[hooligan3000]

is there no chance to update server.met and nodes.dat from emule-project.net @ first start?
my wish is a little bit more uncomplicated speed up and down.

[domdom75]

Hello all.


I don't understand some users who are talking about credits, who are happy to upload to fast users but not to slow ones, getting nothing in return to do it, etc...

But dudes, what is this kind of thinking? It is despicable. I'm from France, there I had in good old times a 512Kbits/128Kbits adsl line. Maximum upload was 16KB/s, I was uploading myself at 12, to not saturate the download rate.

Then I had 8Mbits/800Kbits, so my maximum upload was 90KB/s, I was uploading most of the time at 80KB/s, day and night. I never shut up my machine. I upload 24/24 and I also calculate for worldcommunitygrid.org No need to let a computer idle 99% of the time.

Now I'm on fiber. I get 500Mbits/200Mbits, so my maximum upload is 200/8= 25MB/s or 25.000KBs/s if you prefer. It is an enormous upload. At full speed of 25MB/s, I would be able to upload a 700Mb CD in 28 seconds!!! Your need a very fast hard disk drive, not cluterred, as otherwise your upload rate is limited by your HDD.

So today in fiber at 25MB/s or 15 years ago at 12KB/s, I was always uploading with joy to anyone and nowadays, since some months I'm on fiber, I DELIGHTED to upload at incredible speed to slow and very slow uploaders. I don't care about credits. What is this way of thinking some people are using? You put stuff to share or you download stuff and you wonder if some people are slow and are going to lower your credits? You use emule to see your credits increasing? Myself I'm using Emule, Bittorrent, to share stuff and upload the MOST I can.

What I'm proud of, is that I always, in the past, maintain a ratio of 2:1 MINIMUM and 3:1 most of the time on Emule and was targeting, with lot of difficulty 4:1. 4:1 means I download 1MB, I give back to others 4MB. So I give back 4 times more to the network than what I took. Now with my fiber line, my emule ratio exploded suddenly. I was near 4:1 since many years, now I'm on 15:1, but I only share rare stuff, I rarely upload more than 1MB/s on emule, sometimes 3MB/s but it is rare.

The main problem is people get stuff and goes away or remove files from they shared directory. So it is always the same who upload. If you just upload 1:1, with all the egoists ones, the general speed is not good, except on new stuff. So this is why I always tried to upload much more than I what I downloaded. So with my small upload, I kept to upload 3:1 aiming at 4:1.

Same on Bittorent. I was for years at 4:1. Now when I look at my ratio, I'm at 22:1. And it keeps growing as I upload with my fiber line much much much more than what I download. My total download since the last 5 years is 5.8TB. I uploaded 128TB. My Dumeter program indicates me I upload around 18TB per month. I could upload much more because I do not saturate my upload pipe.

On the contrary on Bittorrent (uTorrent) which is running 24/24 with Emule, I have at least 50 torrents running and I'm uploading 6-7MB/s continuously, sometimes 10-12MB/s. Why would I reduce my upload? Rather I upload a few or a lot, it doesn't change anything for me, so I open the pipe 100%. I'm happy and delighted to serve hundreds of people. I have old movies files, rare, 800Mb size, I already share them up to 450 Giga bytes!!! And I don't intend to remove them. I uploaded the file like 500 times completely to others.

Most people do not stay and leave, or they are still on the network but remove the files. This is the main problem for slow speed, but with my speed, I don't care sharing a file 6 months more or even more. I don't care about credits, I don't care about figures, I care about uploading and sharing, that is all. Because as someone already said earlier in this thread, if you are disappointed to upload to slow users and you ask for a new system which would allow you to avoid to upload to them, to be on a slow line is already a punishment, now to receive less from others BECAUSE you are on a slow line, is a double punishment and a complete, yes COMPLETE, stupid idea.

The only idea which must prevail is: UPLOADING to others and FAST !!!

