Critical components to open source

[Locutus]

Originally Posted by kamelito

Why not compile AROS 68k to 68k asm the all OS and then hand optimize and enhanced it?
Kamelito

Do i need to link another Picard_Facepalm.jpg ?

[Olaf Barthel]

Originally Posted by kamelito

Why not compile AROS 68k to 68k asm the all OS and then hand optimize and enhanced it?
Kamelito

Because the whole process does not scale.

Reading and manually tweaking machine generated code is a exceptionally challenging considering how much code would be involved.

Well, at least you can check the original source code and compare the output to what the original intention was, but that's no help for a few hundred individual source code files, producing thousands of lines of assembly code to pore over.

A compiler is a tool which is intended to make you more productive by letting you express machine instructions with less effort, conquering complex tasks more easily. Think of the compiler as a means to "compress" programs. The assembly language output which it produces is by design less expressive, much larger and requires much more effort to understand.

And these are massive problems which balloon before the equally challenging step of integrating the changes into the target system are considered.

[matthey]

Originally Posted by kamelito

Why not compile AROS 68k to 68k asm the all OS and then hand optimize and enhanced it?

Originally Posted by Locutus

Do i need to link another Picard_Facepalm.jpg ?

I don't know why this would be a face palm moment. If the price of open sourcing P96 (or CGFX) is too high, the money may be better spent paying for backporting and optimizing AROS RTG for the 68k. AROS 68k could benefit with a more usable AROS on the 68k and I believe existing P96 drivers can be used. The biggest hurdle is that the graphics.library is in kickstart (P96 and CGFX are patch hell to avoid) but this is also a problem with just open sourcing P96 (or CGFX) and not the whole AmigaOS so new kickstarts can be made. The classic Amiga would benefit from new kickstarts (ROMs) with support for larger devices, bug fixes, no double reboot, etc. Development and testing would be a plus here and even open sourcing the AmigaOS does not guarantee this (more forks are just as likely as one new cooperative development release).

The overwhelming question is, is continued Amiga development worthy of investment and development? There are only 2 Amiga continuation options I see. Switch to another architecture (ARMv8 would probably be easiest) using commodity hardware or develop custom hardware (using FPGA technology). Switching to another architecture is probably easier and appears to be less risky, at least on the surface. Most of the development would be to the AmigaOS (or AROS) and compatibility would be through a sand box Amiga emulator. If this route is chosen, there is little advantage to wasting resources to open source P96 (or CGFX) as these are already integrated in AROS and we would be abandoning the 68k architecture. It would be better to fund further progress to the AmigaOS (or AROS) toward new architectures. Funding AROS really puts the pressure on Hyperion and A-Eon to compete (risk on instead of milking the cow to death). The AmigaOS can become open source more easily (cheaper) if/when they are irrelevant but it won't matter much as AROS will have replaced it. Choosing to stay with the 68k and develop it provides the best compatibility, retains the most Amiga users and allows for Amiga innovation again but likely would need an initial investment of millions of dollars for a mass produced affordable hardware. Open sourcing the AmigaOS could be helpful in this scenario but so would paying to improve AROS 68k (if no official 68k AmigaOS development). A Clone-A with ARM CPU emulating a 68k is not likely to be enough with the Apollo Core applying pressure through competition to Jen's business. Too bad we can't all work together and pool resources to bring the Amiga back the way it should be done. I would invest in a proper 68k project with AmigaOS development but just funding open sourcing is not enough for me even though I prefer for the sources to be available.

[plasmab]

Originally Posted by matthey

The overwhelming question is, is continued Amiga development worthy of investment and development?

Of course its not worthy of investment (at least in my opinion). But that isn't going to stop us. As the man said.. he's a grown adult who can spend his money how he wants.

I've probably spent (in the 2-3 years) 1-2000 GBP on Amiga kit directly and another 2000 on equipment and parts relating to creating accelerators. This means i've spent more than I would have on a top end laptop to enjoy a 30 year old platform going slightly faster. Its not because I think Amiga is useful or something to invest in for the future... its just something I love(d) doing.

