Emule Future Development
#41
Posted 09 February 2016 - 11:32 PM
We all have our opinions here but the trick is suitable compromise that we all can accept so we can move forward. As David wants to abolish CS where Wizard wants to revise CS. At that point nothing should be excluded from project until better solution is on the table for replacement. But for now we all agree that CS is just another myth as its false presumption that uploaders will show how be rewarded later period of time for their contribution to the network under global file sharing scheme. As CS now currently works to some extent but only for file trading scheme for selected pool of users with similar contents of interest or during active rapid chunk exchange but not very effective under file sharing scheme. The chances of ever seeing that known client is remote, the chances of finding contents to download from that client is very remote. So no one believes in it as it wasn't designed for global transactions thus holds no merit as proven not to be true since day one.
Far as QS goes it is needed as it should be designed for small and large scale solution in mind since its dynamic. Where small number of files being uploaded compared to large scale of uploads facing around 100+ sometime during initial release. Thus making 100+ active upload is not actually wise so QS and QR is needed. But at this point the chunk distribution system should kick in which would keep track of uploaded chunks to which client into the swarm. So those chunks that already downloaded can be redistributed from them to others. Where the main client can focus on equally distributing reset of it in nonlinear manner. Thus QS and QR goes hand to hand for it and mixing CS to it make it little complex. Nothing is easy but it all can be translated and be done.
#42
Posted 10 February 2016 - 05:12 AM
Stulle @ Dec 7 2005, 06:16 PM
#43
Posted 10 February 2016 - 08:51 AM
I just noticed that toll now no one mentioned the elephant in the room:
Should eMule development be continued based on the current MfC framework, a.k.a. compiling only for windows.
Or should the project be ported to for example Qt, such that it can run also on *nix'es?
IMHO it should.
David X.
it is the first client to be able to download form multiple networks the same file.
NL provides the first fully decentralized scalable torrent and DDL keyword search,
it implements an own novel anonymous file sharing network, providing anonymity and deniability to its users,
as well as many other new features.
It is written in C++ with Qt and is available for Windows, Linux and MacOS.
#44
Posted 10 February 2016 - 09:41 AM
Set up a new dev crew.
Keep up maintenance releases.
Separate core functionality from GUI.
Slowly create non-windows releases.
Though all of that is sweet talk about an uncertain future and I also doubt that this thread will change what has been unmodified for years.
#45
Posted 10 February 2016 - 10:02 AM
- feel the lightweight! - featuring Snarl support, the Client Analyzer and tits!
Coded by a Golden eMule Award winner and most people's favorite modder!
..........................................
Music, not muzak:
Progressive Rock :: my last.fm profile
..........................................
eMule user since 0.28 ...
-[ ... and thanks for all the fish! ]-
#46
Posted 10 February 2016 - 10:47 AM
Tuxman, on 10 February 2016 - 10:02 AM, said:
iirc aMule is what wxWidgets, I would say its rather quaint, surly less modern than Qt.
In my opinion a good approach would be a very generic core, like c++11 with stdlibs or boost and a separate GUI done in Qt.
Like iirc qBittorent does.
it is the first client to be able to download form multiple networks the same file.
NL provides the first fully decentralized scalable torrent and DDL keyword search,
it implements an own novel anonymous file sharing network, providing anonymity and deniability to its users,
as well as many other new features.
It is written in C++ with Qt and is available for Windows, Linux and MacOS.
#47
Posted 10 February 2016 - 11:07 AM
Your eMule Fanboyism is uncanny, also did you beg Some Support to remove one of my last posts? Since it was deleted. You asked what was missing in eMule as I stated IPv6, NAT-T. So you think that shouldn't be added why? aMule is going to be dominate figure in *nix as I stated as eMule going to be de facto in your Windows OS. So do not confuse foolishness as cutting edge as you also stated eMule can run on Windows 10 so what? You can deploy 1992's DOS game also now does it mean throw away GPU's that costs more than your whole PC?