Today the system of credits is a COMPLETE BULLSHIT !!!!!!!! If people no longer use eMule, it is because when you add a file to download it, you must wait, a lot of minutes, hours if not DAYS to begin to receive a single byte. Now you install a bittorrent program, you get a torrent file or a magnet link, the file begins to download on your hard disk drive in a couple of 1 or 2 min and you can get it in a couple of minutes or hours, where with the EXACT same people using eMule, you would have needed days to get it.

Some weeks ago I added a file, 3,5GB. 10 people on it, 1 seeds 9 at 5%. I succeeded to get a lot from the seeder after many days and being 24/24 connected, and thanks to my speed, I uploaded 91GB to the others at this day. On this file, it remains 3 missing chunks. One guy I uploaded to more than 3GB had 2 of those chunks, so logic would be he would upload them to me very fast as I uploaded to him quite a lot. I was on his queue list at 900, then 800n then 500, then 700, then 400, then 500, then 800, etc... like that for 15 days. Now he's gone and so the guy with the 100% file.

Same stuff on bittorrent, 2 minutes later I would have get those 3 missing chunks.

In january, I succeeded to finish to get a file, added......... 2,5 years ago !!!!!!!!!!!!!! YEAH, I it is true. I don't know why I never remove this file, it was on my list, forever in red, missing 50%. And suddenly, a miracle.

There are many files which are impossible to get on eMule, BECAUSE of this bullshit credit system. You get a message "Queue full" or you always see a number which never goes below, I don't know, say 500. So you have people you can see, they are here on the network, with the file 100% on their HDD, you want this file too, but the credit system is so a piece of shit, you can never get ANYTHING from them. I don't know why, but you can never get anything from some people, you are always high on the queue. So people stating you can get anything on eMule, yes you can find more rare stuff on eMule, I agree and don't know why exactly, but this rare stuff also mean only a couple of seeder, if not only one and with this bullshit credit system, this fucking queue, you are doomed. The file is here, you wait, you don't close your machine like many people and come back next day, no, even if you are connected like me 24/24, 7 days a week, 365 days per year, the eMule system of credits, of queue, etc... makes it impossible or close to impossible to get rare files. Or you can, sometimes, after 1, 2, 3, 6 month or a year or, like me, 2,5 years. But who is going to wait even 1 month???? The masses want the stuff now, not in 1 month.

And the exact same rare file on bittorrent, with the exact same unique seeder, I would have get the file little by little everyday or maybe in the day, because there is not bullshit credit system, no queue system. You come, you enter the room file and the download start and you upload immediately after you get the first Mega Bytes.

eMule has to be revamped COMPLETELY.

Where is the uPnP system? Do you realize how many Low Ids there are on the network? Who is enough smart to configure the UDP and TCP in the NAT of their their boxes? You install Bittorrent, the stuff is working. You install eMule, you need to open a configuration page on your browser to configure NAT UDP and TCP. Impossible masses can do that. So why isn't there no option in eMule by default? It is beyond imagination to have a program like eMule without this implemented.

Who need a Connection Button as already said earlier? Do you press a connect button when you launch Bittorrent?

There should be like on utorrent the possibility to immediately not share a file anymore. Right now you have to remove the file physically from its shared directory. It's really heavy. The UI must be completely reworked and redone.

Well there is a lot of work to do, I don't understand why nobody worked on emule since this time? This is a very well known program, almost everybody knows its name, it is open source and nobody even created a fork to improve it, remove the queue system, the credit system, create something closer than bittorrent. Unless if you remove the queue system, the program no longer work with other official eMule?

So, I'm still using it, to get some stuff I don't succeed to find anywhere else, but clearly I get 95% of my stuff on Bittorrent.

Last one thing about open source. Someone said earlier than people like and use emule because it is open source. Please, come on, be smart instead of thinking and writing such non sense. Microsoft Windows is on 90% of PC, Linux 2% something like that, Apple 5%. People, the masses do NOT give a shit for open source. Libreoffice is a wonderful program, but people keep installing Microsoft Word to write..... a simple letter. People still keep installing Adobe Photoshop to.... open images and remove red eye. You can do the same with the very complicated but open source The Gimp or for the masses and enough for 90% of people, the Paint.net program. Etc... etc.... People want a program FREE and a good one. That is all. If it is closed source or open source, they don't care, they are not geek like us, tuning the settings to get the most, they install the program, they launch it, so the program must work immediately, not long and obscure configuration stuff, you launch, it works.

