What a superb program this is turning out to be, however I dont know if anyone has had the same trouble as me with it.

Specifically, a file with a limited number of sources, say 10 or less is in process of being downloaded. For some reason or other the download gets interupted by the usual rigurs of everyday life. When you open the program, the search it performs for the already started download is pretty lame.

Is it possible to have included and automatic search for started files that is comprehensive and also repeats at say 5 minute intervals for more sources.

I am aware that this will add a slight increase in traffic loading over the net, but surely if a file is able to be downloaded faster from more sources, up to the recommended 8 connections per file, then it frees the system off a lot faster.

At the moment if I get stalled downloads I have to reinitialise the search manually then "Oh my god look guys theres the file we should be downloading in that search window".

Just a thought, if I could program in Java Id add it myself, fraid im stuck on C and Basic at the moment.

automatic requeries were banned about a year or more ago, because they were discovered to be HORRIBLY expensive. The developers were very reluctant to give up this obviously desirable feature that is so easy to code, but the network's survival was being threatened.

Basically, IIRC, auto-requeries by hash were found to be crowding out tranfers.

Attempts were made to limit them to 1 per client per hour, but even those had to be banned.

Occaisonally someone on the Gnutella Developers Forum http://groups.yahoo.com/group/the_gdf/ will try to suggest another technical change to the network to allow automatic researching, but so far no feasible solution has been found.

Currently the network topology and modern clients have managed to get more generally effective downloads in other ways.

Sorry to have asked, I wasnt aware of the problem, ill stick to manual searches now i know whats happening.