[technician23]

Why not replacing Emule code with Imule code with the next release?
If you are not relaying on your code, but on having a cool ed2k system, dont you think that this time for a change has come and would give ed2k Links a great comback for the next 10 years? please, it is a small step for the coder, but a big step for mankind?

[hooligan3000]

and whats your reason to put in imule code?
it makes emule slower...whats cool on this?

[tHeWiZaRdOfDoS]

"A small step for the coder" - fantastic! I guess you're a coder? So I suggest to take that "small step" and release a modded anonymous eMule version!

[Stulle]

I would rather suggest a small step to a board that is concerned with the whole i2p bollocks but apparently that is not a popular view around here...

[technician23]

It would be a GREAT Comeback of the MULE system, in case the Emule code is replaced with the Imule code (as amule is quite dead as project, this would harmonize all three under this domain).
Emule has good downloadcounts still, but for germany and many EU countries it is not usable anymore.

Imule would change this, and the question is only, if the developer is stuck to his code or can introduce such a CHANGE.

I think he can, because he wanted CHANGE from Edonkey, and now we need CHANGE again, to be compatible with the environment development.

Kind of Housekeeping. This will come so or so. Emule can stay as it is for the next years, a new release with new features is unlikely and makes no sense.
There will be only in that case no action, if the developer has no spirit anymore. The chance is, to have the domain and channel used for the most downloaded p2p app with future, or, to not listen to the actual needs of the market. maybe the point is not now in his eyes, but he already knows, that he is in a phase of a moratorium. lets see, how long that will last and if he is still able to make decisions.

If there is no release in 2012 for emule, we all know: emule is dead.

http://i2p-projekt.de/imule/

This post has been edited by technician23: 21 April 2012 - 12:49 PM

[Meuh6879]

i2P idea ... well, i prefer the real P2P.
I don't want lose bandwidth for others clients... i2P is perhaps anonymous but if you must upload 3 times the upload for a single request, it's stupid (in a private area strategy).


(in my language, the i2P project have simply close the kademelia dev. because ... they don't understand how they can be use ... http://fr.wikipedia.org/wiki/I2P ... well, i'm a stupid guy but i don't trust stupid dev. that they can't read a code writted since 5 years )

This post has been edited by Meuh6879: 21 April 2012 - 09:44 PM

[Some Support]

technician23, on 21 April 2012 - 01:47 PM, said:

...

[DavidXanatos]

Does I2P have a good user basis?

Are there enough users using I2P to support filesharing at decent speeds?

I might add I2P support to my multi platform client.

but only when it would have major advantages over implementing an own darknet system.


also is there some specification what protocol changes to emule are used tu support i2p, i imagine teh adress length is not the same so you ned to change more than just the socket code.


David X.

This post has been edited by DavidXanatos: 25 April 2012 - 10:09 AM

[Stulle]

Will it also feature doing the laundry?

[DavidXanatos]

Stulle, on 25 April 2012 - 11:23 AM, said:

Will it also feature doing the laundry?

No.

but it might feature an OFF-System like bright net.

emule and bittorent and hoster downlaods seam to run quite fine and it is currently under a closed alpha testing.

David X.

[sonoro]

DavidXanatos.


what you call decent speeds? 10KB/s per slot?

Improve your upload in your account, and then make the question to other p2p system..


You told me you speed is 10kb in ed2k.

I hope you see that people want more than this speed, good provided they have internet connections at home because it limits you to gather a stupid amount of low-speed slot?

If you are imule slotfocus as standard upload believe will have a super success!

Your customer will have more because it's mix of networks you earn points not because there is no equal to her!

But it loses much in terms of upload speed ..

[DavidXanatos]

@sonoro
not wanting to go oftopic, but:

upload speed per slot is irrelevant, you will still get one chunk at a time.
And looking globally you wil while getting les speed per slot get mor slots so that a downlaoded amount averaged over a few hours will be the same no Mather what speed a slot head.

also the spot speed will be a setting so if you fill like you can sellect 100 kb/slot.

David X.