So Emule will be used less and less if it remains like this. You have 40 or 50% of people who are Low Ids and they don't even know it for most of us, you have a terrible queue system which oblige you to wait, wait and wait. Why wait? You don't know but the program tells you you must wait before you can download, so you wait. And meanwhile millions of people using bittorrent do not wait at all and can already enjoy their stuff while yourself you're still waiting, waiting, you're still in the queue, waiting, 0 byte have been transfered but you can expect to get your first chunk soon. Others on Bittorrent are already entertaining their stuffs.

So my question is very simple: what people behind eMule are waiting for?

Myself I have absolutely no knowledge of coding, so I can't help at all and I regret it. I was coding years and years ago, I'm 40+, on Amstrad CPC 6128. And as said earlier, me too I was typing those endless pages of code in those wonderful magazines. The good old days.

Salutation.

[Some Support]

niRRity, on 25 March 2015 - 08:09 PM, said:

@Some Support
The answers to your questions are right in front of you. Look at the interfaces of successful P2P clients and websites and you will see what I mean. uTorrent has no connect button, no server list, no message tab and no nonsense.

There is no reason to compare it to another network if we have such a client made out of eMule already as I pointed out. And again I don't really like it.
I also don't think eMule needs to imitate uTorrent - what is the point? A user who prefers the quick download style will always prefer the uTorrent client as due to the nature of the network it will be faster. eMule will not be a alternative to that even if it changes to the causal users interace. On the other hand users who like being able to control and manipulate everything will be disappointed. One small change we probably really should do is stop making the server screenlist the default one on the start up, but instead the download screen.

We could add an easy interface mode which hides most buttons (and everything related to it), but the problem is eMule is a complex network and once something wents wrong it will be much more difficult to figure out what the reason is, without being able to see the log, not having the irc screen to get help, not seeing whether Kad has no nodes in its list, or being able to tell that there are no real server available.

fox88, on 25 March 2015 - 06:23 PM, said:

Is there anything wrong with this idea?

It's a good idea yes. Although we might have to put a expire date on the nodes.dat too, so it will actually not try using long gone IPs after a year or so.

[Zangune]

The main problem is users want to download fast and fixing the upload mechanism is the best you can do for this, if we want to talk about eMule, otherwise it is another program.
At the moment (and for years now) a big problem is eMule is not ready to work **out of the box** because of the outdated servers list and Kad nodes list, this is not easy to fix for the average user.
Considerations about the user interface, credit system, slot focus and similar may be interesting for a **small** number of users.
Read the forum, read help requests, join IRC to check it or trust me as I really know it about this.

[fox88]

domdom75, on 26 March 2015 - 02:15 AM, said:

eMule has to be revamped COMPLETELY.

I was good and sensible writing up to this phrase.
But your flaming after that phrase... let me quote again:

domdom75, on 26 March 2015 - 02:15 AM, said:

Please, come on, be smart instead of thinking and writing such nonsense.

That was your wording, not mine (nonsense was corrected).
It is better to choose sides using knowledge, leaving emotions in the background.

uPnP is there, but this is not the solution even it worked flawlessly. Because lots of users have no routers yet have low ID.

Blaming Connect button, CS and queues for all evil in the world - that would not be smart at all.

There are tons of mods: good mods and bad ones. Have you ever knew that? None of the mods delivered breakthrough performance so far.

PS. Adobe and M$ will be thrilled to know that their Photoshop, Windows and Word are now free.

[fox88]

Some Support, on 26 March 2015 - 12:55 PM, said:

Although we might have to put a expire date on the nodes.dat too, so it will actually not try using long gone IPs after a year or so.

