It doesn't just send out a query once and wait for responses to trickle in, but does something more active? I always figured the time was based on the HTL times the timeout (likely the HTL in minutes then), at which time the query has propagated to as many hosts as it ever will.

Of course, if that is what's happening and some hosts are slow in responding...but the timeout means a maximum additional minute.

Should I therefore wait 1 minute past when the lime stops spinning to be sure of having received all the results that are going to be received?

Or is it in fact more complex than that?

I do notice that some searches seem to quit unusually early ... which if Limewire does take an active role until that point would be cause for concern, since it means the horizon will effectively be closer for such searches, which will make it harder to find rare content in such cases. Add to that that if a single host has hundreds of matches you only seem to see a couple dozen. And add to that that the frigging ipod spammer's 50-odd bogus results seem to compete with real results for limited bandwidth, causing 50 or so legitimate results to get dropped on the floor...

Any suggestions on how to make results more balanced? I'd rather get one or two sources each for dozens of files (with the remaining sources still discoverable via the mesh, or Find Sources) than results like these, which are all too common:

169 Common File 1.xyz
44 Common File 2.xyz
36 Another Fricking Spam.wmv
12 Another file.xyz
1 Rare File 1.xyz
1 Rare File 2.xyz

and nothing else. :P

Of course if I know of some naming pattern to some of the rare files I can snag a lot of them with a narrower query, but some miscellaneous rare files might potentially elude searches for years, with daily searching!