>> >> Lovely. > > Stinky. > > Freenet URLs should not include the server. The server is a > setting. Setting it in URL will confuse users, and makes as much > sense as having your Web proxy in the URL. > > The URL is supposed to locate a piece of data. It is supposed to > decribe locater necessary to find that data. The node used to enter > Freenet has NOTHING to do with this. It doesn't even have to be an > Internet host.
It seems to me that it would be better to think of freenet requests as queries not URLs. A URL definitively tells a user agent where to go for a piece of information. As Oskar said, In freenet the entry node says nothing of the true location of the data. And ultimately you are not guaranteed to find a given piece of freenet media even if you know the key -- other issues come into play, specifically the volume of material in the local node and the nearby trusted network or nodes. I really think this is a bigger issue than how to construct a URL. If folks are going to treat freenet as a static resource for embedding media in the web, i.e. use URL's, then it makes perfect sense to specify the entry node. The consequences of this are that a small number of large nodes/hosts will likely be the entry points preferred for requests and inserts (if allowed). The end result of this is that folks can host all the media for their b2b or whathaveyou websites on someone else's machines, akamized no less. There seems to be a bit of schizophrenia here about whether freenet is essentially a static or dynamic system. If it is to be a static system certain things follow -- URLs with entry nodes make sense. A hierarchical key syntax makes sense. human readable keys make sense. The nature of freenet is changed as it becomes a free hosting service for web media no one would actually request if it weren't embedded in web documents. They Akamize their site using freenet and the images get 100,000+ requests/day propagating all over the place. <img src="freenet:genericorp/images/nav/corplogo.gif"> If freent is considered dynamic in nature a few things follow. URLs don't make as much sense as queries. human readable keys make a lot less sense than strong support for metadata, although the two aren't mutually exclusive. human initiated queries for things people are actually looking for are more common than static requests from web documents. The take-home here is that if Freenet emphasizes the static element over the dynamic it will be used much more as a free as in beer than free as in freedom Freenet. Rick >> _______________________________________________ >> Freenet-dev mailing list >> Freenet-dev at lists.sourceforge.net >> http://lists.sourceforge.net/mailman/listinfo/freenet-dev >> > > > _______________________________________________ > Freenet-dev mailing list > Freenet-dev at lists.sourceforge.net > http://lists.sourceforge.net/mailman/listinfo/freenet-dev > _______________________________________________ Freenet-dev mailing list Freenet-dev at lists.sourceforge.net http://lists.sourceforge.net/mailman/listinfo/freenet-dev
