On Mon, Apr 16, 2001 at 09:23:11PM -0500, Owen Williams wrote:
> On 17 Apr 2001 02:14:15 +0200, Oskar Sandberg wrote:
>
> > If you are not going to delete data in that situation -
> > what are you going to do with it?
> >
>
> My problem with this philosophy is that it is misleading. If I have a
> node that needs to delete X bytes of data, it doesn't mean all of
> freenet is totally overloaded. It just means my own node is out of
> space. The question is, how does my node find another node that can
> spare those X bytes for my file? If my node contains the only copy of
> the data, it is still deleted needlessly.
You are confusing global and local loss of data. The idea is that the
algorithm spreads the data in such a manner as to keep it around as long
as possible, even though it is constantly being removed from individual
nodes. This may or may not be the case at the moment, due to the fact that
our system is far from perfect. Nobody here has been attacked for well
reasoned ideas as to how we could improve the networks behavior in this
regard, and many such ideas are indeed in the air (selective caching,
reinserting on removing, using the routing table mean to weigh up objects
that seem to belong to the node, etc etc) - but people who come in
claiming we suck for not caching infinite amounts of data on finite
resources get what they deserve.
> btw, when I do an insert of a
> file does it get uploaded to anywhere except my own node?
RTFD for fucks sake...
--
'DeCSS would be fine. Where is it?'
'Here,' Montag touched his head.
'Ah,' Granger smiled and nodded.
Oskar Sandberg
[EMAIL PROTECTED]
_______________________________________________
Devl mailing list
[EMAIL PROTECTED]
http://lists.freenetproject.org/mailman/listinfo/devl
