Sounds good! What can we do to achieve/support this goal? *asking*

To be more specific, how can we build up an effective development community? Is the GDF as is works now an ideal place for protcol improvement, client implementation, new developer support and for learning from other p2p systems? Can we make the new Developer Forum more attractive, find more ideas?

Personally I think a friendly place to meet and a better support/documentaion for programmers new to Gnutella could be a step forward. The current V0.4 documenation does contain only a small part of what modern clients do and possible extensions (e.g. multisegmented doenloads, researches, superpeer, hashs, swarming) are very undocumented or hard to find.

Greets & Aloha, Moak

First off. Decide what you want, what is in it for you.

Then when you are committed, decide how you can best achieve what you want. If you want to do certain things alone. Do it alone. When finished, when you want to give it to the community, do so. If you want to work together on things, work together and agree who does what, when.

Keep your focus; and be open what you are in it for,

When you work together, you never get the best hand, the best people, all the resources. So when you can work together with angels and *******s, you are a great TEAM player.

When the current documentation is the thing that holds back. You might want to work on that. When you want to get to a common codebase work on that. But do it like Linux, Publish often, make obvious that something is happening. Make sure that the excitement is there.

I am not sure what GDF means But when you have doubts be clear what is wrong with it. Discuss it AND do not try to make it personal (hard enough).

An other thing; find a pet project that is COMPLETELY legit. That will help to have a lively Gnutella in a year's time.

Thanks,
have fun
Gerard

Hi Gerardm,
the GDF is a mailinglist of some Gnutella client developers (founded by Ex-Clip2), they coordinate client development and discuss new protocoll implementations.

To read online or to subscribe check this thread:
"Simple question on Gnutella"
http://www.gnutellaforums.com/showth...&threadid=4638

Hope it helps, Moak

First I want to thank Moak, he is helpfull.

I have read that people equate a new client to new ideas.

And that to work on an existing client requires that the existing people must allow you to work with them.

It was also said in a reaction to my "read about IBM's userv" that it does not include the Gnutella things like antispam and that is why it is no good. (obvious that it does not ) It is however a p-p appplication that companies might buy. It is also functionality that, given some posts, can be part of Gnutella as well.

I read that one of Gnutella's "strengths" is that nobody controls what and how. I am sceptical, that it leads to anarchy is something you may not mind. Thing is that every new implementation of a client will lead to network clutter by bug that need fixing.

And actually it is not what I proposed. When you have a framework client, with all basic functionality that is build modularly, well documented, everyone can build their client on top of that. When it is found that one module does not fit completely, it is time to work on THAT module again. That is the Linux way of developing. Linux is about the core, the kernel. After that it is up to you. Contribute back and it may become part of the package.

A personal question, I have spend some time on Gnutella, I think it has promise. But do you think that the way I contribute is helpfull ?

Thanks,
Have fun
Gerard

That kind of "anarchy" is why you can read this webpage or download your email. Open standards are why the Internet exists today and all those proprietary online services of the 80's have disappeared or adapted.

There are plenty of closed P2P protocols out there. Many are quite good - perhaps better than gnutella, but they're limited to the company that owns them. Most will be dead and gone in a few years time despite whatever techincal merit they may have.

A true P2P protocol must be open and client agnostic just like HTTP, FTP, TCP/IP, SMTP and all the myriad of other protocols the Internet runs on. Anything else is not only self-defeating, but pointless. If you want a yet another proprietary P2P network go and make one, but don't try and exploit gnutella for your own company's personal gain.

SRL,
Sorry you are ***-uming. Yes, I work for a company, I work in IT professionally. That does not mean that I do this for my company reasons and monetary profit! (the company I work for is not into P-P)

If you want to know WHY I am doing things, ask.

My question is, is my cooperation helpfull and I do not find that in your answer. I only get some !@#$% on anarchy.

Software development, protocol development is about evolution. It is brutal. When Gnutella is only about sharing music. It will die. If you want Gnutella in a year's time it must have the ability to share whatever and may have the option of anti spam. AND IT MUST BE FREE. It must have a living flexible protocol that allows for whatever you all come up with.

Thanks,
have fun
Gerard

As the others already mentioned there are already a couple of P2P clients out there used comercially.

One good web site on your discussion is

http://www.peertal.com/

which is a page for "Business and Technology of Distributed Network Systems" in their own words. There you find a lot of interesting news what is going on in P2P development in the business sector.

I think one has to mention Groove http://www.groove.net, which I consider as one of the leading P2P products that are developed at the moment. Also the web page there gives qu
ite a lot of interesting information on the subject also numbers how much money could be saved with P2P technology in company networks.

I hope this is what you look for and that I could give you some new information ...

If I am wrong in some things please correct me =)

In my personal opinion I think there are quite some indications that P2P networks will survive. If they will be based on the Gnutella protocol as it is today? I doubt. I think the tremendous number of P2P network users that came up in such a short time shows how important and useful these networks may be/are. I say "may be" 'cause file sharing (Mp3, Movies, ...) is shurely not the long time goal and the only thing you can be done with P2P networks.

I am happy to discuss that a little bit more here, also the business models that can be though of, since once in a while people will like to earn money or not loose money ... not only Napster showed this (of course there have been other reasons - I don't want to discuss that here).
.
Hope I could help & Cheers

Felix

Sorry if I sounded harsh, but I've seen other authors try and "take over" the gnutella network. The excuse was always that the "other" peers were buggy and flawed and everyone should just be happy to accept whatever proprietary twist they decide to add.

There may very well be money to be made with P2P, but I don't think anyone should expect to make a fortune from gnutella. It's a grungy, imperfect, mutt of a protocol that has survived were other, sometimes better, protocols have failed largely because *anyone* can write a peer.