@David & Wizard
As we all know currently we have aMule for our *nix which de facto and next thing that comes close is mldonkey with much more superior control. No one is going to deploy WinE/apiC to just run eMule on it as its foolish so they use aMule. The multi platform development can happen in the future but for now it shouldn't as all resource should be allocated to reviving eMule to current common core value for today's world. BTW David yes aMule uses wxALL (xWidget Base) so it already works and can be deployed nativity for Windows platform. I remember the days when I was using UNIX and had to do ~elf for multi platform support until we started on JAVA.
@RejZoR
Don't mind Tux as you have seen he is confused fanboy who loves it so much and afraid of changes as its for better but he refuse it since he is scared. You know there is a problem when it takes 2years to download a file lol and what was it near 1gb? Where it was shared near 100GB+? This is why all my files is mirrored. Since I have been back on our network I delivered some content that was gone from the network since my forceful shutdown by Media Corp. I have been collecting stats and very soon it will be populated.
@hooligan
Something like DLP is needed to combat bad mods.
#48
Posted 10 February 2016 - 11:12 AM
eMule should be unbundled into UI + Daemon + NetAPI.
UI = User Interface (or "FrontEnd") will be the place where you click your mouse, push buttons, type on keyboard. It can be CommandLine, MFC, wxWidgets, Qt, .NET/Mono, WebUI. Whatever you want or need that can be cross-platform and/or multi-arch.
Daemon (or "BackEnd pt1) is a background application that communicates with the UI and uses the network API to upload/download data.
NetAPI = Network API (or "BackEnd" pt2) is a dynamic library that provides the API that implement the ed2k protocol (github: libed2k is a good place to start).
Advantages of unbundling:
- Developers can focus on which part of the software they want to work on, this is useful if you want to integrate with new operating systems, all you need to do is just change the UI.
- Daemon and NetAPI can be easily made cross-platform and multi-arch, with little effort this part of the software can work on Windows/Linux/MacOS and run on x86/amd64/arm/MIPS/etc architectures.
- More choice of UIs, somebody may want a basic window that just show download bars, somebody else may want to control every config, there can be UIs for dummies and UIs for experts.
- More customization, with conf/ini files one can setup every function of the Daemon, like it happens on aMule.
- One NetAPI to rule them all! The API can be used by any ed2k client, making all of them virtually compliant to the same protocol in the same way.
Disadvantages of unbundling:
- Different NetAPIs can come out, potentially creating uncompatibilities. (I'm no expert on this, is it possible?)
- Users may find even harder to install and configure eMule, this kind of problem can be avoided providing easy guides and installers that can make everything work out of the box.
- NetAPI bugs would break many clients instead of only one.
- The lead developer of the NetAPI needs a mission (just like companies) which isn't stuck in the 200*.
Some real life examples of software bundling:
- mldonkey = Daemon + NetAPI bundled in one exe, very hard to use for average users
- eMule = UI + Daemon + NetAPI bundled in one exe, excessive configuration = confused users + doesn't work out of the box
- aMule = UI separated from (Daemon + NetAPI) bundle, same as eMule
- Deluge = UI / Daemon / NetAPI; everything is unbundled, works out of the box, has less configs than eMule but it's more intuitive
#49
Posted 10 February 2016 - 11:59 AM
xSTHNSx, on 10 Februar 2016 - 12:07 , said:
Sorry for defending eMule in the eMule board.
xSTHNSx, on 10 Februar 2016 - 12:07 , said:
No, but I guess he had a reason.
xSTHNSx, on 10 Februar 2016 - 12:07 , said:
I agree that IPv6 should be added some time in the future. IIRC "the eMule team" had already agreed to that too. I, personally, don't think NAT-T is a viable solution for the problems it tries to solve though.
xSTHNSx, on 10 Februar 2016 - 12:07 , said:
This is a filesharing software.
