What Software does 68K Amiga need?

[OlafSch]

@ReadOnlyCat

There is only one problem that it is not easy to develop a game with one person. That was possible to some degree in the 8bit time where graphics was less important than gameplay or ideas but even that changed later. Normally you need at least one programmer, one graphician and one music/sound specialist. Then your calculation no longer works.

[ReadOnlyCat]

Originally Posted by OlafSch

@ReadOnlyCat

There is only one problem that it is not easy to develop a game with one person. That was possible to some degree in the 8bit time where graphics was less important than gameplay or ideas but even that changed later. Normally you need at least one programmer, one graphician and one music/sound specialist. Then your calculation no longer works.

Some games are more forgiving than others regarding graphics.

Many coders could probably draw SOTB or Lotus (in game graphics, not static screens). Lionheart forget it since it is clearly way above but many games are not that demanding.

Audio is a tougher beast though, especially these unintuitive unfriendly frackers.

[Thorham]

Originally Posted by ReadOnlyCat

Many coders could probably draw SOTB or Lotus (in game graphics, not static screens).

That's REALLY not true. Many programmers aren't artists, and couldn't draw their way out of a wet paper bag. I wish I could do reasonable game art, but I just can't, and many coders with me.

[ReadOnlyCat]

Originally Posted by Thorham

That's REALLY not true. Many programmers aren't artists, and couldn't draw their way out of a wet paper bag. I wish I could do reasonable game art, but I just can't, and many coders with me.

These two sentences can be true at the same time:
- Many coders can't draw their way out of a wet paper bag.
- Many coders can draw SOTB/Lotus in-game graphics.

Most coders can't draw to save their lives but many can.
Even if the fraction of graphic-able coders is just 1/10th of the total that is still many kittens.

This said, like every other skill out there, if you train you will get better. You have a brain and neurons, you just need to make the effort needed to create the appropriate neural connections.

This is coder art btw: .
Pretty good IMO.

[meynaf]

While some coders might be able to draw things, others really can't. I can edit graphics : resize, change number of colors. But for building new graphics from scratch, really, no.
The other way around is also true : most graphicians are unable to code.

This is life. Not everyone can do everything.

And 1/10th of the coders doesn't look like many to me.

[DrBong]

Originally Posted by idrougge

How does it compare to ArtEffect, Photogenics 5 or FXPaint?

Off-hand, GIMP probably is most comparable to Art Effect/FX Paint - simple yet effective and elegant tools IMHO. I use GIMP a lot as a lightweight tool for image processing and touching up of material for HOL (see, for example, boxscans for Der Clou!).

Originally Posted by ReadOnlyCat

This said, I see no reason for the Vampire market to work better than the 060/PPC one in the long term. Good software emerges naturally when the user base is large enough and for productivity software I doubt it will ever be large enough. As usual, ports of games will arrive but I am not so sure that they will have anything Amiga specific to show. Will we see original Vampire games relatively as impressive as SOTB was (graphically, I mean) or will there be only ports of existing stuff?

I'd be surprised too if the Vampire market worked any better than the Cyberstorm/Blizzard PPC market. At least the PPC market in 1996/97 had a few commerical players still in the mix for software development. Moreover, and this is nearly always missed on Amiga forums these days, they still had a relatively large pool of young, enthusiastic university students / coders - particularly from poorer European nations - in the Amiga ranks (e.g. Sam Jordan, Steffen Häuser, Ablaze Entertainment, Hurricane Studios). At the time, they were perhaps a bit naive or masochistic in thinking that they could make a few bucks in the diminishing Amiga market, to support themselves while studying or breaking into the IT job market.

Times have changed and anyone in that situation now is looking to code the next big app for Android or iOS. Just about all the decent Amiga coders left are older and have real lives/responsibilities to attend to and pay for, not to mention don't believe that there is any 2nd, 3rd or 4th coming of the Amiga (Vampire, NG or otherwise)!!! Consequently, most software that will appear for the Vampire - assuming it establishes some sort of user base over time - will follow the status quo IMHO and consist of ports of existing software, predominantly open source as too few end-users will be willing to pay for licensed products.

