It's great to see that work is going on to improve the download success rate,
but I haven't seen anything about addressing another major problem with Gnutella:
in my (limited) experience, downloading requires "active management", meaning that
I need to keep an eye on what is happening. The result is that using Gnutella
eats up a lot of my time. Of course, the necessity of being at the computer
imposes a kind of discipline that may be desirable to discourage irresponsible
consumption of bandwidth, but for the long term I would like to see a "servent"
that is more of a "servant", that automates the process of obtaining specified
files or material pertaining to a specified subject, so that I can just give it
a "want list" and go away for hours or days while the agent works on its own.
Following LimeWire's design philosphy, such an agent could be designed to support
the network e.g. by searching at times when it finds that the Net is not
overburdened.

I believe this sort of thing is being developed in other realms of the Internet,
and ultimately an agent would search using multiple protocols or intelligently
choose the ones best suited to a given request.

Thats a positive feature request. But the reason why download succession is so slow on the network is b/c of broadcast traffic. A lot of queries get dropped and/or the host doesnt have enough bandwidth to support their uploads. They timeout or become extremely slow. Thats why we need to focus on avoiding/reducing broadcast traffic on the network first. Broadcast traffices comsumes ~50% of your over all bandwidth. That means the other half is divided for downloading and uploading bandwidth. For example a average modem user gets ~5kbps.

2.5kbps = gnutella network traffic (queries, ping, pongs)

2.5 kbps is left for upload and download

say you're downloading and someone is downloading from you (so you're uploading)

2.5 / 2 = 1.25 kbps

1.25 kbps = for either downloading or uploading. Most of the time it just timeouts b/c the network traffic comsumes more than 50% at times. Thats why we must implement LimeWire proposed Ping Pong Scheme and Query Cache. LimeWire developers said they're going to implement it into LimeWire within the next couple of weeks.

So when broadcast traffic is only comsuming 20% of your bandwidth at the most there will be a higher download succession rate b/c people will have more bandwidth to play with.

Furthermore, when broadcast traffic is reduced you will be able to connect to a larger horizon (an estimated measurement of how much data you are connected to within your network). Because you're only connected to a segment of the over all gnutella network. Resulting in more search results (higher download succession rate as in grouping wise), then maybe segmented downloading and other technologies could be implemented.

My point exactly! I have been using LimeWire for about half a year, and I have tried to discuss this issue earlier (http://www.gnutellaforums.com/showth...&threadid=2892). I even came up with two simple and easy to implement suggestions which would make working with LimeWire a little less troublesome - but nobody has commented on it yet.

LimeWire uses all the political correct ideas: its' free, its' based on an open standard, it's written in java, it's cross platform and it doesn't contain any spyware (though it isn't open source) - but it is also way behind other current file sharing programmes.

It is still struggling with basic connectivity problems while it's competitors offer more advanced search options, they are not limited to using the “old school” file information like file name, file extension and file size but will show you bit rates, length, image resolutions, codec format etc., they allow you to download the same file from multiple sources simultaneously, they can automatically resume your broken downloads and redo your searches...

I seriously doubt we will see much of this in LimeWire this year. It might be implemented at some point of time, but other file sharing programmes already have it, and there is no reason to believe they won't continue to work on making them even easier to work with.