This is not a software which requires sophisticated multimedia parts or whatever hogs up your system resources.
xSTHNSx, on 10 Februar 2016 - 12:07 , said:
Was your deleted post similarly aggressive? If so, could you please stop being aggressive? This is highly disturbing.
xSTHNSx, on 10 Februar 2016 - 12:07 , said:
That's why I'm here. However, I wonder why you are here - obviously not because you like eMule at all.
xSTHNSx, on 10 Februar 2016 - 12:07 , said:
Yes, I started to develop eMule mods because I was scared of changes. You totally got me here. Sorry, everyone, I just learned that I don't want changes in eMule, so I'll sadly have to stop development. Goodbye.
xSTHNSx, on 10 Februar 2016 - 12:07 , said:
Lolroflmao. P2P software allowing uploaders to have a small bandwidth. Roflrofl! Sooooo 90s!!!!!
Please, seriously, go troll elsewhere.
xSTHNSx, on 10 Februar 2016 - 12:07 , said:
Bad mods which are identifiable by DLP are not bad enough.
- feel the lightweight! - featuring Snarl support, the Client Analyzer and tits!
Coded by a Golden eMule Award winner and most people's favorite modder!
..........................................
Music, not muzak:
Progressive Rock :: my last.fm profile
..........................................
eMule user since 0.28 ...
-[ ... and thanks for all the fish! ]-
#50
Posted 10 February 2016 - 12:02 PM
That's not bad idea but as eMule is still the leading client for ED2K/KAD network, which happens to be for Windows platform its will remain so as targeted. As aMule has done nice job separating it's core (aMuleD) in modular way thus allowing External Connection (ECv2) for bi-directional interface to communicate with external aspects of aMuleCMD and aMuleWEB by disabling monolithic parameter during compiling for aMuleGUI.
aMule is not going to replace eMule as its there to complement eMule for other platform including Windows. At the same time aMule usersbase have increased due to people slowly transitioning from intrusive Windows to LINUX distros. Those users still remain in our network may it be they are on aMule as they are still with us. Only client that is not visible as it once was on our network is Shareaza which was virtually wiped out in half decade during its debacle.
#51
Posted 10 February 2016 - 12:23 PM
Does common sense elude you? you should take your own advise and stop these "aggressive" posts as you continue to display "disturbing" childish anger as we all can see it as fanboy "trolling". I'm here since eMule is the dominate client on ED2K/KAD. I don't need to use Windows nor eMule to be part of this network as I have done my part in development and content sharing on this very network long before which you will never surpass. At the same time we all here agree and see the problem that eMule has caused to the network in past decade. As just by modifying few attributes of it we can tackle the LowID problem which is crippling ED2K as it crippled KAD. It use to be below 30% now we have 60% LowID users.
"Lolroflmao. P2P software allowing uploaders to have a small bandwidth. Roflrofl! Sooooo 90s!!!!!"
How did you participate in the thread and fail to read its context as I was referring to someone elses post where User A (main source) uploaded to User B (downloader) as he slowly overtime came to 99% and the main source disappeared. Thus leaving User B with incomplete file for 2 years where it shared 100% more than it downloaded to User X+ (other downloaders) as they even had the missing 1% they didn't send to User B as they downloaded rest missing chunks from User B and went offline as well. Just do forum search you will know what i'm talking about.
#52
Posted 10 February 2016 - 12:36 PM
xSTHNSx, on 10 Februar 2016 - 01:23 , said:
How does your aggressive trolling relate with common sense?
xSTHNSx, on 10 Februar 2016 - 01:23 , said:
You are the one calling me names here.
xSTHNSx, on 10 Februar 2016 - 01:23 , said:
Please bring some evidence.
xSTHNSx, on 10 Februar 2016 - 01:23 , said:
A fanboy of what? Of eMule? What is "fanboyism" to you - to like using something? In this case: Yep. As most people here probably are, because else they'd use something different. Which other software are you a "fanboy" of?
xSTHNSx, on 10 Februar 2016 - 01:23 , said:
So you lied in your deleted post where you said you never used eMule? Good to know that you are not only a childish anti-eMule troll but also a liar. (I won't name your numerous insults here, don't worry.)
xSTHNSx, on 10 Februar 2016 - 01:23 , said:
No eMule - no network. Its "problem" is its maturity, there's simply no need to redo everything every few weeks.
xSTHNSx, on 10 Februar 2016 - 01:23 , said:
This is how P2P works, it is not related to eMule at all.