Originally Posted by Ratteler

There is a fundamental divide in the entire Retro community. A Luddite movement who thinks anything but a 1.3 KS with a floppy drive is kind of sin, and those of us who want a neo-nostalgia.

Overly-simplistic and too black & white (things the Ami retro community has never ever been coz they just can't help being 50 shades of grey!), especially if you've spent any length of time here on EAB.

Originally Posted by Ratteler

The fact is, the only way worth doing anything NEW for the Amiga, is the Neo-nostalgia path. Otherwise development is too much of a wasted effort for too much of Niche Market.

Complete and utter nonsense. That's akin to saying that developing for anything but SuperCPU set-ups or equivalent on the C64/128 is just a lot of wasted effort!

Originally Posted by ReadOnlyCat

I read a lot of people dreaming about small and efficient programs of the old days but you cannot really compare the feature list of DPaint and Photoshop, the latter offers literally thousand times more features than the former so of course it is going to be bigger.

Apples and oranges really for a host of reasons, the most notable of which is that the x86 architecture is very much alive and well and, seemingly, always in next gen development, while R&D of the 68K series in the 1990s was more uncertain and expensive. The upshot of that is that Adobe has little incentive to reduce the bloat of Photoshop when adding features (whether they're useful or not), as they know that PC users can cheaply upgrade to a faster system and that Intel/Nvidia will cover their arses in the longer term.

In contrast, coders like Paul Nolan (of Photogenics fame) had little choice but to optimise his code as the more features he added to Photogenics, the smaller the Amiga market would be that had the specs to run his product (as obviously far fewer Ami users could afford to upgrade given the expense of 68K accelerators/gfx cards in the 90s). Believe me, there is a big difference in the list of features between early releases of Photogenics (v1.x-3.x) and the last major release V5.x, but not nearly as much of a divide in bloat and required/recommended specs (cf. early and later versions of Photoshop). IIRC Paul Nolan did such a good job that just about all versions of Photogenics ran in as little as 2MB ram (under KS 3.x), and he actually reduced memory requirements by ~1MB when he released the V4.x update.

I am not sure why we should regret that the movement continued after the Amiga? There are no reasons to think that programs would not have grown in size on that platform as well.

See above, but yeh more likely than not......assuming the Amiga would've migrated properly to more powerful architecture like x86, PPC or ARM in the 90s and then travelled along a path parallel to Windows/MacOS with AmigaOS.

Python has lost a lot of its popularity IMO, it is a nice language but on the Amiga it is going to be slow as hell, especially on 68k.

Why do you say that it's lost a lot of its popularity? I thought there would've been a bit of an upturn in the last few years - at least amongst hobbyists - given the popularity of the Raspberry Pi. Haven't noticed too much of a drop-off in use for scripting purposes either TBH because, as you say, it's quite a nice language.

Slow as hell on 68K? Probably for V3.x, but how many apps suited to the Amiga would actually require V3.x as opposed to V2.x? I ask this as early Ami versions (V1.4-2.0) required 020/030 (see Aminet), so I'm guessing 040/060 would be adequate for V2.7. As you said, it might be all academic as apps suited to the Amiga may not have all that much Python code anyway.

[Ratteler]

All these points about the Amiga being a "dead" platform are why my wish list are all OpenSource tools with a cross development history. We're not getting Amiga specific development anymore, but if we can benefit from some of FOSS community developments at least that is something.

With the Vampire FPGA not only might there be the necessary speed at a NON-ridiculous price for performance level, but the very nature of the board allows for upgrades to adapt the processor, graphics... if the 500's are going the way I hope with some kind of USB I/O in the redesign, we talking about a pretty serious game changer. An end to the decline of the platform.

