Sorry, did'nt know average click price was that expensive. I knew some clicks were actually very expensive but (mesotholomia for instance), but I avoid them.
Options is removed from the add-on anyway, now only click on non-commercial results are possible. But the point is that... well Google will now that when you click on ads, it's for real...
Moreover, I wanted to point out that pay-per-click can't hold... 15$, how could that be that expensive.
Other comments ?
I'm making a plug-in like TackMeNot called SquiggleSR (http://squigglesr.free.fr). In addition to issuing queries, it also clicks on results and ads. I think it points out a serious problem of the pay-per-click economical approach.
Moreover I'm trying to make more coherent queries (using keywrods extracted from RSS flows). See the site for more details and do not hesitate to post comments.
Well TOR and other proxy actually works, but they are not compliant with any kind of personalized search, so you have to make a choice...
Furthermore, these solutions will be useless for Phone using android. If Google can get your queries at the OS level, proxy won't bring any kind of privacy.
As I already mentioned, I'm developing a privacy enforcement plug-in which generates false queries (it's quite like TrackMeNot, but with more coherent and personalized queries see http://squigglesr.free.fr/ for details).
As my plug-in click also on ads, it may cost money to Google since advertiser won't be happy to be charged for simulated click. So, if tomorrow Google asks you to pay the bill for your privacy, how much are you willing to pay?
I'm not sure we'll agree to pay for all the services if they had to be paid with dollars... if we are not even ready to pay just by switching search engine.
"Don't be evil" statement does not mean Google wants to be good. They are simply neutral; their business is gathering profiles and providing contents. If a government needs a service (obtaining personal information or closing an account) they'll provide it otherwise they may be censored and loose a market
Considering the number of Wifi access poorly protected (Wep) or totally open, I think that "Illegal Downloaders" neighbours should prepare for hard days.
Well my plug-in is not TrackMeNot (actually it's named SuigglesSR).I'm aiming to simulate accurately a search engine user behaviour, that's why I use RSS flows, I think it's more representative of our interest, and you can modify your OPML file to simulate fake interests (noisier profile).
I add a log visualisation, so you can see in SquiggleSR preferences the generated queries and the clicked results. The results are not too bad now, queries are quite realistic.
Are the search engines able to detect it? I think it depends on the number of queries issued by Squiggle, but till now I have no problem with my Google account. The think is that the pattern is strongly dependant on the RSS flows so it's quite hard to detect it.
But if we are able to simulate real user behaviours with clicks, it's clearly a threat to the "pay per click" model. However Google is aiming to make us react like automates to personalized ads...thus an automate could act like us.
Now they will use other data provided by Android, Youtube and co to detect our interest...
When you think about it, Google and Microsoft beating for health files, knowing that the highest bid for pay per click is "mesothelioma" (a cancer), this clearly points out that our privacy and their interest are conflicting.
Well my plug-in is not TrackMeNot. Actually TrackMeNot generates queries from search engine statistics, while I generate queries using words extracted from user RSS flows.
The generated queries depend on the RSS flows you selected; so I don't think the queries "raping virgins" will be generated. If they are, it's because it's related to news. So it depends on RSS flows you pick up. The default profile corresponds to me and the queries I generated are not perfect but I could have make most of them.
In the last version, I added a log, so you can directly see what queries are generated.
Basically, if you import your own OPML file and then edit it, I think the queries will correspond to your profile. So the "noise" will depend on how you edit your profile.
Ideally I'd like to let users modify the regexp used to extract key words to provide more realistic results.
For those of you who want to protect their privacy, I've made a light Firefox add-on which generates randomly some queries on Google to make your search profile noisier and less exploitable. The queries keywords are extracted from RSS flows so you can personalize them. Moreover, the program simulates some clicks on Google search results (and ads).
For further information go on: http://sourceforge.net/projects/fuzzy-search/
It's a beta version and any comments are appreciated.
Another case where it could be used is subway, where density is high enough to be sure that the network is always connected.
Look at the number of 802.11 capable devices which are used in subway : IPhone (and now IPod touch), PSP, DS and smartphones.
You can connect all these devices to create a local network where you can share your favorite youtube videos, podcast and blog or simply play video game.
Clearly this should not be possible using classical GSM or WiFi connections which can not handle the traffic. But mesh networks can.