It's amazing how you totally miss the point that I contribute to active eMule evolution with things like beba because I'm not happy with everything.
This post has been edited by Tuxman: 10 February 2016 - 12:38 PM
- feel the lightweight! - featuring Snarl support, the Client Analyzer and tits!
Coded by a Golden eMule Award winner and most people's favorite modder!
..........................................
Music, not muzak:
Progressive Rock :: my last.fm profile
..........................................
eMule user since 0.28 ...
-[ ... and thanks for all the fish! ]-
#53
Posted 10 February 2016 - 12:57 PM
Bring some "evidence"? Sure, what you doing now is "aggressive" and borderline "trolling" as you trying to derail this thread in attempt to have it locked? We all can see it, maybe you should make another thread with poll where it asks user from this thread if your trolling or not. It's bit ironic you talk about name calling yet you call me "liar" lol when I openly have said I never ran eMule in my life, you would know this if you actually read this thread. Unless you can show proof I have which I would love to see since I been posting here for from xMule now aMule yet you saw otherwise?
"This is how P2P works, it is not related to eMule at all."
That is not how P2P works, that is how YOU think how it works. I was sharing content long before you had a PC where we manually converted utf8 to binary that is 2decade before you downloaded first copy of Napster (1G-P2P) and as you using eMule now which supports ED2K (2G-P2P) with mock up version of DHT for KAD (3G-P2P) network. I just went and checked your thread regarding your eMule.APK mod for Android using Eclipse where I wrote and released AOSP ROM with VI. How does your mod solve the problem when only way to do it is educate LowID users to obtain HighID and implementing IPv6 for those who doesn't have IPv4 allocated to them by their ISP and NAT-T for those who doesn't have access modifying NAT/PAT for full cone setup.
#54
Posted 10 February 2016 - 01:25 PM
But you xSTHNSx come in here, being registered few months ago, say that you'll never use emule and start randomly insulting people. You should piss off from here back to wherever you came from. You're doing absolutely nothing useful and have no right to voice your crap in the manner that you do.
#55
Posted 10 February 2016 - 01:55 PM
@eremini
What are you babbling about? I didn't come here few months ago as I made this very clear I was here over more than a decade. Now you would actually know this if you participated in these threads, just search the forum and you will find me. However I was gone since all my servers were shutdown by Media Corp back in 2010 by force as I operate within USA. Also is it hard for you to grasp I can use other client other than eMule to be on this network? As I use UNIX/LINUX why would I run eMule over WinE/apiC when I can just deploy aMule?
I'm here due to eMule being the main dominate client on ED2K/KAD network and main cause of the slowdown for everyone. Now its your turn A you registered here in 2002, since then what you have done? what did you contribuate to the network in terms of sharing or development?
This thread was not flamebate until Tux started his trolling and you started to babble about ANTI-Torrent. I was going to point out that you lack basic understanding in first page to fully grasp how Torrent exceeded surpassing ED2K/KAD in every aspect thus its de facto file trading/sharing protocol. That is not a debate my boy that is actual proven fact for over half decade now with its footprint in network traffic to speak for itself. This thread was about basic internal changes of eMule to speed up the network thus ensuring healthy survival.
#56
Posted 10 February 2016 - 02:01 PM
IPv6 only will be really bad as it will partition the network in those with IPv4 only those with booth (basically next to no one) and those With IPv6 + ISP NAT. Sure all these people in IPv6+NAT lines will be able to somewhat share with old IPv4 users but only as very intentionally handicapped low ID's. Since it is I think out of question that LowID users have worst DL rates than high ID users that will mean that in this particular case the IPv4 users will be systematically leeching the IPv6+NAT portion of the users.
And that's a shit! Tux do you really want to be a leecher?