I'm not kidding myself that there is a future here. Its fun/hobby/niche.

[wawa]

Originally Posted by kamelito

Why not compile AROS 68k to 68k asm the all OS and then hand optimize and enhanced it?
Kamelito

lets start with, that what you call aros68k, except maybe very few asm inlines in code, that may happen to use some 68020 instructions, are currently being compiled for plain 68000. there are seome patches installed by setpatch at runtime, if the user invokes it, for instance in the s-s, but thats it.

[wawa]

Originally Posted by matthey

the money may be better spent paying for backporting and optimizing AROS RTG for the 68k.

exactly, except i wouldnt backport it but take as a whole, fix and optimize. rtg is not the one and only advance feature, aros has. it has built and and improved tcp stack, amiga applications can transparently take advantage of, it has usb stack, it has skinning if one wants it, ttf outline fonts support, it has built in posix compatibility lib, something comparable to ixemul, i would say, it has opengl in form of mesa/gallium and appropriate interface for gallium drivers, and and and..

and all this in a system that takes 10mb as basic package, and needs few mb to run.

if you think about it right, there are very few amiga features for which you would want to keep the original os in comparison. currently it is stability, or rather iwould say compatibility, sicnce aros is not randomly instable from my experience, and obviously speed. other than that pal/ntsc overscan, screen dragging with different resolutions and support ofr some other unique feats of amiga chipset is missing.

The biggest hurdle is that the graphics.library is in kickstart

it is a feature! you can have rtg available at boot time, same as various other interfaces.

The overwhelming question is, is continued Amiga development worthy of investment and development?

with aros this problem is moot. the development for the genuine platform and anything you would laike to switch to, happens already in parallel.

look, i made a number of commits within last few days, which in aros weekly report are considered being 68k/gcc4.6 fixes. in fact only few of these are 68k exclusive.

[matthey]

Originally Posted by plasmab

Of course its not worthy of investment (at least in my opinion). But that isn't going to stop us. As the man said.. he's a grown adult who can spend his money how he wants.

I've probably spent (in the 2-3 years) 1-2000 GBP on Amiga kit directly and another 2000 on equipment and parts relating to creating accelerators. This means i've spent more than I would have on a top end laptop to enjoy a 30 year old platform going slightly faster. Its not because I think Amiga is useful or something to invest in for the future... its just something I love(d) doing.

I'm not kidding myself that there is a future here. Its fun/hobby/niche.

Raspberry Pi has sold over 11 million units
NES classic edition has sold 1.5 million units (despite shortages)
Natami MX bringup thread had 761487 views
Your dueling banjos video has 6860 views
lantus360's CannonBall video has 5252 views

You don't think a 68k computer/console which can play Amiga/CD32, Atari ST, Neo Geo, Sega Genesis/Mega Drive, x68000, 68k arcade games, etc. with 68k+FPGA while providing Raspberry Pi like computer features would sell? The NES classic was only the 3rd best selling console in January 2017 and none of these other 68k systems offer that much retro goodness now do they? A 68k system with 1 GB of memory like the Raspberry Pi would never be enough either would it?

I guess we have different opinions on what is "worthy of investment". I don't want to "donate" money for something without a future. I am an investor, Amiga fan, 68k fan and retro computer fan and I want to invest in something which will further those causes. I see the retro interest and all the FPGA talent and work which already exists in the Amiga community but is wasted because of lack of investment, cooperation and leadership. Mass production is the only way to lower prices enough to bring in new customers and a new generation which are necessary to sustain the Amiga. The greatest risk to the Amiga is the current status quo of deterioration followed by certain death. It is this which is not worthy of investment, IMO.

