Brandon wrote:
> 
> > How bout this idea:
> > We take the server in the URL as a *suggestion*.
> > The priority of who to contact first for the data is as follows:
> > 1. Localhost
> > 2. entries in ~/.freenet
> > 3. entries in /etc/freenet.conf or appropriate master file
> > 4. Suggested server in URL.
> > Fair enough? 
> Lovely.

I strongly disagree here, there is no difference between specifying a
server in a URL and just hosting the data using HTTP on that same
server, in fact, HTTP would probably be quicker since it lacks the
handshaking overhead.

If people want to browse data on Freenet then they MUST have access to a
Freenet node through one means or another.  Specifying a node in a web
page, for example, will completly defeat the point of Freenet, even with
your suggestion above, since it will tempt people to get lazy, and then
if a piece of information gets popular, the Freenet node specified will
grind to a halt, and the data will be unavailable - exactly the scenario
we hope to avoid!

When you specify a mailto: link in a URL do you specify a mail server? 
Of course not - that would be silly.  Similarily for someone wanting to
run a Freenet server - access to their own (or a local shared) freenet
node is as essential as access to a mail or DNS server.

To make my position clear: Freenet servers have *NO* place in URLs, even
as a fallback mechanism.

Ian.

_______________________________________________
Freenet-dev mailing list
Freenet-dev at lists.sourceforge.net
http://lists.sourceforge.net/mailman/listinfo/freenet-dev

Reply via email to