Even better, the OpenSource tools will actually benefit from the Amiga hardcore developers figuring out how to solve the problems of providing the features without the code bloat that the FOSS tools might have from the platforms they were originally developed for. Development from Amiga can trickle back UP the OpenSource code base making those apps more portable. This helps THEM get other resource restricted platforms like cell phones and tablets.

[Thorham]

Originally Posted by ReadOnlyCat

This said, like every other skill out there, if you train you will get better.

Yeah, of course, but it takes years to become really good at it.

[idrougge]

Originally Posted by Ratteler

All these points about the Amiga being a "dead" platform are why my wish list are all OpenSource tools with a cross development history. We're not getting Amiga specific development anymore, but if we can benefit from some of FOSS community developments at least that is something.

Yeah, let's have bloatware with 70s user interfaces and dependency hell. Add a package manager to handle your dependency hell and there you have it – Linux with half the performance of a Raspberry Pi at ten times the price.

Originally Posted by Ratteler

With the Vampire FPGA not only might there be the necessary speed at a NON-ridiculous price for performance level, but the very nature of the board allows for upgrades to adapt the processor, graphics... if the 500's are going the way I hope with some kind of USB I/O in the redesign, we talking about a pretty serious game changer. An end to the decline of the platform.

The Amiga's decline ended several years ago, when it hit the rock bottom.

Originally Posted by Ratteler

Even better, the OpenSource tools will actually benefit from the Amiga hardcore developers figuring out how to solve the problems of providing the features without the code bloat that the FOSS tools might have from the platforms they were originally developed for. Development from Amiga can trickle back UP the OpenSource code base making those apps more portable. This helps THEM get other resource restricted platforms like cell phones and tablets.

As far as I can tell, the absolute vast majority of Vampire buyers are people who like shorter loading times while playing Cannon Fodder in WHDLoad or watch the icons display faster in Workbench – just like most 060 cards have ended up in the hands of rich collectors. I don't see any of those supposed hardcore developers standing in line to get a Vampire so that they can help solve the problems of the open source world using a non-existent toolchain.

[ReadOnlyCat]

Originally Posted by DrBong

[about Python]Why do you say that it's lost a lot of its popularity? I thought there would've been a bit of an upturn in the last few years - at least amongst hobbyists - given the popularity of the Raspberry Pi. Haven't noticed too much of a drop-off in use for scripting purposes either TBH because, as you say, it's quite a nice language.

It is a very subjective impression of mine (hence the IMO) which might not have much basis in reality. I used Python a lot and liked it from 2007 to 2012 but I always felt that v3 would cause a dip in the popularity of the language because of the necessity to re-code a lot of legacy Python code. Also, I am under the impression that Python also serves as some kind of a gateway introduction to even higher level languages such as Scala, Haskell and the likes and eventually, many Python adept will end up using these more powerful alternatives instead.

But as I said, this very well could be wishful thinking.

Originally Posted by DrBong

Slow as hell on 68K? Probably for V3.x, but how many apps suited to the Amiga would actually require V3.x as opposed to V2.x? I ask this as early Ami versions (V1.4-2.0) required 020/030 (see Aminet), so I'm guessing 040/060 would be adequate for V2.7. As you said, it might be all academic as apps suited to the Amiga may not have all that much Python code anyway.

I meant 68000, since my target are the stock OCS/ECS models.
You are right that there probably aren't many Python apps on the Amiga and it is a bit of a shame that the language is not used more for scripting since it is so versatile and has such a rich library. I never managed to get interested by Arexx even though I tried at least two times but Python would be really great for quick prototypes and glueing code.
Ideally though I'd wish Haskell was available on the Amiga since it can produce really good low level code but we would need a 68k backend for LLVM and it is not there yet.

Which makes me think! -> this is what we need! A group effort to finish the 68k backend of LLVM.

[alpine9000]

Originally Posted by ReadOnlyCat

Which makes me think! -> this is what we need! A group effort to finish the 68k backend of LLVM.