Slashdot Mirror
Oups, didn't know it was that serious, I removed the program until I can fix this.
Sorry, did'nt know average click price was that expensive. I knew some clicks were actually very expensive but (mesotholomia for instance), but I avoid them. Options is removed from the add-on anyway, now only click on non-commercial results are possible. But the point is that... well Google will now that when you click on ads, it's for real...
Moreover, I wanted to point out that pay-per-click can't hold... 15$, how could that be that expensive. Other comments ?
I'm making a plug-in like TackMeNot called SquiggleSR (http://squigglesr.free.fr). In addition to issuing queries, it also clicks on results and ads. I think it points out a serious problem of the pay-per-click economical approach.
Moreover I'm trying to make more coherent queries (using keywrods extracted from RSS flows). See the site for more details and do not hesitate to post comments.
Well TOR and other proxy actually works, but they are not compliant with any kind of personalized search, so you have to make a choice...
Furthermore, these solutions will be useless for Phone using android. If Google can get your queries at the OS level, proxy won't bring any kind of privacy.
As I already mentioned, I'm developing a privacy enforcement plug-in which generates false queries (it's quite like TrackMeNot, but with more coherent and personalized queries see http://squigglesr.free.fr/ for details).
As my plug-in click also on ads, it may cost money to Google since advertiser won't be happy to be charged for simulated click. So, if tomorrow Google asks you to pay the bill for your privacy, how much are you willing to pay?
I'm not sure we'll agree to pay for all the services if they had to be paid with dollars... if we are not even ready to pay just by switching search engine.
"Don't be evil" statement does not mean Google wants to be good. They are simply neutral; their business is gathering profiles and providing contents. If a government needs a service (obtaining personal information or closing an account) they'll provide it otherwise they may be censored and loose a market
Considering the number of Wifi access poorly protected (Wep) or totally open, I think that "Illegal Downloaders" neighbours should prepare for hard days.
Well my plug-in is not TrackMeNot (actually it's named SuigglesSR).I'm aiming to simulate accurately a search engine user behaviour, that's why I use RSS flows, I think it's more representative of our interest, and you can modify your OPML file to simulate fake interests (noisier profile).
I add a log visualisation, so you can see in SquiggleSR preferences the generated queries and the clicked results. The results are not too bad now, queries are quite realistic.
Are the search engines able to detect it? I think it depends on the number of queries issued by Squiggle, but till now I have no problem with my Google account. The think is that the pattern is strongly dependant on the RSS flows so it's quite hard to detect it.
But if we are able to simulate real user behaviours with clicks, it's clearly a threat to the "pay per click" model. However Google is aiming to make us react like automates to personalized ads...thus an automate could act like us.
Now they will use other data provided by Android, Youtube and co to detect our interest...
When you think about it, Google and Microsoft beating for health files, knowing that the highest bid for pay per click is "mesothelioma" (a cancer), this clearly points out that our privacy and their interest are conflicting.
Well my plug-in is not TrackMeNot. Actually TrackMeNot generates queries from search engine statistics, while I generate queries using words extracted from user RSS flows.
The generated queries depend on the RSS flows you selected; so I don't think the queries "raping virgins" will be generated. If they are, it's because it's related to news. So it depends on RSS flows you pick up. The default profile corresponds to me and the queries I generated are not perfect but I could have make most of them.
In the last version, I added a log, so you can directly see what queries are generated.
Basically, if you import your own OPML file and then edit it, I think the queries will correspond to your profile. So the "noise" will depend on how you edit your profile.
Ideally I'd like to let users modify the regexp used to extract key words to provide more realistic results.
For those of you who want to protect their privacy, I've made a light Firefox add-on which generates randomly some queries on Google to make your search profile noisier and less exploitable. The queries keywords are extracted from RSS flows so you can personalize them. Moreover, the program simulates some clicks on Google search results (and ads).
For further information go on: http://sourceforge.net/projects/fuzzy-search/
It's a beta version and any comments are appreciated.
Another case where it could be used is subway, where density is high enough to be sure that the network is always connected. Look at the number of 802.11 capable devices which are used in subway : IPhone (and now IPod touch), PSP, DS and smartphones. You can connect all these devices to create a local network where you can share your favorite youtube videos, podcast and blog or simply play video game.
Clearly this should not be possible using classical GSM or WiFi connections which can not handle the traffic. But mesh networks can.