(For those who don't want to re-read what's below, fish is worried that the gateway isn't anonymous because the initial request for a proxied resource will show up in web-logs, hence isn't anonymous)
Firstly, the weblogs do not reveal who browsed the url. What makes this gateway extention anonymous is that the freenet node that gets the data via another protocol is not the freenet node where the initial request is made. The first opportunity for the request to happen is on the node where the request runs out of htl. If that node is not running the gateway, then the first node in the path back to the reqesting node that is running the gateway fetches from the other protocol. Thus it is anonymous in the same manner that peekabooty is. You request the key for the first time, and the content is fetched from the other protocol by another nearly random node several hops away. Secondly, all links and images in html that is fetched by the gateway are indeed re-written, so they will be proxied and you don't get external links. Thirdly, yes, this will be slow. It will, in fact, be exactly as slow as freenet is. Browsing content this way will be roughly as painful as, well, browsing freent. The solution to this is to make freenet faster :) Lastly, it's nice to read some moderately positive reaction. Now if only you had commit privleges... :) On Fri, 08 Nov 2002 15:40:42 -0800 fish <fish at bovine.artificial-stupidity.net> wrote: > >Hi! This is actully something that I've thought of quite a few >times, so >it's great to see it implented :) (I also have a similar idea about >gnutella queries, but we'll go there another day - i've already >got two >freenet related projects right now, that's enough :-p) > >I do, however, have some issues with it based on your architecural >notes >and my reading of your diffs (I was unable to retrieve the source >from >that key). This, however, may be completly uninformed bullshit. >(I do >excel at that :-p) > >Firstly, you can't call it anonymous, since my initial request will >show >up in your web logs ;). If anonymity is actully important to you, > you >have to proxy through something else as well :-p. As a result, >I'm sure >many, many people would be upset if it was committed to freenet >CVS right >now, sorry ^_^. > >The other issue I see, is that freenet timeouts, running at about >5 >minutes for a reasonable node (from cruft's page) or longer for >a less >reasonable one (like mine ;)) will make web browsing pretty painful >for >files which aren't in freenet (multiple times). > >Are you rewriting links and image refrences? Obviously, since I >couldn't >get the source to do this, I can't see. If you havn't, i can send >you a >diff from another java project that I did that you can use for that, > and >if you did, then you probably should mention that in your arch description > >:). > >Given the lack of anonymity of this, I'd be significnatly more comfortable >if it ran on a different port or something with a big fat warning >the >first time you use it. > >like I said, cool idea, tho - i really am with you on this one, >despite >the reservations :). Anyhow, I was supposed to be at work two hours >ago, >so yeah, that's it :). "the american government lies to you. more >to >come." > > - fish > >p.s. being bitter around freenet is like being hot in a sun (no, > not >the computers) :-p. > >On Fri, 8 Nov 2002 thirty at hushmail.com wrote: > >> >> Hi there, >> >> I've written a a framework for allowing freenet to get data from >other >> protocols and insert it when it can't find that data in freenet. >It >> does this in a distributed anonymous manner. >> >> I've also written code that uses this framework to get data from >> http/ftp, thus allowing freenet to function as a distributed, >> anonymous, cacheing webproxy. >> >> Archetectural notes, diffs, and new files are at: >> >> SSK at S2orFS~u679fz02IHOkLM8tJJZoPAgM/gatekeeper/description.html >> >> With luck, someone with CVS access will commit this. If not, >I can be >> bitter :) >> >> I will try to post the architectural notes, diffs, and a zip with >the >> new files as follow-ups to this post. >> >> (note: this particular hushmail account has been created just >to post >> this stuff, and probably won't be around very long afterwards) >> >> >> >> >> Get your free encrypted email at https://www.hushmail.com >> >> _______________________________________________ >> devl mailing list >> devl at freenetproject.org >> http://hawk.freenetproject.org/cgi-bin/mailman/listinfo/devl >> > > > >_______________________________________________ >devl mailing list >devl at freenetproject.org >http://hawk.freenetproject.org/cgi-bin/mailman/listinfo/devl > > Get your free encrypted email at https://www.hushmail.com _______________________________________________ devl mailing list devl at freenetproject.org http://hawk.freenetproject.org/cgi-bin/mailman/listinfo/devl