This is a huge task. I started doing an LLVM backend for the SH-2 processor which has small instruction set and simple register set compared to the 68k. And gave up when I realised that I would not be able to make it generate code even close to the quality GCC 5 does.

Unless we can find a similar backend to copy (one with separate data and address registers would be a good start) it's a massive job to get something that would be even close to VBCC.

And then, we are really talking about doing X backends (vanilla 68000, 68030, 68030/68040 with FPU, 68060, Vampire ?)

[ReadOnlyCat]

Originally Posted by alpine9000

This is a huge task. I started doing an LLVM backend for the SH-2 processor which has small instruction set and simple register set compared to the 68k. And gave up when I realised that I would not be able to make it generate code even close to the quality GCC 5 does.

Unless we can find a similar backend to copy (one with separate data and address registers would be a good start) it's a massive job to get something that would be even close to VBCC.

I am not surprised. When I looked into LLVM's backend documentation I was disappointed to learn that one needed to write code to support a new CPU because they had not designed the code generation system to be fully data driven.
I was pretty pissed because the whole promise of LLVM had been to be as modular and flexible as possible to fix gcc's design mistakes which also prevented its backend to be 100% data driven.

This said, there is already a team working on it but I'm not sure what the status is.
I must admit that I would not care too much about code quality, speed or size in the beginning as long as I can compile Haskell or c++ code. One can always recompile later when the backend has improved.

Originally Posted by alpine9000

And then, we are really talking about doing X backends (vanilla 68000, 68030, 68030/68040 with FPU, 68060, Vampire ?)

I guess the wise thing to do would be to support the 000 first, then add subsequent models in the chronological order of their release.
My opinion is probably not representative but 000, 010 and EC020 would be more than enough for me (stock, stock, stock, did I mention stock? ).

[Ratteler]

Originally Posted by DrBong

I'd be surprised too if the Vampire market worked any better than the Cyberstorm/Blizzard PPC market. At least the PPC market in 1996/97 had a few commerical players still in the mix for software development.

I wouldn't. 060/PPC boards started with the disadvantage that for the same price per performance you got for an upgrade to discontinued Commodore hardware, you could get a whole PC system that ran current software.
$1200 USD for PPC accelerator vs a >$1000 complete PC with superior features was a ridiculous notion in the first place.
When they were trying to push PPC Amiga's, a $500 PC ran circles around the nearly $2000 PPC machine.

This was, in my opinion, the most anti-Amiga/Commodore thing they could have done. What made Commodore was the fact that it offered the best performance for the dollar. The Amiga's niche that kept it valuable was it's compatibility with NTSC/PAL.

The 060/PPC were way to expensive, had serious compatibility issues, and we not even top of the line for PowerPC machines.

I could see the 060 interest because it was still 680x0, but it was end of line and in such limited production the price point could never be feasible.

PowerPC was just retarded. The AmigaOS would have been better off if they just made it run on PPC Mac's. At least they were top of the line PPC machines for the price. They did NOTHING better, and NOTHING as good, for nearly 3x the price of the competition.

The Vampires looks to be around >$300!
There is no longer any competition expectation BECAUSE the Amiga is considered dead. People getting the boards most likely have a PC, or Mac.
Because it's an FPGA, your hardware investment can keep being optimized indefinitely simply by upgrading the core. It can also emulate other things by simply swapping the core.

Times have changed and anyone in that situation now is looking to code the next big app for Android or iOS. Just about all the Amiga coders left are older and have real lives and responsibilities to pay for, not to mention don't believe that there is any 2nd, 3rd or 4th coming of the Amiga (Vampire, NG or otherwise)!!! Consequently, most software that will appear for the Vampire - assuming it establishes some sort of user base over time - will follow the status quo IMHO and consist of ports of existing software (predominantly open source as too few end-users will be willing to pay for licensed products).

I agree with that. But ports of OpenSource software is nothing to turn your nose up at. Blender alone, coming with the Blender Game Engine can offer enough entertainment value to make it worth it.