Nodes.dat in 0.50a is of April 7, 2010
I did the test in January of 2013 and still managed to connect.
Of course you could add expiration date, but there is better option: more regular updates.

I also was writing about lack of progress display.
One of the most trivial ways might be to fill the list with nodes and update it while nodes get eliminated - very much like it happens with ordinary nodes when connection is lost.
When count reaches zero there should be a message box: no active nodes; check network connection or get new nodes.dat file.

[inman]

@domdom75:

I think you missed my point about uploading to slow uploaders entirely. I do not want to (further) "punish" people with slow lines at all.
Here are some examples of why the credit system does not work well (nowadays) to reiterate my point:
1. You are downloading a file and one of your co-downloaders has only one chunk (5% of the file size) that you need, and they are the only user sharing that chunk.
2. You have a fast upload speed, so in the current session you upload the remaining 95% to that user.
3. The co-downloader now has the file and unshares it, leaving me stranded on 95%.

Several years ago this scenario would not have happened, correct? Why? Because line speed was more evenly balanced. So - with help of the credit system - I probably would have got the 5% I required either before, or during, the time period it would have taken for that user to get the remaining 95% from me, or others sharing the file. You can sometimes get punished for uploading too fast. Remember, the faster you upload the faster they can unshare. If the user is incomplete then at least they have to share the partial download.

Which brings me onto my second example:
1. There is only one user sharing the complete file. It is a large file and they have a very high queue. Many users try to download the file, one with a fast upload speed.
2. The user with a fast upload speed proceeds to upload the same chunks to all co-downloders. The latter user has a better upload and shorter queue than the user sharing the complete file so they spread the same chunks to many users.
3 a). It takes so long to get the complete file that many co-downloaders cancel the file, wasting the user's good upload speed. Some eventually get the file but it is an extreme taste of patience.
OR
3 b ). The user sharing the complete file unshares it or goes offline for a long time, leaving many users stranded with the incomplete file.

In both examples, clearly I am uploading fast but having no real effect. I am being punished by either uploading too fast & consequently users unsharing too quickly; or I am uploading fast but spreading chunks that are likely going to be eventually cancelled, as they may never be completed.
If one of the co-downloaders has a significantly greater upload than the rest, with only 1 complete source, then often the credit system is largely ineffective.

This is not the only effect download and upload limitations has on the network. With Torrents you can always max out your DL and Emule you can always max out your UL. Most users (I know there is exceptions like yourself) want to max out their download, so they go and grab a few Torrents. As users with fast download capacities (who want to max it out but can't on Emule) also likely have fast upload capacities, and so Torrents are further benefited.
The reverse occurs on Emule, a user with a slow download capacity (say max 200 kb/s) won't notice much difference from use of either Emule or Torrents. However, a user with poor download speeds by today's standards also will have slow upload, and will more likely to use Emule than someone with fast UL and DL (as it makes little difference, assuming they want to max it out.) This further compiles a snail's pace throughout Ed2k and Kad.

You made some good points (and some similar ones) to mine but I just wanted to clear up that I don't want to further punish slow uploaders (as they are already punished with a slow line.) Remember that punishment applies to all uploaders, as the dedicated and fast uploader is punished because they want to see their files spread efficiently and at least eventually get their file! That is "punishment" for them if it doesn't happen, as per the examples I gave!

This post has been edited by inman: 26 March 2015 - 11:50 AM

[tHeWiZaRdOfDoS]

fox88, on 26 March 2015 - 12:34 PM, said:

There are tons of mods: good mods and bad ones. Have you ever knew that? None of the mods delivered breakthrough performance so far.

That - basically - is not true.
Remember the "golden" age of leecher mods? Exploiting the credit system (which has been made harder, lately) drastically improved the download rates.
That put aside, there is basically nothing a mod can actually do because the bottleneck is the ed2k network (by design) and the main user base (who use vanilla eMule). So you either have to advertise your mod like crazy or change the problem at its roots.
eMule now supports way higher upload speeds but still has a 3kB/s per slot target speed. That speed allows ~1 chunk to be transferred in 1h which is - in turn - the time needed for a fresh file download participant to be able to start sharing himself. This is a problem that should be targeted.
Higher upload speeds would also mean faster queue rotation which - in turn - means faster start of download process.

