Re: [Gnu-arch-users] [FOSDEM substitute] "The Document" is Only a Hypothesis

[Andrew Suffield]

On Sat, Feb 25, 2006 at 04:34:27PM -0800, Thomas Lord wrote:
>   A document is always manifest in the real world in the form of a
>   signal.  There is no other physical manifestation of a document.
>   Books, recordings, IP packets -- all are carriers for signals 
>   moving and spreading through space and time.

That's some strange variation on quantum theory... in the less-cute
but more-technically-useful form, a document is data representing
information, a signal is a (potentially noisy) method for conveying
the data from one location to another, and the problem is to perform
this operation with acceptable accuracy. This is the information
theory model - which you'll need to solve the problem on the internet,
which is a large collection of slow, lossy, highly redundant channels.

Almost all existing attempts have been naive ones that assume
non-lossy perfectly reliable communication, and they've suffered for
it; notable exceptions are overlay networks like kademlia, gnutella,
and freenet. I'm pretty sure that I know where you're going with this,
and the 'right' answer is going to smell rather more like freenet (if
you think anonymity is necessary) or kademlia (if you think it isn't)
than it does like git. Notably, it'll be addressed purely by content
and not by location, and storage will not be controlled by authors.

The construction of such systems is a subject of current research. All
but the file sharing systems are stone age technology.

signature.asc
Description: Digital signature