RE: [Blocks-development] Authentication and Ratings

[Paul K. <pa...@nk...>]

But who would want to download the contents of a whole file if the quality
is unknown? I know I wouldn't want to sit on a 56K modem, download a 10MB
file and only find out it's bogus. The rating couldn't be generated by the
party who added it in the 1st place, since, if they're creating bogus data,
the file could be marked, in the same way, as being just fine.

On the other hand, how can anyone tell if a file is valid (other than being
corrupted or completely bogus) by only 2-3 blocks of it. In the example of
an MP3, what's stopping a given bogus file from being "valid" for the 1st
few blocks and just garbage from that point on? Even if it's been hashed,
the hash itself could be generated invalidly.

Maybe I'm just missing something from Micheals explaination. I understand
the web-of-trust model, but I'm not understanding the hash model that he's
proposing fully.

For the hash to be valid at least one other (valid) signature must sign it?

-----Original Message-----
From: Erik Moeller [mailto:mo...@sc...]
Sent: Monday, July 31, 2000 1:51 PM
To: blo...@li...
Subject: Re: [Blocks-development] Authentication and Ratings


I cannot emphasize enough how important authentication and ratings 
are for a system like Blocks. Just think of the "Stop Napster" crap --
so far, not a single file-sharing system has come up with a solution 
to avoid intentionally mislabled or corrupted content.

I'm not sure I understand your web-of-trust concept completely. 
Here's my understanding of a possible implementation, correct me if 
I'm wrong. I'm using the term "announcement" instead of 
"advertisement".

- Every announcement includes 
		- the filename and location
		optionally:
		- a human-readable nickname
           + if it already exists, a new one with a 
             running number is created, e.g. Flaggovic_2
	 	- the public key belonging to this nickname
		- a short string rating the file's quality
		  and describing its content (should be done
		  using pre-defined labels and categories
	      to save bandwidth)

If this metadata exists, it is encrypted and signed with the public & 
private key.

Does this make sense? How do we make sure that an IP address is not 
associated with a nickname?

Regards,
Erik Moeller
-- 
Scientific Reviewer, Freelancer, Humanist -- Berlin/Germany
Phone: +49-30-45491008 - Web: <http://www.humanist.de/erik>
The Origins of Peace and Violence: <http://www.violence.de>
Please donate free food NOW: <http://www.thehungersite.com>

_______________________________________________
Blocks-development mailing list
Blo...@li...
http://lists.sourceforge.net/mailman/listinfo/blocks-development