Re: [Mldonkey-users] Tuning for search for sources with large download queues?

[Neil Sedger]

Matthias Witassek wrote:
> Neil Sedger schrieb:
>
> > It would have to be slightly more intelligent though - e.g. if a rare 
> > chunk popped up it should resume that file. And it should probably go 
> > looking for rare chunks first, and maybe look again for those chunks 
> > every x minutes...
>
>
> But how to identify a rare chunk?

(ml)donkey knows how long a file has been in its queue, and it'll know 
how long that file has been sitting at 93% waiting for the last 7% to 
appear.
Now if this time has been a while (over a month for some things in my 
list) that chunk should be marked as 'rare'. If it happens to spot this 
chunk somewhere it should make it a priority to get this chunk - taking 
up all available bandwidth - at the expense of any other downloads - 
which can probably be paused and resumed later.

Neil