cannot get to be in the top search results, they'd hammer even stronger - kind of like what happens with email spam: now we have better spam filters so what do they do? Send even more spam in the hope some of it might get to someone.
Posted at 4:52AM on Dec 19th 2005 by RBA