Comments on: How do you deal with competitors in open systems? http://www.jimgilliam.com/2009/05/how-do-you-deal-with-competitors-in-open-systems/ by Jim Gilliam Sun, 07 Mar 2010 16:54:53 +0000 hourly 1 By: Daniel DeLorme http://www.jimgilliam.com/2009/05/how-do-you-deal-with-competitors-in-open-systems/comment-page-1/#comment-46 Daniel DeLorme Sun, 31 May 2009 09:27:37 +0000 http://www.jimgilliam.com/?p=77#comment-46 Maybe there's a way to make these people work *for* the system, despite their efforts to the contrary? If somehow the effect of their actions could be measured, and those actions were found to have a consistent bias *against* the objectives of the system, then you could internally label them as saboteurs and do the opposite of whatever they say; i.e. if a saboteur votes for X that should be counted as a vote *against* X. Of course the trick is figuring out how to measure this consistent "saboteur's bias". You could apply statistical clustering techniques to separate users into two groups and label one of them as saboteurs... but which one? Or you could use a baseline group of "trusted" users and people with behavior opposite to that core group would be labeled saboteurs... but that would risk excluding all dissenting opinions. It's a tricky problem, no doubt about it. Maybe there's a way to make these people work *for* the system, despite their efforts to the contrary? If somehow the effect of their actions could be measured, and those actions were found to have a consistent bias *against* the objectives of the system, then you could internally label them as saboteurs and do the opposite of whatever they say; i.e. if a saboteur votes for X that should be counted as a vote *against* X. Of course the trick is figuring out how to measure this consistent "saboteur's bias". You could apply statistical clustering techniques to separate users into two groups and label one of them as saboteurs… but which one? Or you could use a baseline group of "trusted" users and people with behavior opposite to that core group would be labeled saboteurs… but that would risk excluding all dissenting opinions. It's a tricky problem, no doubt about it.

]]> By: Nick Tesla http://www.jimgilliam.com/2009/05/how-do-you-deal-with-competitors-in-open-systems/comment-page-1/#comment-41 Nick Tesla Sun, 31 May 2009 02:01:56 +0000 http://www.jimgilliam.com/?p=77#comment-41 You may want to read "smart start-ups" or "The Social Network Business lanModel," both by David Silver. In the former, Silver argues for reputation management rules to block "sex perverts and other psychopaths," "corporate flacks" and others who don't cooperate. Matter of fact, he says: "Sure, people can register under multiple IP addresses and thus appear to be sheep, when they are in fact wolves in sheep's clothing. But through aggressive collaborative filtering, data mining, exchange of information between communities through the reputation management organization and pattern matching, people who change their online pseudonyms can be discovered." While this sounds extreme, perhaps only by exploring the extremes will we finally deal with the three-cornered problem of IDENTITY, which is required in order to establish trust, ANONYMITY, which is required in open systems to which predators have access, and PRIVACY, which the internet has yet to effectively address. You may want to read “smart start-ups” or “The Social Network Business lanModel,” both by David Silver. In the former, Silver argues for reputation management rules to block “sex perverts and other psychopaths,” “corporate flacks” and others who don’t cooperate. Matter of fact, he says:

“Sure, people can register under multiple IP addresses and thus appear to be sheep, when they are in fact wolves in sheep’s clothing. But through aggressive collaborative filtering, data mining, exchange of information between communities through the reputation management organization and pattern matching, people who change their online pseudonyms can be discovered.”

While this sounds extreme, perhaps only by exploring the extremes will we finally deal with the three-cornered problem of IDENTITY, which is required in order to establish trust, ANONYMITY, which is required in open systems to which predators have access, and PRIVACY, which the internet has yet to effectively address.

]]> By: Jim Gilliam http://www.jimgilliam.com/2009/05/how-do-you-deal-with-competitors-in-open-systems/comment-page-/#comment-36 Jim Gilliam Sat, 30 May 2009 22:44:05 +0000 http://www.jimgilliam.com/?p=77#comment-36 Thanks for your thoughts.... I do remember reading recently about the registration/moderation approach having the opposite effect of what was intended. What about the people who in fact want the *general* project to fail. Let's say I'm running Craigslist in an open fashion, anyone on the internet can be involved on how the company should be run. It's quite reasonable to expect that people working at newspapers might view the success of Craigslist as directly hurting their ability to make a living. So they participate in how Craigslist should be run, and suggest things like charging for listings so their classified ads wouldn't be undercut. I suppose more and more transparency does help that, although it might be hard to actually get that out of people. But if you know that 70% of the people supporting this one idea all happen to work in the newspaper industry, then you can take that into consideration. So then, how does that manifest itself on a website in an easily digestible way. We can't expect everyone to delve into everyone else's background. Maybe there's a means of tagging other people? In politics, it could be something like "20 people tagged this person a lobbyist"...something like that. Thanks for your thoughts…. I do remember reading recently about the registration/moderation approach having the opposite effect of what was intended.

What about the people who in fact want the *general* project to fail. Let's say I'm running Craigslist in an open fashion, anyone on the internet can be involved on how the company should be run. It's quite reasonable to expect that people working at newspapers might view the success of Craigslist as directly hurting their ability to make a living. So they participate in how Craigslist should be run, and suggest things like charging for listings so their classified ads wouldn't be undercut.

I suppose more and more transparency does help that, although it might be hard to actually get that out of people. But if you know that 70% of the people supporting this one idea all happen to work in the newspaper industry, then you can take that into consideration.

So then, how does that manifest itself on a website in an easily digestible way. We can't expect everyone to delve into everyone else's background. Maybe there's a means of tagging other people? In politics, it could be something like "20 people tagged this person a lobbyist"…something like that.

]]>