Re: URI Request on LimeWire UltraPeers

Quoting neglacio <[email protected]> from ml.gnutella.dev-forum:
:When searching with a magnet which only contains the bitprint or SHA1
:hash of a file in LimeWire, it simply doesn't work. 
:That's because LimeWire UltraPeers ignore URI Requests. This isn't
:good IMO. 
:If there are any reasons why not, I would like to hear them and change
:my opinion :)

One of the reasons is LimeWire is stuck using small 64K-slot QRP tables,
where every other servent can use up to 2M-slot tables.  So adding SHA1
in the small tables fills their space and makes QRP-routing useless.

Another reason is that SHA1 lookups must not be done using the same logic
as keyword querying, due to the mere identifying nature of SHA1.  Queries
by SHA1 are only going to be successful through a dedicated structure, like
a DHT can offer.

LimeWire has implemented a DHT, but there are no public specifications about
it not about the policies for storing / querying, so it's hard to tell when
this is going to be widely deployed.

Meanwhile, no luck, SHA1 searches will not work.

Raphael