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

[Mike Jones]

>>  I vote for: all 
>>  
>>  I want a credit system.  I want slot assignment based on client
>>  type. And I want the standard upload/download ratio speed limiter
>>  from the original edonkey client.
>>  
>>  Have them all in mldonkey with a proper upload queue and let the
>>  user decide which she wants to use. (All completely separate of
>>  course and easily selectable.)

M> Yes, I want a no-upload option, and let the user (why not he ?) choose
M> if he wants to use it ! No, the credit system and the infinite queue
M> are not good (maybe they are good if you are sharing the lattest
M> windows softwares, and because emule users want them).

Uh, you miss the point completely. :)

M> It is not a question of satisfying the users. If we were not concerned
M> with that, we would simply not release mldonkey, we would keep it for us,
M> with connections to 250 servers and no-upload. We are concerned, but
M> we have to choose some "defaults" features that don't hurt other
M> users.

Hey, that is what I do!  I modify source to no upload, connect to all
servers in server list simultaneously, and auto-download all files with
"xxx" in the filename.  Giggle, giggle.

Oh wait.. no it isn't.  I've better things to do with my life.  And so
do most people.  (uh, that was sarcasm up there... I hope that was
obvious :) )

Who cares about idiots who sit in their basement thinking they are
"getting away" with something on a p2p network.  Bits and bytes .. they
have their own pitiful way of living, why focus on them all the time?

>>  Plus, consistently offer what every other client offers.

M> I agree that there should be a "monitoring other clients" team in the
M> project (someone wants to volunteer ? never have got a lot of success
M> with the question before... send a mail on the list every month (or
M> for important events) with such info !), that would try all other clients 
(emule,
M> overnet, etc) for all networks (edonkey, gnutella, ...) and see what's
M> new and what's good.

M> But I don't think we should add all the features: just the
M> best/interesting ones. It is not possible, and if only one user is
M> interested, he can probably hack it himself. 

>>  It is easy to modify source, easy to do lots of things, but most
>>  people are lazy and too busy to care about modifications if what
>>  they think is fair is already easily accessible to them in a
>>  default setup.

M> Easy ??? writing code takes some time, but debugging it takes a lot
M> more... introducing many new useless features makes a software
M> unmaintainable. We have to be careful with that.

My point was as follows:  It is easy to modify the source to become a
no upload client.  You can't hide from this.  Emule has no-upload
patches, edonkey client does, the new Hybrid overnet+edonkey client
will have one soon enough.

If you constantly think about the 0 uploader people you will forever
be frustrated.  Don't focus on the lowest of the low.  Instead, you
must lead.  Don't waste time battling cock roaches.

Most people upload.  Most people don't modify source - they click and
click and use whatever everyone else is using.  If you inform/remind
the user in a simple and direct way how networks suffer without sharing, and 
then
offer a fair "Share System" they will be content. They just want to ensure
they are not getting shafted more than their neighbour.  If other clients
have a certain system (upload/download ratio speed limiter, credit system,
etc) and mldonkey users can have the same, they won't feel like they are
getting shafted and will choose a system they think is fair and go with it.
Maybe people even like uploading.

Some people on emule forums complain about the credit system.  It is a more
complicated system than one can sum up in one sentence, but whatever
it is, and whatever it becomes, should be offered in mldonkey.   Maybe
it is the unlimited queue that is worse than the credit system?

>>  People on emule like to press a button -> set priority to release
>>  mode.  It is fun and "feels" good to the user.

M> Well, you should understand that we simply cannot add all the buttons
M> every user would want (I want a button with my name, and one to my web
M> page !!). There are already options that can be modified, and so on...

M> The main problem is not to give the users many fun ways to do the same
M> thing, but to add things that were not possible before (adding a new
M> shared directory from the GUI for example). This is our priority.

It isn't about buttons.  It is about ease of use while maintaining
that the end user feels like she is in control.

If the GUI had options in a chart:

Option Name | Default Value | Current Setting | Explanation of what
this thing did

Then it would be a step forward.

If it had a few simple settings that were just a front end for the
more complicated way of setting them:

Per File Priorities [Release/Normal/WhateverEmule has]
Share System [Credit/Speed Ratio/No Upload, I want to leech]

(uh, just kidding on that last one)

It would be an even bigger step forward.  You don't get rid of all the
advanced settings, you just make it easy to set them up.

BTW, Queue's don't have to be unlimited.  I don't remember saying
that.

>>  mldonkey needs to understand more from a user perspective, not
>>  always exclusively from a smart algorithm perspective. (although
>>  these are great too from an "under-the-hood" point of view) You
>>  will never come up with a single "best" solution that works for
>>  all.

M> That's the reason why I use mldonkey and Simon implements the features
M> I want :), when he is not introducing bugs somewhere else ... (I will
M> not forward my mail to him :) ) I'm a dummy user with wishes, and he is
M> my god.

I've never seen Simon post on here.  He showed up at just the right
time?  Jekyll/Hyde?  Who knows... Yes, whomever works on sources is
great.