We suffer from a complete lack of "corporate memory" here.

What happened to the whole file-splitting plan that would make this
sort of problem irrelavant since, if you did things right, most of the
separate blocks would collide and you can have as many "metadata
first blocks" as you want?

Now we move on to the recurring discussion on the infinite correct
methods of how to implement file splitting ...

Is anyone taking notes? ;]

Mike


> 
> Subject:
> 
> Re: [freenet-devl] the encryption key for all control docs on Freenet is..
> From:
> 
> Brandon <blanu at uts.cc.utexas.edu>
> Date:
> 
> Mon, 2 Apr 2001 21:43:34 -0500 (CDT)
> To:
> 
> devl at freenetproject.org
> 
> 
>> Think about the consequences of allowing two files on the network under
>> the same CHK but with differing payloads..  whoever inserts first decides
>> what metadata everybody is going to get forever.
> 
> 
> Yep. That's already been said by BC. It's pretty sucky.
> 
> I think the only solution is to not put metadata in files. This is also
> pretty sucky. When I get a file I'd like to get its metadata as well. But
> let's look at our options of bad stuff we can have:
> 
> 1) First inserter dictates metadata of file forever.
> 2) Multiple versions of a file split popularity and double space usage.
> 3) Files don't contain their own metadata.
> 
> Option 3 looks the least bad to me. Does anyone agree/disagree?
> 


_______________________________________________
Devl mailing list
Devl at freenetproject.org
http://lists.freenetproject.org/mailman/listinfo/devl

Reply via email to