Forums
| BadPhorm - When good ISPs go bad! :: Forums :: Phorm Discussion :: Tech Talk |
|
<< Previous thread | Next thread >> |
| Phorm, Discover, and Relevant Links. | ||
|
Moderators: Jim Murray, narcosis, felixcatuk, Sammy, revrob
|
| Author | Post | ||
| felixcatuk |
| ||
felixcatuk   Registered Member #95 Joined: Wed Mar 05 2008, 12:03AMPosts: 2626 | I am wondering where this database of 'relevant sites' is coming from. I don't think Phorm have spidered or stored the content of the internet, nor do I think they have a PageRank style algorithm iterating over the content of the internet, computing those links of relevance to a given term. So, presumeably, they have partnered with a competent search engine (because if they are rewriting search engine results without consent it would be a 'bad idea'TM). So who is the search engine partner who is providing this relevant content for the widget? (What is the search engine spider 'user-agent'? If Phorm are operating the search engine, what is their spider user-agent?) | ||
| Back to top | | ||
| revrob |
| ||
| Registered Member #372 Joined: Wed Apr 23 2008, 04:09PMPosts: 640 | User agent? What's that? Phorm don't DO user agents!! | ||
| Back to top | | ||
| VforVendetta |
| ||
 Registered Member #281 Joined: Wed Apr 02 2008, 11:59PMPosts: 620 | Which leads us back to the question, who is funding Phorm? | ||
| Back to top | | ||
| madslug |
| ||
  Registered Member #266 Joined: Tue Apr 01 2008, 01:11PMPosts: 800 | It seems very probable that Phorm are holding a cache of various pages. This was suspected when the scraper bots were first seen shortly after domain owners contacted Phorm to have their sites opted out - interestingly within the 48 hour deadline. Richard Clayton's report on Webwise indicates that they have a programable crawler bot and system capable of caching the robots.txt file for up to one month. My logs show that the bot which requests the robots.txt file does not display a useragent. Practically that leaves 2 methods for recording the content of publisher sites for various channels. 1. Using an opensource script like Nutch to build a database of 'media' content plus the top 50 popular sites. They could daily scrape any site that hosts the widget script. 2. Expand on the existing channel feature where URLs are recorded within the channel analysis which would make channel matching very quick and mean that users are being used to update the content library in real time. The one thing that makes me doubt that only 2 is an option is that the widget images shown provide snippets from the content and that would only be possible with a database as per #1. It is well known that the major search engines provide scripts that allow 'anyone' to include their search results within the web page. As this is accompanied by lots of PPC type URLs which provide revenue to the search engine and site displaying the results, and there is mention that there is no revenue involved with the display of the free widget, I discount any major search engine as the source of the content database. One of the features of KindSight, during its testing last year, was that it was sending out a bot (which was based on Nutch and identified as coming from ProtectRialto). Whether or not the same is happening during the current testing is not known - nothing spotted on any forums about the crawler being active again. I was going to type up some conclusions. However I would rather not make them public. I work on the principle that most people reading this will work it out for themselves. Anyone not too sure is welcome to PM me. | ||
| Back to top | | ||
| felixcatuk |
| ||
| felixcatuk Registered Member #95 Joined: Wed Mar 05 2008, 12:03AMPosts: 2626 | Dephormation 'Recover' widgets launched! After nearly 7 minutes in development (with no source of revenue) Dephormation "Recover" Widgets will transform your dull and tiresome website from a tedious and uninteresting sales pitch, to a place with links to more interesting stuff on other more exciting web sites. Watch in amazement as your visitors leave your site for better content somewhere else on the internet and never return! See Dephormation Widgets in Action No download required! Doesn't use Flash/Java/ActiveX! And no DPI required either! | ||
| Back to top | | ||
| badphormula2 |
| ||
  Registered Member #653 Joined: Thu Apr 30 2009, 10:32AMPosts: 106 | You need to follow that 7 minute widget nugget up with a press conference and a bogus survey! Sorry you can't have 150 employees in a big circle and ask them many times what they think, until you reach a good sample number. "Vladimir have you got your friends from the Russian wrestling academy ready?" "Norman, are them undergrads from your old Cambridge college ready?" "Telegraph marketing department, ready?" [ Edited Thu Jun 04 2009, 10:53PM ] | ||
| Back to top | | ||
| felixcatuk |
| ||
| felixcatuk Registered Member #95 Joined: Wed Mar 05 2008, 12:03AMPosts: 2626 | You know what? Having created that page I get a very bad feeling about this widget idea. Even worse than Phorm and DPI. Phorm scam your traffic using DPI to profile you, thats a given. To get 'relevant' links, they have to reveal part of that profile to a search engine (on the assumption they don't operate the search engine) as keywords. So hypothetically assume that search engine is Google/Yahoo/LiveSearch. The search engine is potentially now receiving both what you explicitly search for, AND implicit data from your profile too. It can potentially link all of that data to your IP address courtesy Phorm. So if you think you have a privacy problem with Phorm, or a privacy problem with search engines, let me introduce you to something much worse. In effect Phorm leaking your communication profile to a search engine. This needs to be watched very carefully indeed. | ||
| Back to top | | ||
| clanger9 |
| ||
  Registered Member #40 Joined: Wed Feb 27 2008, 06:31PMPosts: 69 | Yep, the Widget (if implemented as described) appears to completely destroy Phorm's claims of "preserving anonymity". The original Phorm system had quite a clever feature in that the advertiser never actually got to find out what your interest categories were. Assuming you trusted Phorm more than the advertiser (*cough*), this could actually have been construed as a privacy enhancement over existing BTA methods. But that one "advantage" of Phorm is now gone. The other weird thing is that the Widget appears to be free. If that really is the case then if effectively allows world+dog to gain access to the results of Phorm's snooping without paying the OIX subscription. Which means that thay just cut off their only revenue stream too. This is a truly bizarre company... [ Edited Tue Jun 09 2009, 05:28PM ] | ||
| Back to top | | ||
| felixcatuk |
| ||
| felixcatuk Registered Member #95 Joined: Wed Mar 05 2008, 12:03AMPosts: 2626 | The more you consider this bizarre idea, the more it sucks even more than vanilla Webwise. They are explicity linking the results of profiling with the IP address you use (by including the widget in pages you view) thereby allowing the organisation presenting the webpage and/or the ads and/or the search engine providing the relevant content to gain access to the profile data linked to IP address (as well as Phorm cookie). [ Edited Sun Jun 07 2009, 10:46PM ] | ||
| Back to top | | ||
Powered by e107 Forum System