I've been thinking about this some more.

It's clear that some people expected more. But I still feel it's quite
unfair to blame Sonic. He communicated very well what he would do here
on MZ. If this wasn't what people wanted, we should have reacted
before, not now.

I also think it's unfair to put all the blame on the bounty
administrators. Maybe they did a mistake, but they surely could have
gotten more and better advice here on MZ.

In the end, the weird thing about this is that while everyone knew
that there was already a TCP stack being worked on for years, tested
and improved by the most knowledgeable MorphOS people for a long time,
most people still thought it was a good idea to give Sonic the
go-ahead to port an old stack. Again we can't blame Sonic - writing a
completely new stack is a *huge* job and the old AmiTCP sources he
used was pretty much the only option.

But ok, for some reason most people seem to have agreed it was a good
idea. Only a few voices had anything different to say.

Anyway, it's not the end of the world. This bounty system is a tricky
thing and we can't expect a total success every time.

But if we really want a better native TCP stack, I think we should
make a serious attempt to persuade laire to finish and release his
effort. Maybe we could start by making some poll to see what people
want, if people are willing to pay, and which project they would
rather support in that case. Too bad laire's DVI poll didn't get
enough interest, but evidently there is more interest in a TCP stack..
I just hope this debacle hasn't put people off too much. And I'm sure
there are people who didn't contribute to the previous bounty, who
would put in something to get laire's stack released. At least I
would.

I've paid for Miami, MiamiDX quite a lot of money and
I've donated a few bucks to MOSNET ...

And I think I got in both cases what I deserve ...
From my point of view the legal part of the
bounty is filled ...

I've donated nothing for MOSNet, and I got a TCP/IP-Stack which works for me as good as Miami(dx) did, so I've got no right to complain. And I don't need the extensions Sonic suggested, so I won't donate to his second bounty either.

@ Sonic

I fully agree what Acill wrote here:

*****************************************************
I know you are reading this and may feel a bit upset at some posts. Just keep in mind this in know way is an attack on you. Your work is very much needed and we all are glad you do what you do. Keep it up, just please do a bit more on MOSNet before calling it done and moving to another bounty!
*****************************************************

I cant say though if a 2nd bounty is needed for implementing some functions?? Though I would need the one suggested by you as e. g. SpamFryer cant be used anymore with MOSNet and I find it is slower as MiamiDX.

Regards

Christoph

http://www.freelists.org/archives/mosnet/03-2006/msg00000.html

And now some words about the future.
Now the project is able to move forward in any direction you would
like. Some tasks can be done on a free basis, some of them (large
ones) require some sponsorship.
A free projects (will be done by me in any case):
1. TCP protorol upgrade to support T/TCP and window resizing.
2. Configuration editor.
3. Implementing missing miami.library functions - is done per
request. If you find some software which generates "Not implemented"
messages in the syslog, please tell me about this and it'll be fixed.
A projects requiring bounty sponsorship:
1. Total stack code upgrade based on recent NetBSD code.
2. IPv6 implementation.
3. PPP (+ PPPoE, PPTP) implementation. Requires co-developer with
ability to perform excessive testing.
4. Development of new network driver interface, probably based on
current BSD driver API.
5. Implementing some other protocols (IPX? Appletalk? What else can
be still in use today?)
A volunteer-based projects (requiring volunteer work, other
conditions depend on who takes them):
1. Porting to any other Amiga-alike (well, why only Amiga-alike, may
be some embedded?) operating systems.



I guess some gui/config editor are in the works..

I'm still using AmiTCP/IP Genesis. No problems with anything. Stable, way faster than Miami or MOSNet, inetd's and all work... but anyway it's good that there's freeware option too.