On an other side note, I don't give a flying F*ck about ideology, imho pragmatism is the best way to go. And for the question about CS and Queue, the simple fact is that the modern user knows BT before everything so if any new (old revived) client wants portion of the market you have to compete in therms of download experience with BitTorrent, even if that rapes you sharing ideology.
You have to adapt to the market or fade into obscurity!
David X.
it is the first client to be able to download form multiple networks the same file.
NL provides the first fully decentralized scalable torrent and DDL keyword search,
it implements an own novel anonymous file sharing network, providing anonymity and deniability to its users,
as well as many other new features.
It is written in C++ with Qt and is available for Windows, Linux and MacOS.
#57
Posted 10 February 2016 - 02:03 PM
- Simplify GUI arrangement. Currently it feels like it was designed for computer experts. Sure, allow all the "clutter", but by default, keep it nice and clean.
Effort needed to resolve this: Half an hour and I'm being very generous here.
- Fix the darn UPnP. Works on one PC but repeatedly keeps failing on another within exactly same network. Identical OS used, only difference is wireless and wired connectivity. LowID is the biggest issue for having a healthy network and no one seems to care about it. It's UPnP that lets eMule sniff on the internet past the routers which are in every single home these days.
Effort needed to resolve this: I'm not a programmer so I don't know, but since function is already here, tweaking it can't take that freaking long.
- Update server.met and nodes.dat
Effort needed to resolve this: 5 minutes tops
Just 2 things out of 3 not taking even an hour in total to fix would already make a MASSIVE difference to eMule experience and usability. If UPnP gets a bit of a touch up, maybe a mechanism that automatically tries to resolve connectivity issue and it would be like day and night. And we haven't even started talking huge core changes. So, why the hell is nothing being done here? I appreciate all the work they've done so far, but I can't appreciate all the work they haven't done lately. Especially because we are talking about small things that make big changes and require nearly no testing.
Stulle @ Dec 7 2005, 06:16 PM
#58
Posted 10 February 2016 - 02:11 PM
You hit the nail right on the head. Just by modifying few current aspect would change the network in direction where it would yield positive result compared to current state. But the problem is these foolish fanboys choose and support in stance to bog down the whole network where they can leech of others and refuse to share with others. When I left our ratio of LowID user was around 30% now its double that around 60% of the network.
@DavidXanatos
I'm surprised you didn't just talk about it but you actually backed it up by making NeoLoader and proving to everyone that it can work as it has under your project. I actually came across client's who uses your Mod thus looked into it and found it supports LowID2LowID, NAT-T, IPv6. Why it wasn't added to mainline is beyond us all but thanks for your conurbation as you do have userbase which use your mod.
#59
Posted 10 February 2016 - 04:16 PM
This post has been edited by eremini: 10 February 2016 - 04:17 PM
#60
Posted 10 February 2016 - 05:02 PM
Everything around eMule is evolving, just not eMule itself. We now have huge hard drives, faster connections than ever and yet eMule hasn't really adapted to any of it. Sharing 5000 files totaling 150GB and people say "that's a lot" and that's the reason why KAD is totally stuffed and takes literally decades to propagate files within network. Connections have boosted dramatically. I used to have 1024/256 Kbps when I started using eMule. Today I have 20/2 Mbps line. And yet everything feels like back in the days of 56k and "fast" 256 Kbps cable connections. I still remember when KAD was first released. I was like a small kid on Christmas day, waiting to give it a spin. Fastforward to today and KAD is pretty much then exact same thing it used to be. It hasn't progressed anywhere, I'm not even sure I can say it serves any purpose. You can't really rely on it without servers because it's so god damn slow to actually get anyone to download files from you. That's why I love the concept of kMule, but due to nature of how KAD works, it's not what it should be. And that sucks really bad.
Some follow the rule: "Don't fix if it ain't broken". I live by the rule "If it ain't broken yet, you haven't tweaked it enough". Current eMule devs are not doing enough of the second...
Stulle @ Dec 7 2005, 06:16 PM