> Yes, this would be a way to address multiple documents, but it would be > a very inefficient way to access versions of documents.
Yes, it depends on what information can be used. If version number or insertion date is a piece of public metadata then it is _exactly_ as efficient as file by name/versionnumber. Falling back to filtering based on CHKs is quite silly. I just mention it because, in the absence of any information other than the file name, you can still find what you're looking, albeit in a slow and annoying way, but it's better than nothing. Of course, if you have extra information such as the version number attached to the files then, again, it's _exactly_ the same as knowing the exact and unique key in terms of routing and efficiency. > I don't see the problem with what I proposed - It's a good way of doing versioning, all in all. It has the flaws of requiring updatable data (not that it won't work, but the more thing an idea relies on, the worse off it is until those things have been stress tested in the real world), and it allows key stealing. > But is spam really being combated effectively if it causes you to need > to keep downloading stuff until you get through the spam? I'm glad you asked! In fact it is, for two reasons. First, it's better to wade through the spam and get to what you want than to never be able to get to what you want. Second, for many keys there probably will not be very much spam, so it's not that big a deal. Lastly, if we can filter based on signatures then all you have to do is obnoxiously go through each piece of spam until you find the right thing, _once_. Then you can add the people that signed the good piece of data to your list of trusted people and filter based on their signatures in the future. This might not get rid of all the spam the first time, but the spam will decrease as you continue to use Freenet. Not only will you see less spam, but as you request less of it, it will fall out of the system. _______________________________________________ Freenet-dev mailing list Freenet-dev at lists.sourceforge.net http://lists.sourceforge.net/mailman/listinfo/freenet-dev
