On 02/11/15 23:47, Bob Ham wrote: >> Why do you think TCP still relies on packet loss to signal congestion, >> rather than ECN? Because there are thousands of buggy TCP firewalls >> which drop packets with the ECN bit set, or worse, home routers which >> crash completely when they see it! > > You would seriously argue that there should be one single implementation > of TCP? > > If there is a problem in the Freenet protocol, would you rather another > implementation highlighted it sooner or later? > >> We gain considerably from being able to change >> stuff without worrying too much about legacy incompatibilities > > You said you're gaining something. This implies some set of values > which determine this to be a "gain" rather than a "loss". What are > those values? What is the goal you're working towards that has made it > clear this is valued as a gain and not a loss? >
Personally, I think that we, as in people working on new internet architectures, should take a step back and consider the problem in a grander scale and with greater foresight. Learn from other new projects, and think about lessons that Freenet can teach other projects. At the moment we (the world) have teams working without much global co-ordination that would be required for this sort of large-scale engineering infrastructure project to succeed. I ranted about this on IRC a while back, and I can go into it again, but I'll stop here. I agree with Matthew that a second implementation is a waste of time, given the current status of the project. But my reasons for this position are guided by the above principles, whilst I would assume Matthew's reasons are different. Second implementations are useful in certain situations, but *not this situation*. But anyway, all of that is irrelevant compared to the MAIN point about a second implementation - who is going to do the work for it? Talk is cheap, show me the code man! However, I agree with Bob's original point that there should be more documentation. Again, I refer to the principles from the first paragraph. Currently all of these "new internet" architectures are silos that don't talk to each other much. Lots of modern projects are running into problems Freenet semi-solved several years ago, running into the same stumbling blocks, making the same compromises, blah blah blah. Documentation and constant comparison and awareness of other projects is a good thing worth spending a significant amount of time on. Fucking around with Java 1.6 gives you diminishing returns. Don't get trapped by the sunk costs fallacy. There are lots of common problems to be solved; we should take advantage of the FOSS nature of the project and try to figure out and define unified statements of these problems, that the whole world can then gather together to solve. Looking inward just for the interests of The Freenet Project won't ever scale, it's not taking advantage of the power of FOSS, and we might as well be writing proprietary software. Like a certain large company when they do/did throw-it-over-the-wall "open source" software releases. I have a shit ton more to say about this, but I'd prefer to have a constructive discussion than to debate the finer semantics of what I said, so forgive me that I will leave this thread and not worry too much about responding to any specific criticisms. You can catch me over private email or IRC. X -- GPG: 4096R/1318EFAC5FBBDBCE git://github.com/infinity0/pubkeys.git _______________________________________________ Devl mailing list Devl@freenetproject.org https://emu.freenetproject.org/cgi-bin/mailman/listinfo/devl