We all want to improve eMule and the simple way to improve the uploading and downloading experience is to let new users join the processes ASAP.

[inman]

Powershare is essential for releasers, otherwise a new/rare file spreads for too slowly. I would argue that is breakthrough.
Plus, MorphXT has better (or different at least) UPnP which works better for me (when I can't port forward as I am on someone else's Internet). Dreamule also solves lowID to lowID.

[Zangune]

inman, on 26 March 2015 - 01:23 PM, said:

Powershare is essential for releasers, otherwise a new/rare file spreads for too slowly. I would argue that is breakthrough.

It is prone to abuse.
We all should know how people abuses of the "*Release*" priority.

inman, on 26 March 2015 - 01:23 PM, said:

Dreamule also solves lowID to lowID.

This is not a proper solution.
In that case a low ID causes the same problems to the network without any drawback.

[inman]

@Zangune: I don't want to take this debate any further, but I would rather 'abuse of the powershare function' (not the same as abuse the 'release priority') than the below screen captures:

In both cases, the complete source to the file has a very large queue. A problem easily solved with a powershare. Furthermore, there is an auto powershare feature on MorphXT so, depending on the rarity and requests of the file, the user does not have to keep making any adjustments themselves.

As for Dreamule, I know nothing of the Mod other than lowid can transfer to lowid (providing both use Dreamule, I think.)
I would also argue, that all the dedicated Emule users I know make use of mods. They all have some essential features, some of which should be added to Emule, in my opinion. I myself use MorphXT which, when I don't have access to the router when I am in a hotel/at work and so forth, makes opening ports much easier. MorphXT gives me highID via UPnP when Emule does not. In this case, just a small tweak makes all the difference.

We can only voice our opinions in the hope that someone will listen and makes these changes for the next version!

[Zangune]

I believe we have to think to the network survivor, these two "features" can disrupt a network.

[inman]

And you can confirm that the latest Emule bans, for good, users of the Xunlei mod?

Or do I need "Dynamic Leecher Protection" provided by mods?

Xunlei users need to be BANNED by all clients. They bring more disruption than features provided by MorphXT, Xtreme etc.

Also, read carefully what I said, MorphXT has the auto feature which can prevent any powershare abuse. Maybe only allow this option in Emule? I fail to see how this is disruptive. Elaborating on your point would help.

[gigatoaster]

I remember The Black V8 or V7...that was a great mode

Concerning your discussion, I agree with most reasons why eMule lost interest. During 5 years most modders pushed their features and I regret some like slotfocus for instance were never implemented. That would have helped a lot the whole network.

[fox88]

inman, on 26 March 2015 - 02:47 PM, said:

Several years ago this scenario would not have happened, correct? Why? Because line speed was more evenly balanced. So - with help of the credit system - I probably would have got the 5% I required

You would or you would not get those 5% because of higher line speeds while influence of CS here again would be in the range from minor to negligeable.

[Some Support]

fox88, on 26 March 2015 - 10:50 AM, said:

I also was writing about lack of progress display.
One of the most trivial ways might be to fill the list with nodes and update it while nodes get eliminated - very much like it happens with ordinary nodes when connection is lost.

(aside of the other discussion) That's actually a good idea too. I'll put it on the list to look into for the next beta update.

[inman]

fox88, on 26 March 2015 - 07:43 PM, said:

inman, on 26 March 2015 - 02:47 PM, said:

Several years ago this scenario would not have happened, correct? Why? Because line speed was more evenly balanced. So - with help of the credit system - I probably would have got the 5% I required

You would or you would not get those 5% because of higher line speeds while influence of CS here again would be in the range from minor to negligeable.