[plasmab]

Originally Posted by matthey

Raspberry Pi has sold over 11 million units
NES classic edition has sold 1.5 million units (despite shortages)
Natami MX bringup thread had 761487 views
Your dueling banjos video has 6860 views
lantus360's CannonBall video has 5252 views

You don't think a 68k computer/console which can play Amiga/CD32, Atari ST, Neo Geo, Sega Genesis/Mega Drive, x68000, 68k arcade games, etc. with 68k+FPGA while providing Raspberry Pi like computer features would sell? The NES classic was only the 3rd best selling console in January 2017 and none of these other 68k systems offer that much retro goodness now do they? A 68k system with 1 GB of memory like the Raspberry Pi would never be enough either would it?

I guess we have different opinions on what is "worthy of investment". I don't want to "donate" money for something without a future. I am an investor, Amiga fan, 68k fan and retro computer fan and I want to invest in something which will further those causes. I see the retro interest and all the FPGA talent and work which already exists in the Amiga community but is wasted because of lack of investment, cooperation and leadership. Mass production is the only way to lower prices enough to bring in new customers and a new generation which are necessary to sustain the Amiga. The greatest risk to the Amiga is the current status quo of deterioration followed by certain death. It is this which is not worthy of investment, IMO.

Ok i'm probably missing something but I really dont see how what you describe relates to investing in the original AmigaOS. I thought that this thread was about that. IMHO if you invest in Amiga you'll never get your money back. But thats just an opinion.

My dueling banjos video has made less than $1 of ad revenue. Which considering it took me 2-3 hours to setup and shoot is well below minimum wage. My TF530 card was between 1-2000 hours of work. If i was serious about making money on this project the cards would need to sell for 150-200 GBP per card. Which wouldn't be viable.

[matthey]

Originally Posted by plasmab

Ok i'm probably missing something but I really dont see how what you describe relates to investing in the original AmigaOS. I thought that this thread was about that. IMHO if you invest in Amiga you'll never get your money back. But thats just an opinion.

My point was that donating money to open the AmigaOS sources or components isn't going to change much without a plan for new affordable hardware which can attract outside people. The retro gaming angle for hardware is one way to do that and might attract more investors, at least me. High risk is much better than guaranteed no return.

Originally Posted by plasmab

My dueling banjos video has made less than $1 of ad revenue. Which considering it took me 2-3 hours to setup and shoot is well below minimum wage. My TF530 card was between 1-2000 hours of work. If i was serious about making money on this project the cards would need to sell for 150-200 GBP per card. Which wouldn't be viable.

The Amiga market has been neglected for a long time but there is still interest. How much advertising or promoting have you done for your videos? It is interesting that your dueling banjos has the most views considering it is fun and not really useful (it does attract Amiga and Atari ST people which shows the value of cross platform retro appeal). The Natami bringup thread is really the biggest indicator to me of Amiga hardware interest as the number of views was without advertising or promotion and the price was thought to be too high for most customers. The Natami motherboard could have been cost reduced and mass produced though. Your accelerator is cheap and simple for small production runs in a small market. Competition and requiring an old computer probably limit any chances of mass production and limit how much money could have been made. It is necessary to attract retro computer/console customers outside of current Amiga customers to make affordable and profitable hardware.

[wawa]

really? you want to develop wholesale strategies on how to bring amiga back to relevant market? and most curiously have the community of sworn individuals (except os4 crowd) to obey it? amiga is anarchy. live with it!

[PeterK]

Originally Posted by grelbfarlk

PeterK's icon.library, we can't have one person (however benevolent) distributing his critical piece of software as a "freeware". It MUST be purchased and GPL'ed and handed to an undying consortium in a perpetual safe trust for all of eternity.

I consider PeterK's icon.library critical and that's one library not needed to license or rewrite. Even though it's freeware as you say it might not be GPL as you desire.

You can do whatever you want with my icon.library and the source code. You are allowed to modify it, you can sell it (if you find an idiot who would buy it) and you may GPL it for free to make it foolproof, but please, then don't distribute it as "IconLib 46.4", use another name or version/revision number. That's all.

[wXR]

Good attitude there PeterK.

