mldonkey-users
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Mldonkey-users] eMule Mod to emulate mldonkey


From: Roland Arendes
Subject: Re: [Mldonkey-users] eMule Mod to emulate mldonkey
Date: Wed, 8 Jan 2003 23:16:16 +0100

hi Goswin Brederlow,

am Mittwoch, 8. Januar 2003 um 21:08 schriebst du:

>> people here giving upload mostly to emule. THE WHOLE NET IS NEARLY
>> COMPLETE EMULE!
GB>                new mldonkey:             143043 (70.65 %)
GB> 70% mldonkey clients. The net is MLDONKEY (+ emule fakes).
GB> Thats changed a lot in the last days it seems.

You don't believe this, do you? You really think that over a period of
3-5 days about 100'000 people switched from Windows easy eMule to
mldonkey?

Here's a quote from a previous mail from me:

--
These stats are not accurate anymore. AFAIK there is no good way to
recognize an overnet peer - and emules are changing md4's to the
mldonkey one to get the last slot (thats why you get this boost of
mldonkeys).

Even eDonkeys will be more in the future. I think the next version
will have the md4-Random-Generation as the default setting at startup.

So, forget about the brands. They're nonsense as of today.
--

I'm quite sure a great amount of clients you see as mldonkeys are
actually modded emules.

GB> But still: 20% of my clients get 40% of my uploads but only provide
GB> 20% downloads. Sharing seems OK but leeching is too high. And I have
GB> the minimum emule slots.

Hm. What if, as mldonkey already mentioned, we only have a problem
asking emules right? Perhaps because of this they're giving us less.
Just a assumption.

>> Remember the way lugdunum walked some time ago? In the first step, he
>> blocked mldonkey specially. Then, he blocked *aggressive* clients, not
>> mldonkey especially. And now, we're making the same damn mistake: We
>> exclude a specific client, because some people here don't understand the
>> net, can't read the stats or are no friends of emule. Sorry, but this is
>> BULLSHIT in my eyes.
GB> How do you differenciate between a sharing client and a leeching
GB> client? The only way at the moment is to see if its emule because
GB> emule has its internal scoring system that hurts everyone else.

The credit/scoring system is not the right way, and it probably has
bugs. But, is there any other effective method to get "hacked" clients
out of a closed sharing system?

GB> Also we are not talking about blocking emule but about distributing
GB> the downloads according to the uploads or according to seen or
GB> something. The emule slots are against emule. THIS IS NOT. Its just a
GB> reaction to reality.

I think there should be a more unspecific solution. And the actual
"solution" only rises countermeasures. And you're actually
experiencing them. They are meanwhile cloaking as mldonkeys, to get
out of this max_emule_slots and even more, to get even the last slot
(which is reserved for a mldonkey client).

>> No, I don't like the idea of favorising one specific client. Imagine what
>> happened if eMule would have introduced this feature before us. What would
>> be our reaction, if emule prefers other emules? (It does NOT)
GB> They did. Thats exactly what they are doing with their scoring system.
GB> And thats what the "allways give mldonkey a slot" does in a way.

I think the right way to answer this scoring system is to be
compatible and act in the same way - and the problem is gone.
Unfortunately I don't have enough time to learn ocaml. :-(
And we lost the author as a coder. The only people who hold up atm are
pango and Simon.

>> We cannot fake an emule, but we can use their features to be one. I
>> thought in the first place, mldonkey and emule would cooperate and the
>> world would be fine. Nope. :-(
GB> Go, get the source of emule and start coding the protocol for
GB> mldonkey. Don't say someone should put out the fire, put it out
GB> yourself.

I really *WOULD*, but I don't have the time to do so. If mldonkey was
coded in C, I'm quite sure many people can contribute, find bugs and
write hotfixes (f.e. me).

>> If this way of thinking from some people on this list doesn't change, it's
>> likely that some of the people on this list will switch to
>> linmule-development. I can't see this stupid bitching about emule anymore.
>> I really HOPE some people will understand THIS TIME.
GB> Looks like you didn't understand what is being discussed.

Look, I don't want to start a personal war, but - hm. I'm quite sure I
know what I'm talking of and what is beeing discussed.

GB> Its about coming up with a fair way to balance up and down. Give way
GB> more back to overnet if they want it for example. Also balancing
GB> between donkey and the other protocols if anyone uses them. It can't
GB> be that we leech 90% of stuff from gnutella and upload 90% to the
GB> donkey network. Same problem.

Ack. But, releasing 2.00 with only download capabilities from overnet
was wrong, too. Ok, Jed did not cooperate, but you have to understand
also his point of view.

But try to understand that there is no good way to recognize specific
clients. eMule is very versatile and will camouflage as an overnet
peer if needed, so giving more back to overnet won't last long
(they're already masking for mldonkeys).

Resistance is futile, they will adopt. ;)
The only way is to get a comfortable position with emule. But, as I
already mentioned, I can't put my thoughts into code.

-roland





reply via email to

[Prev in Thread] Current Thread [Next in Thread]