Overly-simplistic and too black & white (things the Ami retro community has never ever been!), particularly if you've spent any real time here on EAB.

Well, you're entitled to your opinion, but that's what I see in both the Commodore and Amiga community.
You don't see any one complaining that it's not "real" if it doesn't run on a Mac 128K, and demanding that the 400K floppy be the defacto standard for retro Mac's.
The 8 bit Commodore market WORSHIPS the 1541 instead trying to wipe it from history and overcome it's limitations.

It's the same with Amiga 880k floppy.
The fact is the community is divided by two very different goals. One group that just wants to play old games, and one that likes push the limits of the hardware and software.

Complete and utter nonsense. That's akin to saying that developing for anything but SuperCPU set-ups or equivalent on the C64/128 is just a lot of wasted effort!

It is. It's a complete waste of effort because you are programming either solely for yourself, or for such a small and declining market that it might as well be for yourself. Until someone makes an FPGA SuperCPU clone for the 64 and 128, programming for it mental masturbation.
That's fine. Cranking out a LOAD"*",8,1 is normal and healthy.
But let's not pretend it's anything a community can enjoy.
If there only a few thousand Amigans, I bet there are less than 100 SuperCPU users.

Apples and oranges really for a host of reasons, the most notable of which is that the x86 architecture is very much alive and well and, seemingly, always in next gen development, while R&D of the 68K series in the 1990s was more uncertain and expensive. The upshot of that is that Adobe has little incentive to reduce the bloat of Photoshop when adding features (whether they're useful or not), as they know that PC users can cheaply upgrade to a faster system and that Intel/Nvidia will cover their arses in the longer term.
In contrast, coders like Paul Nolan (of Photogenics fame) had little choice but to optimise his code as the more features he added to Photogenics, the smaller the Amiga market would be that had the specs to run his product (as obviously far fewer Ami users could afford to upgrade given the expense of 68K accelerators/gfx cards in the 90s). Believe me, there is a big difference in the list of features between early releases of Photogenics (v1.x-3.x) and the last major release V5.x, but not nearly as much of a divide in bloat and required/recommended specs (cf. early and later versions of Photoshop). IIRC Paul Nolan did such a good job that just about all versions of Photogenics ran in as little as 2MB ram (under KS 3.x), and that he reduced memory requirements by ~1MB when he updated to V4.x.

This is the kind of things that excites me about development on the Amiga. It's sad that he hasn't taken this code to Android, IOS and OS X.
Lean, tight code has a place on mobile. I'm on my 5th or 6th android phone and I still don't have a decent paint program.
I would buy Photogenics multiple times if it ran on a platforms I use.

This is why I'm pushing Python for he Amiga. Make it work there, and it should scale nicely.

[alpine9000]

Originally Posted by Ratteler

It is. It's a complete waste of effort because you are programming either solely for yourself, or for such a small and declining market that it might as well be for yourself. Until someone makes an FPGA SuperCPU clone for the 64 and 128, programming for it mental masturbation.
That's fine. Cranking out a LOAD"*",8,1 is normal and healthy.
But let's not pretend it's anything a community can enjoy.
If there only a few thousand Amigans, I bet there are less than 100 SuperCPU users.

I think writing software for yourself is on of the most rewarding things you can do as a programmer. Hardly a waste of effort.

Porting crap unix software to an even worse closed source dead operating system just because someone made an awesome hardware emulator seems like more of a waste of effort to me.

[Ratteler]

Originally Posted by alpine9000

I think writing software for yourself is on of the most rewarding things you can do as a programmer. Hardly a waste of effort.

Porting crap unix software to an even worse closed source dead operating system just because someone made an awesome hardware emulator seems like more of a waste of effort to me.

Well, I can't argue with your opinion of what makes you happy.
To me, it's like creating a painting that no one will ever see.
Even if you have a unique problem to solve, some one MIGHT have the same problem and want to solve it.
I just hope most of the community doesn't agree with you, because unlike a painting, code that is never shared is unlikely to be seen even after you're dead and your SuperCPU has been sold on Ebay.