And wawa you made me laugh out lout with that, "Amiga is anarchy. Live with it.". So true that it should be a t-shirt.

[grelbfarlk]

Originally Posted by PeterK

You can do whatever you want with my icon.library and the source code. You are allowed to modify it, you can sell it (if you find an idiot who would buy it) and you may GPL it for free to make it foolproof, but please, then don't distribute it as "IconLib 46.4", use another name or version/revision number. That's all.

Unacceptable. I feel money should change hands before it gets GPL'ed. I would suggest a single payout of no less than 750 EUR to you for all current past and future versions and then it can be released under GPLx.

Branding yes I agree is very important how about Icon For4ks as a name?

[nobody]

Even if they hand over everything it won't get you anywhere without porting everything to ARM which I guess is a massive job.

[PeterK]

Originally Posted by grelbfarlk

Unacceptable. I feel money should change hands before it gets GPL'ed. I would suggest a single payout of no less than 750 EUR to you for all current past and future versions and then it can be released under GPLx.

Branding yes I agree is very important how about Icon For4ks as a name?

No, I would refuse any payment. I was always glad to get so much free software and open source code from the Amiga community that I regard my little contributions only as a very small refund for what I got. I like the idea of open source, but I'm not a fan of all these rules that are included in GPL and don't see why I should need that.

Future versions? What are you talking about? I'm an old man and my job is done! (Bugfixes are an exception, of course) Icon For4ks? No, I've no more plans for the rest of my life to buy another new PC and a 4k display. That would also need a better upscaling function. My Amiga screenmode is still 1280x800 on my 1920x1200 monitor. Enough pixels!

[michaelz]

As a simple message on this moment, I contacted Cloanto yesterday and got a message from Mike Battilana that at least gives some space to start negotiations. I asked if I'm allowed to put the messages in verbatim on this thread and will do so if he agrees.

[grelbfarlk]

Originally Posted by PeterK

No, I would refuse any payment. I was always glad to get so much free software and open source code from the Amiga community that I regard my little contributions only as a very small refund for what I got. I like the idea of open source, but I'm not a fan of all these rules that are included in GPL and don't see why I should need that.

Future versions? What are you talking about? I'm an old man and my job is done! (Bugfixes are an exception, of course) Icon For4ks? No, I've no more plans for the rest of my life to buy another new PC and a 4k display. That would also need a better upscaling function. My Amiga screenmode is still 1280x800 on my 1920x1200 monitor. Enough pixels!

Oh well I just thought I try to get you a cash dump while people are throwing fat stacks around, good rationale you have for not accepting.

I meant ForA4ks, like "For A4000s" since that's what I use it on, actually A2000 and CDTV too but maybe people would just be happy and delighted that it worked on Not-A4000s. How about "Icon of Good"? Or GoodIcon, hang on it's on the tip of my brain. BestIcon library?

[Mr.Flibble]

Originally Posted by michaelz

As a simple message on this moment, I contacted Cloanto yesterday and got a message from Mike Battilana that at least gives some space to start negotiations. I asked if I'm allowed to put the messages in verbatim on this thread and will do so if he agrees.

Much appreciated!

[wXR]

@michaelz

Thank you so much for doing that! I feel the start of some genuine momentum here.

[michaelz]

Okay, I just send another 2 page e-mail to Mike Battilana and there genuinely seems to be a foothold.

In the first place, Mike mentioned bringing up the idea of getting a non-profit to "own" Amiga a few times earlier; [ Show youtube player ], AmiCast #11 and at the [ Show youtube player ].

On this moment I can't place the e-mails in verbatim here, but he really strikes me as someone with the communities best interest in mind. What I can mention is that he is willing to provide "free" beta-test licences to people who participated in this thread until now. This would probably involve giving back some technical feedback for those who participate, but I'm currently asking what strings would be attached. I don't mind some strings, but I want to have those strings clear before I start offering "free licences" just like that. Would people be interested in this (of course; depending on what the strings are)?