Depends on what you're trying to download. I was sharing 200 DivX-Movies for a couple of days and I received 80,000 Upload Requests within 24 hours but within that time I could not upload more than 2 movies.

Now I'm sharing 5,000 mp3s and I'm receiving about 10% of that number in 24 hours, while I can upload roughly 200.

You can see the difference: divx was 40,000reqs to 1ups, mp3 25reqs to 1ups. It's no problem to download mp3s these days, but downloading divX movies from gnutella is really painful (although I've already done so: I downloaded harry potter within 24 hours using mutella)

Freeloaders and modem users were a problem for gnutella just a year ago, but today modem users are becoming rare and people with fast connections who are on the net 24/7 don't have any reason for freeloading. Even if 20% of the gnutella users were freeloaders or modem users (or both) that would not matter so much.

Such postings make me believe I waste my time in making Gnutella better. Perhaps you want to live in anarchismus, personally I belive most people are too selfish and gnutella is too anonymous and selfish today

Mr. Jones- Modem users have a long way to go before they become rare, so this problem will not just disappear on its own. Modem users represent a much greater percentage than 20% of Gnutella users.

I use a Modem, and you'd be surprised how much a modem user can share if they want to. I have several popular movie soundtracks, and I have upload requests almost constantly. As long as I am not downloading anything myself, I can upload several lmedium-size mp3 files withing an hour. To tell the honest truth, if I were to keep an actual record, there would be more bandwidth used uploading than bandwidth used for downloading. I've got over 3 GBs of stuff, and most of it's popular. Never let it be said that all modem users are freeloaders. Judge people (or <b>nodes</b> for you socially impaired) by their character, not by their class.

John does constantly post this unfounded statement about freeloading. Currently we have a statistics from Oct 2000 [1], which says 70% of Gnutella users share no files. Is any other information available?

[1] Free riding on Gnutella - http://www.firstmonday.org/issues/is...dar/index.html

Oh yes that article. That was written over a year ago and gnutella at the time wasnt in really good shape. I mean some news sites have posted stories about gnutella dieing within a few months. There wasnt as many people using back then and the gnutella servents out at that time werent easy to use.

Its a great piece of gnutella history but about thats it. I dont think it would be wise to make decisions about gnutella today based on a year old report using year old statistics. If someone comes out with a new report then that would be useful and that would be better to base ideas off of.

On anti-freeloading i still dont agree with it. The ideas behind it all want to restrict, place limitations that is not what P2P and file sharing is about. Look at napster they never had any anti-freeloading rectrictions and they had one of the highest sharing rates of all. So increaseing the number of people that share can be done without elborate plans that emphasize restricting users. Also as i said in other posts in this thread i think people will share more when gnutella itself is improved and is more efficent.

Sephiroth

Obsviously, you didn't read any of our posts. We are talking about network improvements, and have in NO way schemed to restrict users.

I was commenting on anti-freeloading in general. Realize that when i say anti-freeloading i mean features that do mean to add restriction or other elborate plans to force users to share. Other more logical ideas i consider performance and uploading features that encourage sharing but doesnt force you.

However ill be more than happy to comment on what was recently posted.

I agree with all the points that you and fist brought up.

The problem is that gnutella hasnt really intregrated modem users very well however with things like supernodes im happy to say thats changing. Since around 2-3 kilobytes which is the average i use are used on just keeping 4 host connections that only leaves about 2-3 kilobytes free to download and upload.

I expect supernodes and swarming to drastically change this.

Im on 56k modem and im sharing about a 1000 files. Mostly low file size files like jpg wallpapers and some midi files.

Thank you for responding again, Sephiroth.
:-)

I do agree that the 'current' statisitics report is outdated. We need an updated statistic report to accurately assess the network's current state and what needs to be done.

How would one go about finding new statistics. Do you think this would be a large unertaking requiring many people's involvement? Could we simply mimic the methods used to obtain the older statistics?

I'm working on new Gnutella statistics.
It's a side project, I let you know as soon I have something. Meanwhile, what makes you think freeloading has changed or has become less. It could be even worse. Sephiroth, currently we have 1 year old data, okay, but everything else are presumptions based on what? Seeing the high amount of busy slots, still, I wonder what makes you think things have changed to the good?

Do you know alot of statistics? I took a semester of it and you got to have a representative sample using a random method in order for it to do any good. If not then it wont accurately measure the network and you just wasted a whole lotta time. Someone here is allready posted a message saying that they are doing a study on this very thing.

Anyways the network at the time those stats where take wasnt in good shape. For one gnutella wasnt easy to use, the orginal servents were still out there and probably most importantly web sites searches weren't being blocked. Which im assuming that you dont know that it was these gnutella web search engines that were the major freeloaders and not gnutella users themselves.. Im talking about web searches on gnutella that would leech off everyone. There are some gnutella searches out there now and thats different. They only search users who let them and so its completely different than back then.

Now thats all changed the network is at least 5 times bigger than it was at that time, web sites are blocked by a majority of gnutella programs so thats taken care of, .56 are also blocked because they were also hurting the network i think it was becuase it was generating too much traffic, and the gnutella programs of today are a whole helluva lot easier to use than they were back then and that attrachs more users.

You cannot measure freeloading on the amount of busies you get. That is pretty funny.. Why because they have to be sharing for you to download off of to get the busy in the first place haha. The idea that less freeloading will make less busies although seems like it work i believe its a myth. Even on programs that do have excessive anti-freeloading features busy downloads are still common. Theres a big difference between Encourageing sharing and Forcing Sharing. Encouarging sharing is a good idea but forcing sharing is not.