[Stulle]

I think the point he was trying to make is that dividing the upload among n networks means giving an average amount of Upload*time/n per network while potentially receiving unlimited amount of data from any connected network until your download bandwidth is saturated. This is pretty much the worst flaw in any multiple network client and the fact that you did not see this is a bit worrying. Besides, read the last sentence here:

Quote

One way to reduce that kind of bloat is described by the Unix philosophy: "Write programs that do one thing and do it well".

Now everyone say "Hi eMule!"

This post has been edited by Stulle: 25 April 2012 - 07:14 PM

[DavidXanatos]

Stulle, on 25 April 2012 - 08:09 PM, said:

I think the point he was trying to make is that dividing the upload among n networks means giving an average amount of Upload*time/n per network while potentially receiving unlimited amount of data from any connected network until your download bandwidth is saturated. This is pretty much the worst flaw in any multiple network client and the fact that you did not see this is a bit worrying.

Why should this be a flaw?
You divide your upload among n network, but at the same time you divide what you need to download among n network.
So emule gives 100% upload to ed2k and receives 100% download form ed2k.
A multi platform client for example gives 50% to ed2k and 50% to BT and receives accordingly.
As long as you maintain on a long run an ul/dl ratio of 1:1 it is competently irrelevant in how many networks you participate.

You know you can not download more that 100% no mather to how many networks you are connected.
Having more networks only decreases the time you need to reach 100%.

You know with a closed source client one could actually even enforce a 1:1 sharing ratio but i guess we wont be doing that.

Quote

Write programs that do one thing and do it well

Thats for amateurs.
Or to put it differently if you can write a program that does n things very well, do it!

I know yo think neo will be kind of a new shareaza, but it wont, already now it has full ed2k and BT support with all relevant fancy extensions, shareaza does not.

or to be more direct shareaza suxxx, but this does not mean that any multi platform client has to suck.

David X.

[fox88]

DavidXanatos, on 25 April 2012 - 10:35 PM, said:

Why should this be a flaw?

There is a possiblity of a flaw: in some cases it might work as a pump from one network to another. Besides, you'll multiply overhead traffic by the number of supported networks at least.
PS. I wonder, how well would the client behave with 500+ shared files in BT.

DavidXanatos, on 25 April 2012 - 10:35 PM, said:

Thats for amateurs.

People who created Unix were not amateurs. Calling them so would not make you a pro.

DavidXanatos, on 25 April 2012 - 10:35 PM, said:

Or to put it differently if you can write a program that does n things very well, do it!

That's a long road without a final point.
Look at the OSes: they have grown from kilobytes to multigigabyte monsters and still are far from perfection.

[DavidXanatos]

fox88, on 26 April 2012 - 07:56 AM, said:

There is a possiblity of a flaw: in some cases it might work as a pump from one network to another.

That would be bad design, knowing your ul/dl ration on each network can allow you to balance you uplaod so that you always maintain a simmilar ratio on booth networks.

fox88, on 26 April 2012 - 07:56 AM, said:

Besides, you'll multiply overhead traffic by the number of supported networks at least.

Thats wrong, the overhead depends on the amount of sources you have, emule limits this number to 500 afair, so if the new client would also limit the sources to a given amount you would just substitute ed2k sources for BT sources, but still have the same about of sources in total and those the same relative overhead.

Only for rare files with less sources you would increase the source count, but thats actually desired, thats the whole point, to make download of rare files faster.

Also its the year 2012 and not 2002, my DSL upload in 2012 is twice of my DSL downlaod in 2002, so overhead does not mathers that much nowadays.

fox88, on 26 April 2012 - 07:56 AM, said:

PS. I wonder, how well would the client behave with 500+ shared files in BT.

Well, BT is kind of a one way streat, you can automaticly share BT downlaods on ed2k, but not vice versa.
for sharing in BT you need to create a torrent and publish the magnet links someware and this is a manual job.

David X.

