To all the people who dont quite understand:

I think the next release should be kept off for at least a month. Just to let everything, ultrapeers and the whole ultrapeer topology, propagate before another big change is introduced into the network. Hopefully within the next few months other gnutella developes will start implementing this into their clients. Then gnutella will be the biggest and strongest living P2P to the day :]

I very much agree that there have been a bunch of really significant changes, and that we should let everything settle down for a bit.

All of the changes we're currently working on for the next release are much more low-impact than anything we've been doing recently -- there just little tweaks in several areas that will have a pretty nice effect nevertheless. Even though the changes are fairly minor, however, we probably won't be releasing again for a little while. We've gotten on this crazy release cycle in part to get user feedback as quickly as possible, and also because it's difficult to know what will happen with some of these changes on a distributed network like Gnutella, so we've just been throwing them out there when we've got them as close to perfected as we can.

Glad to hear you've received your t-shirt -- I think they're really nice too!

Sounds good.

waiting??
 

What changes are you holding off on?

Will file hashes be in the next release and what is the point of waiting? BS already has them. They will make swarming more powerfull. They wont hurt the network. I really hope you get this in your next release and get it out soon.

As for some type of auto-requery method what do you have in the works? This is SOOOO important as the user above pointed out. If it isnt implemented by the major client (LW and BS) other clients will do it and they wont be as concerned about network traffic. And I will use those clients because they realize the fact that I own a computer that is supposed to do work for me. I shouldnt have to sit in from of it and manually research the network, thats just silly. Please get this one out there very soon as well.

These are all issues that we're very much working on. Both HUGE and re-querying will definitely be in future releases, although I'm not going to make any promises about when they'll be out.

We're working as hard as we can!! These features will get implemented -- fear not.

As Adam just stated the next release will only be minor changes. I'm assuming fixes to the whole ultrapeer topology (better connectivity and so forth). I'm also guessing since there such a minor change, there is really no point in releasing it any time soon. I'm pretty sure they want to test and test until they get the perfect values to make this ultrapeer topology stable. So in other words, they probably wont have the things you were talking about in the next release.

I dont doubt that you will be adding them, I just feared that you might be delaying these features to the detriment of the network.

What methods are you considering for auto researching the network?

I actually don't want to go into the re-searching feature in too much detail. As far as withholding features, I don't think we would ever withhold features to the detriment of the network -- that's just not in our interest. It's just that features don't get magically implemented overnight.

If you're interested, all of the latest code is always available on our open source site (www.limewire.org).

Please dont get me wrong. I am well aware ofand very greatful for the hard work of the limewire team and all the open source developers. The tone of the above posts seemed to hint that given the major changes to the network recently you might not be as apt to implement more major changes in the near future.

Yeah, I realized I might have misinterpretted you a little bit -- I made it seem like we did not have the same urgency about improving the network as we did before.

This is definitely not at all true. We will just probably not be introducing any really major changes with our next release, although it's hard to imagine anything quite as major as the UltraPeer change. So, basically, the next major release will have a bunch of nice little changes (automated re-querying, for example), but it probably will not have HUGE. This is because it takes considerable time to implement and it probably won't be ready, but we also would like to make sure that the first major changes (UltraPeers, swarmed downloading) are working as close to perfectly as possible before putting something else on our plate. We basically just don't want to end up with a bunch of pretty good implementations of a bunch of really cool features. We'd rather have two really cool features implemented really well before adding another really cool feature with a first-generation implementation.

Ultimately, though, we're trying to get as high quaility a product as possible out to you guys as quickly as possible.