On Mon, Jun 18, 2001 at 03:02:51PM -0700, Ray Heasman wrote:
> On Mon, Jun 18, 2001 at 01:19:27PM +0200, Oskar Sandberg wrote:
> > On Sun, Jun 17, 2001 at 06:49:57PM -0700, Ray Heasman wrote:
> > > Hi Oskar,
> > 
> > Hello Ray! How are Martha and the kids doing?
> 
> I'm sorry my standard salutation annoys you.

I'm just wondering from where we know one another...

< > 
> > > Um. Is that supposed to be an answer? How very Zen.
> > 
> > Yes, it is an answer. The objectives of this project go beyond creating
> > a distributed datastore (which can be done using hypercube routing, see
> > Plaxton etc), and we have many auxiliary requirements regarding the
> > survivability of data, resistance to localised attacks, and at least
> > some level of anomity for publishers and readers. It is clear that
> > highly mobile data is an essential ingredient in these.
> 
> Hmm.
> 
> 1) You make arbitrary statements. Flying can be done using flat surfaces.

Uh?

> 2) "It is clear" does not constitute an argument. Even were your statement
> always true, a naive design nullifies a potentially good concept.

Are you seriously going to argue that replication and mobility of data
is not necessary for survivability and resistance to attack on a network
of untrusted peers? And I did not say it was a sufficient condition.

> 3) If you want to make indefensible statements, do not be surprised when
> people demand you explain yourself.

If you want to jump in on a long term project with radical ideas, don't
be surprised when people expect you to understand the basic goals and
concepts of the project.

> 4) What is the point of having a developer mailing list if we can't discuss
> ideas in a useful fashion? I would like to use this list as a place to learn
> about freenet design details. I want to know how it works, not what the code
> does. If I'm wrong, tell me I'm wrong, but don't bother unless you tell me
> why I'm wrong. If nobody has the time, then so be it. Tell me and I'll
> leave you all alone.

You are welcome to discuss it, just as I am welcome to reply in the
manner I see fitting. I have not flamed you for posting, in fact, for
all the red you are seeing at the moment, I have not flamed you at all.

> 5) If you wish to make appeals to authority, please give full references. I
> have dealt with hypercube routing in parallel processing architectures, but
> am not sure of the context here. I found
> http://citeseer.nj.nec.com/did/19555 , but I can't find anything to do with
> hypercubes by Plaxton. Still, it looks like interesting reading.

C. Plaxton, R. Rajaraman, A Richa. Accessing Nearby copies of
replicated objects in a distributed environment. In Proc. of ACM SPAA,
June 1997.

> Furthermore, the current state of freenet is:
> 
> 1) Survivability of data. Maybe 15 minutes, for most of it.
> 
> 2) Resistance to localised attacks. What resistance? In the current network,
> requesting a file is an attack.
> 
> 3) Well done on the anonimity. Nothing I have suggested would have affected
> it, but anyway. Oh yes, please refer to (1) and (2). Your lost data is
> anonymous. Well done.
> 
> So far, the track record is not very good. Fair enough, Freenet is work in
> progress. It's not a very imposing track record to be used as a source of
> authority, though.

If you had payed attention to my posts you would know that you do not
need to tell me this. Just because the current system doesn't work does
not mean I can't express my conviction that your ideas on how to fix it
are wrong.

> > I may do such a test, but I find it rather futile, since:
> > a) It has no application to reality.
> > b) It won't work.
> 
> Then why do the test? Be consistent. Are you afraid of annoying Martha and
> the kids? Either my argument has merit, or not.

I would do such a test if there was sufficient interest from other
people who do not agree with me regarding the futility. Since I am
currently working on employed time for Freenet Project Inc. I obviously
have to consider more than just my own opinion when weighing in what to
spend my time on.

<snip blablabla> 

-- 
'DeCSS would be fine. Where is it?'
'Here,' Montag touched his head.
'Ah,' Granger smiled and nodded.

Oskar Sandberg
oskar at freenetproject.org

_______________________________________________
Devl mailing list
Devl at freenetproject.org
http://lists.freenetproject.org/mailman/listinfo/devl

Reply via email to