The CS's influence is greater after longer periods of time. The majority of the CS's effect comes in to place with partial downloads & downloads (of popular files) took longer several years ago than today. I think this is an undisputed fact. As my upload was only 30 KB/s several years ago, and not today's ~1600 KB/S, I would like to think I received a ~9MB chunk before I uploaded the rest of the file to the only user who had that chunk shared (and before they swiftly unshared it.)
My point is, today uploads are sometimes too fast for the CS to have any desired & positive influence. It is a roll of the dice as to whether the complete sources you uploaded to chooses to keep & share the file. Of course, the CS also comes into place when you have 2 users who want to trade complete files. The chances of this are quite slim unless you DL a lot of similar files and are a heavy user.
And what about lowID? We haven't even mentioned the effect that has on the CS system. A complete source with highID must upload the file 200% if the 2 requests are from 2 lowIDs only. Therefore the highID is at a huge advantage in terms of gaining credits, are they not? If the same highID user happens to run into the same 2lowIDs, not only will the lowIDs' queues be shorter (as they can connect to less users), the highID should go up the lowIDs' queues faster. This is because of the credit scores the highID gained just because they were the only reachable user at the specific time they got 2 lowID requests.
I think there are too many scenarios where the CS's effect is greater influenced by other factors. Fix lowID and reward credits not only for sharing but simply for leaving Emule running. This evens the effect of credits gained for having greater upload speed and for having highID. This reward system is also less likely to negatively effect those who upload too fast and are left stranded with an incomplete file (thanks to those that immediately unshare.)

[fox88]

inman, on 26 March 2015 - 11:48 PM, said:

The CS's influence is greater after longer periods of time.

Finally it gets so great that blasts through the roof.
I wonder, how you invent all these ideas?
Nothing much changes in the long run: it was a minor factor, and it still would be a minor factor, where greatness value might have changes in order of a single percent.

This post has been edited by fox88: 26 March 2015 - 10:12 PM

[inman]

You still don't get it. Credit system is not the only factor involved, it is the constant of time (that Emule is left running and the files are shared as a consequence) combined with credits gained that gets you up queues. With that quote, I am NOT comparing the influence of the CS in the past and the present. I meant the CS takes GREATER effect with GREATER Emule runtime and share time.

However, downloads and uploads capacities have increased meaning share time can decrease for 'hitters and runners.' Plus, CS is more likely to have a greater influence with your co-downloaders, does it not? If downloads are faster (and filtered towards those with faster uploads rather than those with most credits) then there is less chance of trading chunks between your co-downloaders. Before fibre, there was much more trading of chunks between many users. Now 1-2 users can do all the work for 10 other users. I usually have shared the file 400-1000%, and many unshare, before my file finally completes. Again, why persist with the CS?

Look at what some Torrent sites do. Torrents keep up with the times. They know downloads are faster, and they know not enough is shared by the downloader in the DL time to compensate for this, and they know loads unshare when done. The CS is not as effective - as an incentive to share - as a reward just for keeping a file in share. Certain private Torrent sites reward you for leaving files in share (regardless of how much shared.) Maybe someone will try a reward system for Emule users for leaving files in share when complete, or maybe it will fall upon deaf ears.
...Or maybe I will get into another pointless argument and left with sarcy comments as responses.

This post has been edited by inman: 27 March 2015 - 12:48 AM

[fox88]

inman, on 27 March 2015 - 02:56 AM, said:

You still don't get it. Credit system is not the only factor involved, it is the constant of time (that Emule is left running and the files are shared as a consequence) combined with credits gained that gets you up queues.

Apparently, there are more variables. Just in case you read Docs long ago, there are articles: Credits, Rating and Score and Credit System
Now, do you get that we should mostly look at the difference made by the use of credits in these calculations, leaving all other factors aside?

inman, on 27 March 2015 - 02:56 AM, said:

I meant the CS takes GREATER effect with GREATER Emule runtime and share time.

Waiting time, used in calculations, is neither runtime nor share time. It is the time since last entering queue.
There is one good reason to give more to people with good credits: they probably will keep sharing and provide more sources instead of unsharing right away.