This post has been edited by DavidXanatos: 26 April 2012 - 10:35 AM

[fox88]

DavidXanatos, on 26 April 2012 - 01:33 PM, said:

That would be bad design, knowing your ul/dl ration on each network can allow you to balance you uplaod so that you always maintain a simmilar ratio on booth networks.

Does that imply you would restrict upload in network A if you downloaded the file from the network B?

DavidXanatos, on 26 April 2012 - 01:33 PM, said:

Thats wrong, the overhead depends on the amount of sources you have

You forget the file publishing overhead.

DavidXanatos, on 26 April 2012 - 01:33 PM, said:

emule limits this number to 500 afair

No it does not. Though there are restrictions imposed by servers and by KAD publishing rate.

DavidXanatos, on 26 April 2012 - 01:33 PM, said:

Also its the year 2012 and not 2002, my DSL upload you can automaticly share BT downlaods on ed2k, but not vice versa.

You should have complete file or magnet with ed2k hash for that, don't you?

[DavidXanatos]

fox88, on 26 April 2012 - 04:33 PM, said:

Does that imply you would restrict upload in network A if you downloaded the file from the network B?

I wouldn't do it on a per file basis but globaly according to a total ratio per each network.

So lets say that ma global raio is ul/dl 1:1 and the ratio in network A is 1:2 so in network B it would be 2:1, in such case i would restrict the upload to B, so that more upload would go to A...

Quote

You forget the file publishing overhead.

This one is usually negligible in comparation to the file request overhead.

Quote

No it does not. Though there are restrictions imposed by servers and by KAD publishing rate.

yes it does the variable in the settings is called "maxsourceperfile", in the ini "MaxSourcesPerFile" it is set to 400 by default, the hardcoded maximum limit is

#define	MAX_SOURCES_FILE_SOFT	750

Quote

You should have complete file or magnet with ed2k hash for that, don't you?

no, a typical magnet only has the torrent info hash.
those I can share the BT file on ED2K only after it was completed and hashed compeltly with the ED2K hashing scheme.

For bittorent this does not work as bittorent can habe any number of different torreents with different info hashes for one and teh same file, as it does not hashes only the file hash set but also some additional informations likr file name, etc....

And besides there are yet no keyword searched in BT possible so automatically creating random torrents would be compeltly pointless.

David X.

[Stulle]

Just for the record, MAX_SOURCES_FILE_SOFT does not limit the amount of sources, it just limits the number of sources eMule tries to gather. If that limit is exceeded eMule will no longer try to retrieve more sources from the network.

Your ideas may all be very noble but I think you are just rushing progress and will end up with a bloated pile of software that is beyond anything most users would want to use. The question is just when you will realize that and how you will react. IIRC, NeoMule had very few users and some complained about a variety of bugs and, in the end, the mod was left abandoned. Abandoned to the extent that the only attempts to resurrect the client have presumably consisted of asking me to take over.

I am not saying you are a bad coder. Far from that, actually. The problem is that one person alone simply can not work on a large project in his spare time and do things swiftly and without fail. One person is not even capable of recreating the variety of user input that is possible, especially not, if the same person tries to use the software in a specific way that suites him best but leaves opportunities to others to use it differently.

This post has been edited by Stulle: 26 April 2012 - 07:25 PM

[DavidXanatos]

I Abandoned NeoMule cause I had the new project, and it was clear to me that I wont work anymore on the old one, also I hadn't any more ideas for features that could be build in without profound changes to the underlying architecture, like for example would be needed to properly add BT support.

How bloated a software is depends on what the user gets to see, in agree that having 1000+ settings for every small bleep is not very user friendly, so in the new project I will attempt to automatize as much as possible so that the normal user interface will look simple and easy.

Also I have an associate that will do the GUI so I wont have to handle the user interaction and can concentrate 100% on the core functionality.

My design goals in NeoMule ware to invent new features and have the best personal download experience, so I never cared much about how user friendly it was for others.
With the new project one of the primary objectives is however to recruiter a large user basis to start an own KAD and this makes usability a major issue.

As about the "variety of user input" I wonder how emule devs solved that, other than just hoping for many useful crash dumps.


David X.