Quote:
As I see it, the biggest advantage of MosNET over MiamiDX is that, there's unrestricted version legally (and freely) available. This isn't true for MiamiDX. You can't really say "Get a warez version of MiamiDX" to every new user (as that's the only way to get the non-time-limited version)

Personally, I didn't think of this as a "worthy" bounty (for a temporary solution to the lack of publicly available native TCP/IP)

About GUI... I think the most important thing would be to have a GUI where you can enter all configurations (IP addresses, buffer sizes etc.) for the connection. For the stack itself, some "online/offline" GUI could be done, but I think that settings gui would be more important. I guess such a settings GUI could be created quite easily with f.ex. rxmui. I think I could do it, but well, it isn't my bounty

Quote:


Holy shit!!11 i bet linus torvalds could do that too with his thing....whats teh name....yeah, linux..
He's been working for free for years...poor man

Quote:
Actually, he's employed, as are many famous Linux developers (by companies close to Linux, such as Red Hat, qt, Suse, IBM etc). The big fish don't have to work their magic after 4 pm, unpaid, unshaved, feeding on bread and water. Linux expertise pays off and making open source software can put food onto the table. Not in the MOS corner, though, which is why you should be a bit nicer.

>Not in the MOS corner, though, which is why you should be a bit nicer.

It got nothing do to with being nice, the rules was set and sonic accepted the bounty, dont come here and bring in real world wages here.
If sonic was out of work and was looking for money it would be pretty stupid to use from september to february
porting this aros stack for 1600usd.

Put this in another angle, if the bounty donations was 30.000 usd, would you still say stfu and be nicer?

the stack have bugs , and no gui but it was promised
it should be fixed becaused it was the terms, if its for 100usd or 30000usd shouldnt matter.

Quote:

And thats it exactly! Thats what the GUI should have and what was expected.

Hi,

Here is my point of views on this. And to Sonic too.

What the bounty description says:

------------------------------------
Basic: write, convert, or otherwise get a native tcp/ip stack for MorphOS.

Long: Connectivity to the internet. Compatible with applications using BSDSocket.library. Must support DHCP, have a MUI or other GUI that includes all necessary classes or use default MorphOS 1.4.5 classes. Networking protocoll can be a port of an existing project from another platform, a port from AROS, or even a whole new project.
------------------------------------

I would like to say, its not a very detailed description. This is maybe something to learn for future, to write more detailed bounty descriptions of what the goal, and what should be implemented in the software.

What I can read out from the above description is: TCP-stack that runs on MorphOS 1.4.5 - Compatible with applications using BSDSocket.library - DHCP - GUI.

# The TCP-stack runs on MorphOS 1.4.5

# It is Compatible with applications using BSDSocket.library

# It has a DHCP client

# It has a GUI

For the speed part, I get around 7Mb/s in download from the linux-computer, 5-6Mb/s in upload. Its about the same results as if using Miami. The transfer rate should of course be 10Mb/s as thats what the network capacity is. I have myself never got that result within MorphOS. Maybe Ralph Schmidts TCP-stack will make it this. Dont know as I never used it myself.

As for the GUI, the GUI only consist of a Miami Panel, where you can see traffic and put the interface on/offline. Nothing else can be made in this GUI. However the bounty description doesn't specify what should be possible to do in the GUI, or does it ?

As for the MiamiDX GUI, its GUI is a really advanced GUI, where you can do a lot of things. In matter of fact You can do everything in the GUI in controlling of Your TCP-stack. It is not many GUI's out there for any TCP-stack or on any platform where you can control everything so detailed as you can do in the MiamiDX GUI. It is really a very great GUI in my opinion. You can also easely control several networks configurations independently of each other in the MiamiDX GUI.

I think however the aim with the request for a GUI in the bounty description can be seen as expecting little bit more then what the Miami Panel thats present for now offers. Also Sonic did write when he releassed version 1.0 that he would like to work on some more parts.

------------------------
Quoting here:

And now some words about the future.
- Now the project is able to move forward in any direction you would like. Some tasks can be done on a free basis, some of them (large ones) require some sponsorship.
- A free projects (will be done by me in any case):
- 1. TCP protorol upgrade to support T/TCP and window resizing.
- 2. Configuration editor.
- 3. Implementing missing miami.library functions - is done per request. If you find some software which generates "Not implemented" messages in the syslog, please tell me about this and it'll be fixed.
------------------------

However I think expecting that the GUI (configuration editor) to be as good as the MiamiDX GUI, is to expect little to much, at least for now.

