Unknown Gnutella-metadata in Limewire (like Bearshare?)
 

Hi there,

I wanna ask if Limewire did modify the gnutella protocoll, adding some proprietary and undocumented metadata?
For example Bearshare did this twice (QOS in QueryHit and "New-Bearsahre-Testing-TTL1") and I'm not very lucky about this development. I agree the gnutella protocoll needs various improvement and found a lot of articles on internet (e.g. at limewire.com).

I think undocumented features are slowing down the further development, you'll find my argumentation in another thread:
http://www.gnutellaforums.com/showth...=5261#post5261

I would be very happy about further information (any other known client adds proprietary information to the protocoll?) and serious discussions (technical and no emotional flames like xyz sucks plz). Thx!

Re: Unknown Gnutella-metadata in Limewire (like Bearshare?)
 

You can find more information and discussions at the Gnutella Developer Forum (http://groups.yahoo.com/group/the_gdf).

Thx, great informations, somthing I was allways searching for!

Now I only have to find a way to subscribe without using Yahoo Account.

LimeWire has not modified the Gnutella protocol.

In LimeWire 1.4, we began both using and supplying the "metadata" inserted within query reply packets, as many oher Gnutella cleints do.

This metadata, however, does not represent a "new" protocol. It rather takes advantage of the flexibility built into the existing protocol (whether intentionally, or by accident), although individual Gnutella developers differ slightly in their opinion on this.

The protocol, including metadata, is documented very well in the clip2 document found at:

http://dss.clip2.com/GnutellaProtocol04.pdf.

Thanx!

> The protocol, including metadata, is documented very well

I can't agree with that, let me ask something?
The data structure of the metadata inside the "Private Data" QueryHit is not documented, right? It seems every vendor tries to cook his own gnutella soup, or do I miss a point?
Sorry, I'll try to understand the details of what vendors like Bearshare/Limewire crypt inside the so called "Private Data"... is this top secret/proprietary?

The private area of the QueryHit is not documented, but (as far as I know) BearShare is the only client that uses it.

This private area, far from being debilitating to the network, allows developers to experiment with new types of information that may eventually benefit the whole network. You can think of it as a sort of trial area for new ideas, but most programs don't use it anyway. The standard for the QueryHits was developed in a somewhat seat-of-the-pants manner, but it works.

Also, if you look at the history of other protocols (such as HTTP), many of them were developed quite hastily (and even messily) in response to the pressing needs of the network.

Thumbs up guys!
This is how mature people discuss important issues.
Note that they all are not from the 'bearshare camp'?

It is very important to voice opinion, criticism and opposition, only then will the Gnutella network grow and still (hopefully) stay free. THAT WAS THE IDEA BEHIND IT.
If 'privat data' takes and pop-up ads take hold with gnutella clients, the free spirit behind the idea will become a memory only.
We need some programmers who can develop anti-privacy gnutella clients, to filter out all privat data, encrypted packeting and adware/pop-up data making their rounds.
I can see no other way.
Hundreds of thousands users on the gnutella net are just too much for the hungry minds of commerce. Talking alone does not stop greed.

Oh well, let's enjoy it while it lasts.

Also: Let's start more nets than just the
' GNUTELLA CONNECT/04 ' , which seems to be the only one I can see running.

JD

Hi afisk, I'm glad that limewire does not use proprietary data.

I can't agree with "trial area for new ideas", this is a very enthusiastic definition, hehe. It also isn't what Vinnie says, he sees e.g. Bearshare's QoS as his top secret client advantage without later public benefit to all developers (refering to public discussions, haven't subscribed to developer forums for longer). Also it is used far beyond trial, it's used in public Bearshare versions.

What you say about protocol development in the past: There have been public discussions and RFCs or steps to relase an RFC and all of those ideas have been well documentated (excluding HTML and this suboptimal protocol/client war).

Didn't we learn from the last years internet history:
For the interest of all users only an open standard will be good, proprietary is for single company interests.

I might new in gnutella, please tell me where I'm wrong! I'm willing to learn, maybe I didn't get the advantage of proprietary data in gnutella protocoll or have to accept it. My intention is to understand present and future of gnutella... as a programmer I got facinated by the idea of P2P.

Thx for reading. :)