[alpine9000]

Originally Posted by Ratteler

Well, I can't argue with your opinion of what makes you happy.
To me, it's like creating a painting that no one will ever see.
Even if you have a unique problem to solve, some one MIGHT have the same problem and want to solve it.
I just hope most of the community doesn't agree with you, because unlike a painting, code that is never shared is unlikely to be seen even after you're dead and your SuperCPU has been sold on Ebay.

Just because you write if for yourself doesn't mean you can't release it.

If you take $$$$ out of the equation, I would get more satisfaction writing something I need and enjoy using than I would writing something that I didn't particular care about just because I thought more people wanted to use it.

[Ratteler]

Originally Posted by alpine9000

Just because you write if for yourself doesn't mean you can't release it.

If you take $$$$ out of the equation, I would get more satisfaction writing something I need and enjoy using than I would writing something that I didn't particular care about just because I thought more people wanted to use it.

Ok. I respect your opinion.
You're not hurting anyone, so whatever you do to enjoy yourself is fine.

But getting back to the topic, What Software does 68K Amiga need? I think there is an implication that we are talking about the collective merit of the software, and not the personal joy of the art of programing.

[modrobert]

Not trying to derail the topic further, but since the discussion tends to go this way repeatedly.

Here are some relevant statistics of sales:
https://en.wikipedia.org/wiki/Usage_...vice_shipments

TL;DR: Android has 53% (1.3 billion devices) of the total market in 2015 according to Gartner, all devices that runs an operating system of any kind are included (Yes, PC desktop with Windows included at 14%, decreasing rapidly every year).


When asked, these are the desktop operating systems developers want to use in 2016:

http://stackoverflow.com/research/de...erating-system

[ReadOnlyCat]

It is unrealistic to expect outside coders to suddenly grow an interest in writing software for a now obsolete operating system.
IDrougge derided package managers but they solve a problem which is inherent in building software from modular blocks and AmigaOS definitely lacks dependency management and an enormous amount of modern features which are necessary to attract users and coders.
And the community isn't going to fix these so it is time to put to rest the idea that AmigaOS is going anywhere.

Now, this does not mean that outside coders will not be interested by the unique challenges offered by the equally unique hardware architecture of the machine. Retro coding is a thing and the Amiga has plenty to offer in this domain and it is a much easier target than consoles of the same era.

What we lack are tools which facilitate targeting this retro system. Many exist but to a large extent everyone still seems to rewrite their own toolchain and this is clearly a waste of time.
Hannibal made available his own effort and we could probably standardize around this initiative.
For this we'd need to know what are everyone's need for developments and tailor a toolchain modular enough to accommodate most people's use cases and habits.
Sure there are a few grumpy coders who will stay persuaded that their current setup is the best and won't even consider improved solution using whatever rationalization they can think of but at least newcomers and open minded people should see value in benefiting from one another's efforts instead of reinventing the wheel.

So here are my questions:
When developing on the Amiga, which toolchain do you use?
Why?
What is the process you follow for your iterations?
What parts are you willing/desiring to outsource?
Which tools do you wish existed which would make you faster when coding?

[Thorham]

Originally Posted by ReadOnlyCat

And the community isn't going to fix these so it is time to put to rest the idea that AmigaOS is going anywhere.

AmigaOS can't go anywhere anyway. Efforts of the community are not relevant to this. It's just too far behind, and I'd rather see it go (Amiga hardware can stay forever, of course).

Originally Posted by ReadOnlyCat

When developing on the Amiga, which toolchain do you use?

A heavily customized FrexxEd that calls the Barfly assembler on my A1200 with 50mhz 68030.

Originally Posted by ReadOnlyCat

Why?

Because I like 68k and using my Amiga to program in asm.

Originally Posted by ReadOnlyCat

Which tools do you wish existed which would make you faster when coding?

Full memory protection.