So finally I think it was not wrong too give the bounty funds to Sonic for his work with the TCP-stack. At least there is a TCP-stack now, that anyone can use. The software can only become better I think.


[ Edited by gunne on 2006/4/8 17:00 ]

Quote:
Well, stick your head into the sand and whistle, fine with me, it's your head, I'm not using MorphOS. However, the real world still exists, if you ignore it or not, rewards in it will attract developers away. Maybe you should be prepared to lower expectations, allow programmers leeway and not be a stickler about "the rules". It's not as if you've got a queue of developers lined up before the door. Your position, in a nutshell, is take it or leave it. As to your inability to understand the importance of being nice (in this thread, non-payers went as far as refrering to the developer as "that guy"), I can't help you. To conclude, he has only asked for a bounty, not for your life, maybe you should consider it. It's not as if you would have to pay into it, if you are totally opposed.

@dietmar

again you completely miss the point...but i dont have time for this useless talk.

>attract developers away

well, without users there are no need for developers either, is it ?

I'm very disappointed. I expected a native TCP/IP stack. This means (at least for me) it could transfer data at full 100Mb/s bandwidth (I have a G3, and my Linux CAN download files from internet with 10.5MB/s. With MOSnet the max speed I got was only 3.5MB/s.).

I expected a GUI. This means for me something similar to MiamiDX's GUI. What have we got? An implementation of using miami's panels. Cool. (At least I could leave my gelbesPanel where it was...)

I think that it's sad. Even the config files are exactly the same ugly-formatted ones given with AmiTCP...

Oh, sorry for the question, but why isn't is possible to port MiamiDX to MOS? I think it's very crazy that we have really the best tcp/ip stack of the world, and instead of improving it, we pay for porting much worser ones from Amiga... (What's absolutely crazy is that the MOS developers do an another version...)

Do you guys realize that the way some people (mostly people who where against this whole project in the first place) complain is really destructive ?

My 2 cents.

I greatly appreciate ANY development done for the platform as long as it has some use for some people even if has not for me.

If you don't like someone's work just ignore it.
If you don't like a bounty do not pay for it.
Is this that complicated ?
What about showing a minimum of respect for the people who made different choices than yours ?

This thread is really too much Amigaesque in the worst way of it.
What about we all meet with big guns and fire at each other until the last survivor so we will be done with it ?

Hi GK_LKA,

There is no chance a TCP-stack can load with 100Mb/s in a 100MBit network. Around 10Mb/s is what the 100Mbit network limits too.

About config files, they are similar to the config files You have in Your Linux system.

As for MiamiDX the answer is very simple. Its Holger Kruses software.

cdfr,

> What about we all meet with big guns and fire at each other until the last survivor so we will be done with it ?

------------------

No, no... thats not good .-)

Gunne: 100Mb/s yes, 100MB/s no :) b=bit, B=byte

GK_LKA: I wouldn't say Miami is the best stack in the world :) Seems it bugs sometimes and in any case, the speed is much lower than on amitcp for example.

jPV,

> Gunne: 100Mb/s yes, 100MB/s no :) b=bit, B=byte

---------------

I stand corrected !

jPV,

However, whats really interesting to know, is if Ralph Schmidt could show us an TCP-Stack that can take full benefit speed-wise of the Pegasos onboard 100Mbit ethernet-connector

as distructive as it may be, most do like the stack. I do, but for him to call it done is wrong and even more so comming and asking to take on another bounty to add to it. He didnt do a GUI, he did a interface panal support for some work another person did for another project. Did he give some of that money to the miami panel author?

Quote:
Why not Gb?

A GUI where you at least can set IP/ DHCP mode, DNS and Gateway would be appreciated, so that you can install and setup the stack without tempering with the config files by hand.

call it "frontend" and we will avoid misunderstandings in the future

ah, btw... i use MosNET since a month or so (i'm under a "sort of" router now) and its far faster/reliable than my previous PPPoE emulated solution with Genesis (quite obvious i know).
I understand that most of you, are not satisfied about the speed using LAN so i cannot comment about it (i have no LAN atm here).

So from my POV, MosNET fits all my current needs. (and its very cheap as well...)


[ Edited by Framiga on 2006/4/9 11:37 ]