gnunet-developers
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [GNUnet-developers] namespace


From: Jason Gilbert
Subject: Re: [GNUnet-developers] namespace
Date: Thu, 18 Jul 2002 10:43:06 -0500
User-agent: Mutt/1.2.5.1i

On Thu, Jul 18, 2002 at 01:20:03PM +1000, Glenn McGrath wrote:
> I think its important that a namespace be supported by GNUnet.
> 
> Searching does provided a way to locate documents, however isnt
> satisfactory for dynamic content.
> 
> e.g. One of my goals is to distribute debian metadata via GNUnet, i would
> like to create a signed index file that defines the packages available for
> a particular release. This index file would have a list of hash values for
> each packages metadata.
> 
> Users could search for this index file, but they have no way of knowing if
> the search result refers to the current index file, or an old index file.

Maybe include the publish date as a keyword and then publish a new one every
day.  When people search they can search w/ an AND of the current date.  Similar
to date based redirects in freenet.  I would think that it would be more
efficient than in freenet too since (assuming I understand everything
correctly) if the file hadn't changed the net effect would be to just add
another keyword index.  I think in freenet it copies all the content again.  I
could very well be wrong though.

> Web pages might be another example, im sure tis something that will come
> up again and again.
> 
> A namespace could be formated something like urn:gnunet:<host-id>:<path>
> the path would be controlled by whoever controls the host-id.
> 
> This urn could simply point to a gnunet hash which can be handled like
> normal gnunet content, when content changes the owner of the urn just
> changes the hash that the namespace points to and inserts the updated
> content like normal.

This sounds like Sub-Space Keys (SSK) in freenet is what you want.  The subspace
is tied to a private key so you aren't necessarily guaranteed of who the content
came from, just that it all is from the same place.

While hosting everything in gnunet would be good, I've been leaning towards a
post the index of files on a website strategy.  It seems like the only way you
could guarantee an update or generate the list on the fly.  From what I've read,
it doesn't seem like dynamic content (cgi,servlet,etc) would be publishable
through gnunet or any other anonymous distribution network.

To me, the problem is basically give me the "current" version of this file.  You
could put version keywords on the files, but how do you know which version is
the HEAD of the branch?

jason

-- 
"Only two things are infinite, the universe and human stupidity, and I'm not
sure about the former" - Albert Einstein



reply via email to

[Prev in Thread] Current Thread [Next in Thread]