-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
You are arguing against a strawman. I never entirely discounted Oskar's argument, in fact, if you have been following devl you will note that Toad will shortly be starting work on a simulation of NGR to address some of the valid points Oskar made.
Ian.
On 6 Sep 2004, at 17:15, Newsbyte wrote:
Stop replying to something I have not said. I don't know what you are reading, but I have never said anywhere that anything should be mathematically proven. I advocate experimentation, and what I am saying is that none of it going on here.
Does "Absence of proof is not proof of absence" sound familiar? Anyway:
While you both have some interesting points, I must agree with oscar (even
though we disagree a lot on some other points). I think his basic point is,
that the devl cyclus that the project takes is crap, in the sense that it is
haphazard.
Now, I know You like to potray it as enivitable due to it's experimental
nature, but, frankly, that seems just an easy way out for explaining the
shortcomings. It has nothing to do with your perceived idea that oponents of
the current development are mathimatical perfectionists, as you always seem
to claim.
I agree with you (Ian) that in some more detailed points oskar made, he
might be a bit expecting to much of a Freenet in development (or out beta
stage, even); anonymity can never be absolute, and it all depends on what
effort (cost/benefit) it involves. (Hence my proposal for the
encryptionlayer of the datastorage that has, through the law, more legal
protection to some legal harrassement practises).
This has nothing to do with his basis premisse that things should be done in
a more structured way, and you are wrong in claiming using a more sientific
method can not be done, because of the intrinsic experimental nature of
freenet. A scientific endeavour is NOT the same as thinking out a complete
theory till all details fit, and then making a perfect tool in one stroke,
but it DOES mean, as Oskar hinted, that you have a process of developping a
theory, testing that theory, observing the experiment(s) and looking at how
it holds up to the theory, adapting the theory and/or the experiments and
trying it out again, etc. The process of falsification is paramount to the
scientific progress made, and NO other way has resulted in more succes then
that, including guessing and trying things out haphazardly.
Besides, I find it a bit puzzling you are defending the current
development-process so much, as it is actually your word(s), that we are
doing things too haphazardly. I remember clearly because I was reading your
post (several months ago) where you claimed exactly that, and I had to look
it up because I didn't know for sure what it meant. So, you acknowledged it
yourself, back then...so what changed? We are STILL doing it haphazardly. If
you were criticising it yourself back then, you should still criticise it
today, since nothing fundamentzal changed.
You can claim we have made progress in this or that area, but the bottom
line is, the network is still pretty much crappy. And basides, we maybe
would have made a lot more progress if we had done things a bit more
scientific and structured, instead of trying things out on a hinch or
'intuition'. The example you give is very interesting, but it's not a
shortage of novel ideas we have, it's adequately testing those ideas and
analysing and adapting them so they become novel *working* ideas.
And saying 'make a fork if you don't like it' is pretty weak too. It would
enveriably diffuse the effort and finances we have, and it's not like we
have an infinite amount of those. If a fork would be made, it would
enevitably weaken the two projects, maybe to a point of where both would
completely grind to a halt. That's why I wasn't interested in the offer of
joining the former try of a freenetfork with the 0.5.x branche, if you
remember that episode). I just thought it was a waste of time and effort.
The project does not genereate THAT much interest that it can succesfully
sustain 2 variants in development, IMHO. It does not mean the critique given
is unvalid, though, and, as I said, you once agreed to it yourself.
So why not try to remedie it? The alternative that oscar proposes is NOT
dissabandonning the project, as you seem to think, but rather revert it back
to when it was a in a more simple and managable stage, and try to
investigate every new feature or implementation thoroughly, see if it works,
and if it doesn't, analyse why and try to fix THAT, one step at a time. It's
NOT the approach we use today (and which has nothing to do with the work of
Toad, because I'm sure he's working hard, only he's trapped in the same
non-scientific system-of-guesses).
Another alternative (and maybe even necessary in oskars suggestion) would be
the testnetwork I have spoken of before. Your counteragrument is always 'it
didn't work the time we tried it before', but, frankly, if you remain
consistent and use the same criteria, you should have LONG ago abondend
ratelimiting and NGR, because we tried to make these work zillions of times
(and it still doesn't seem to work)... yet I never hear you complain about
that and use it as an argument you shouldn't try again.
And now you can complain and say that their is no proof that it is NGR that
isn't working, and that maybe it is a bug, but, once again, just the same
could be said about the first testnetwork. Maybe it just wasn't implemented
in the right way, or flexible anough, or buggy, etc. So, you seem rather
selective in your arguments and counterarguments when it suits your pet
(sub)projects, such as NGR, but don't find them valid when the same
arguments are used for promoting a testnetwork, which you mainly see as a
time of waste.
But I agree with Oskar on this: if we don't try to change the way we are
doing things, Freenet will become (and already is, in a way) a vaste waste
of effort, time and money, and the little progress we made and will make,
will be at a great cost to these three items, and far from being efficiently
used. It can't have escaped attention that more and more people, including
Freenetters, are becoming tired of the bad results of Freenet; so how long
can one, with wishfull thinking and going on the same way as we do, keep it
up anyway?
You always say:you can't be sure a testnetwork will help, yet you claim you
can't be sure of absolutes when it comes to anonmity, and you have no
problem with *that*. And frankly, if a dude is drowning in quicksand, it's a
bit absurd to ask if a branche that is offered will hold his weight and will
actually save him. Maybe it will and maybe it won't, but it still is the
best chance to go out the quicksand.
I don't understand your current reluctance to acknowledge there is a major
problem in the way development is done, and your abhorrence to try out some
radical new ways to improve it.
_______________________________________________ chat mailing list [EMAIL PROTECTED] http://news.gmane.org/gmane.network.freenet.general
-----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.4 (Darwin)
iD8DBQFBPXO5QtgxRWSmsqwRAsJ0AJ0d3REnUOmCqq8zTmYerbLX4b96HwCfTzLC kMIJNv0oTEZowHmrS7p6qZA= =tYFH -----END PGP SIGNATURE-----
_______________________________________________ chat mailing list [EMAIL PROTECTED] http://news.gmane.org/gmane.network.freenet.general