If it's to have a future there needs to be some agreed upon standards developed. This has been true for every other Internet protocol, so why should P2P be any different? What's needed is something like a P2P RFC.


I didn't mean you literally - more as a general warning to those that would try, but you were the one who seemed to be saying we should defer the development of gnutella to the beneficence of some company for it to be profitable.

What you seem to miss is no one besides those companies needs Gnutella to be profitable. Corporate sponsorship would hardly have the interests of its current users at heart. The truth is they don't care if you or any company never make one cent off of Gnutella. Actually why should they? It's only bad for them.

For what it is, Gnutella's better off being a garage protocol slapped together by CS students on a weekend lark. It will grow, sometimes painfully, in fits and starts, but in ways the users want - because some of those users *are* the developers, and someone will always be willing to make a peer that's better than the rest just because they want one. The same is true of Linux. It was a cool toy long before it became a "Corporate IT solution", and to be honest that's all it really needed to be. The companies embracing it are only benefiting from the open free-for-all ethic that created it. Thankfully a few even seem willing to give back for now.

Is your co-operation helpful? It depends what you want to do, what you plan to give back, and what you might take away. Too many developers look at Gnutella's popularity and dollar signs flash in their eyes. They don't care about making it better for its users - only finding a way to make it better serve them.

Honestly, your questions seem a bit, well, anachronistic. Surely you must know gnutella has *never* only shared music - it's has always been about as generic a file sharing protocol as it gets. I actually filter MP3's out of my results because I'm looking for other stuff. Also, spam turned out to be kind of a non-issue. It hasn't been much of a problem for a over a year. Currently there's much more pressing things facing the network.

For example, it's getting very hard to find files that aren't very popular and well distributed. This is a protocol flaw, and other P2P protocols have already overcome it. It's just a matter of getting all the current developers to agree on the same implementation.

Maybe you should lurk around these boards for awhile, or read up on Gnutella's current development. Things have changed drastically since a year ago, and some of your questions seem points long moot.



That's nice, but it already is and does.

Actually I may have completely missed what you were after. So I'll ask, what is it you'd like to accomplish with Gnutella? What do you expect it to do for you that some other P2P protocol wouldn't do better?

There's already been a good deal of development in alternate P2P uses with far more commercial potential (swarming comes to mind), so why Gnutella? It's not an attack, it's a valid question - what use does any company have for this mutt? It may even be the *worst* of all the P2P protocols, but people use it because it half-way works for what they want.

I'd expect a company to choose something better, and it always makes me suspicious when anyone talks of its commercial potential. Either they don't know what been going on in P2P, or they've really set their sights on milking its userbase (which is gnutella's only real asset).

I am happy. People actuallly try to read what you say. SRL's answer is a case in point.

I do not want to take over, better still I will not take over as I do not have that in mind. I want to do filesharing on a rather massive scale. I am in the process of creating a community around a database application (GPLish I do add ). This database application is about Taxonomy and pictures of plants.. Really dull when you do not like that

I need a P-P software package that has multiple uses and can be configured as such. Yes, it must have the current Gnutella functiality, Yes I need people to interrogate a file server to get the name of the file they want. Yes it would be best that it is only asked to a subset of all Gnutella clients. Yes, the files are all over the world so a limited horizon is

This is a project that costs nothing and can benefit a whole community. When it works all kinds of people interested in plants will want this. That is a hobby/scientific community.

This is what I am after. What you get is a "window" application. That is a project that you put as an advert for all to see. What I am good at is badger people to cooperate. To define the protocol better, To get people to talk about the core functionality. Cause when I am considered to be helpfull I will follow what happens and comment.

At this moment I can not use LimeWire because of the spy-ware.

SRL, You are right, that a RFC mechanism is needed. Isn't that what a forum can be used for? Have one for chit chat, have one for genuine TECHNICAL stuff.

SRL, the intention was that you would ask, so there..

Linux is not corporate. Companies benefit themselves by contributing to the community. IBM eg sells more mainframes because of it. The corporations that do Gnutella benefit from a mature protocol. They will not make their money from the client. They will make their money from the add-ons. LimeWire shoots itself in the foot with their spy-ware. I CANNOT use it because of it. My community will not stand for it.

So my contributions will be in being professional IT and commenting on what I see. Personally I write RPG, CL, Synon/2 some Cobol some MS/Access. So I will not add to the code.

When I write things that are old hat, call me ill informed and I look for better info next time, I learn..

Why Gnutella; it is well known. It is free. And I choose it for a community that has at this time little money to spend. The basic functionality is there in Gnutella clients. And there is this vibrant community.

Thanks,
have fun
Gerard

Ok, I understand better now. ;-)

You might want to look at this...

http://sourceforge.net/projects/gift/

It's an open source reverse engineering of the FastTrack protocol (used by Kazaa, Morpehus, etc). You can think of FastTrack as being a sort of second generation Gnutella that's actually solved most of the major weaknesses of the original protocol.

The only problem was, until giFT, it was completely proprietary. Once giFT opened it up, FastTrack actually changed the protocol to shut it out again (and in the process destroyed its server-free independence). However, rather than abandon giFT, the authors decided to create an open parallel protocol. If it gets support and good peers written for it, it could beat Gnutella hands down. For an independent project that doesn't depend on the existing network for files, it would be a much more functional choice.

Gnutella is actually moving in this direction, but it has to fight the demons of backwards compatibility (which giFT is free of). Actually I should point out LimeWire is open source, so spyware or not, you could always fork the code and make a "clean" version (in fact the CVS source tree is spyware free already), but gnutella won't work as well for what you describe. By the time it does (think super-peers), it'll look a lot more like giFT/FastTrack.