Should game devs have to support new hardware?

[Valken]

The "developers" should support whatever the heck they want to realize their end goal. If they want every "Amiga" to run the game or application, then let it be.

If they want to PUSH a new design that requires more than stock, then they should.

If everyone catered to the lowest common denominator, there would be no need for faster CPU, GPU or etc..

[grond]

Programmers shouldn't use buggy programming "techniques" that already caused such programs to crash on 020-060 boards back in the day. Leaving these bugs to WHDLoad people to clean up is just lazy. The Amiga never was a computer with fixed timings such as the C64 was but was mistreated like that too often. It's now over 30 later during which programmers should have learned that you simply must not make assumptions about blitter, chipmem or other speed in relation to the CPU speed. If you use self-modifying code (e.g. for speed reasons on a plain 68000), clear the cache if you are running on a CPU with one. It's no magic.

[phx]

[gimbal]

Yes I agree that developer should be "allowed" to do whatever they want, because they're not our personal servants Freedom is the right of all sentient beings, a wise robot once said.

[mcgeezer]

Originally Posted by grond

Programmers shouldn't use buggy programming "techniques" that already caused such programs to crash on 020-060 boards back in the day. Leaving these bugs to WHDLoad people to clean up is just lazy. The Amiga never was a computer with fixed timings such as the C64 was but was mistreated like that too often. It's now over 30 later during which programmers should have learned that you simply must not make assumptions about blitter, chipmem or other speed in relation to the CPU speed. If you use self-modifying code (e.g. for speed reasons on a plain 68000), clear the cache if you are running on a CPU with one. It's no magic.

Programmers don't know they're making mistakes, 30 years past or not and you're assuming they've been programming the Amiga for that length of time. Consider that a programmer only made an OCS game in the past and now wants to come back and make another game.

In your interpretation the programmer needs to go an read up on all the past programming mistakes and avoid them when all they want to do is make an A500 OCS game.

And that's why I put in option 4...

New programmer wanting to make a retro game on Amiga... bugger all that, too much to do... I'll go with the 8 bits or game consoles for making my game.

It's no magic.

[phx]

Originally Posted by mcgeezer

This is easy for you to say though PHX because you're an experienced programmer, But would you expect a programmer new to the Amiga and who wants to make a game to go and understand all of that stuff up front?

No, of course not. But usually a new programmer will take some time to develop his skills, and doesn't immediately publish a complete game as his first project.

I would hope that such a new programmer quickly finds one of the many startup-codes which take over the system in a safe way. So the game will automatically work on most hardware. Or use the OS correctly, then it works anyway.

There is no doubt that there are easier retro systems to learn coding on. But that's how it is.

EDIT: I agree with most others here, that every developer can do what he wants. I am the last one to criticise or discourage a developer, especially new developers. I just want to express my feeling how it looks when a game for A500 just crashes on A3000 or A4000. It simply doesn't make a good impression and can be avoided. It's a sign of quality.

[T_hairy_bootson]

Just do whatever you want to do, if some people want to call you lazy for using outdated techniques let them, then see what they push out.

[grond]

Originally Posted by mcgeezer

Programmers don't know they're making mistakes, 30 years past or not and you're assuming they've been programming the Amiga for that length of time.

Anyone is free to do what they want. If your code crashes on an A4000/040, it's not good code. There was a lot of bad code.

Testing code on different setups and seeing it crash is how you learn to avoid the mistakes that programmers did in the 80s because they never imagined or cared that one day there would be an Amiga with something else than a 7MHz 68000. Simply run your code in UAE without immediate blitter and CPU set at fastest, this should cover a lot of setups.

[robinsonb5]

I'd say the only thing you "have" to do, or at least should do, is decide very early on whether the project is just something you're going to do for your own enjoyment, or whether you're going to charge for it or seek financial support while developing it.

If the former then you're under no obligation to anyone except to yourself, to have fun! If people like the result, great.

If your audience ends up being customers, then things change slightly - and in that situation it's important to manage peoples expectations. Even in that situation it's fine to say "This will work on a base-model A500, or a base model A1200 - it might work on accelerated machines but it might not - I don't have such machines for testing" - provided that's communicated clearly to users, early enough.

Having said that, there are a number of people here with experience of making programs work on faster machines, and I daresay they'll be willing to offer help and advice if and when you need it.

[Galahad/FLT]

Ultimately a developer is free to support whatever they like.

My own personal view is that if what I'm doing works on what Commodore originally released, then I have supported what needs to be supported.

If however it works on 68060 and Vampire and other boards, then great, but if I have programmed correctly, it should work on those configurations anyway, if they don't, then its a hardware issue, not mine.

We are in a good place with regards programming on Amiga, most of us have learnt from the mistakes of the past what not to do with regards programming techniques, and thankfully most people are prepared to listen or ask if they are unsure.

[Minuous]

Originally Posted by mcgeezer

the programmer needs to go an read up on all the past programming mistakes and avoid them when all they want to do is make an A500 OCS game.

Not really, it is only a few pages in Chapter 1 of the Libraries RKRM. There are also a few Amiga Mail articles on the subject but they mostly just repeat the same guidelines.

[roondar]

Originally Posted by Minuous

Not really, it is only a few pages in Chapter 1 of the Libraries RKRM. There are also a few Amiga Mail articles on the subject but they mostly just repeat the same guidelines.

That is a very misleading statement.

[Minuous]

@roondar:

How so?

[roondar]

Because it's not true - even if you follow those guidelines, you'll still likely run into compatibility issues with some of the more exotic Amiga setups.

[Minuous]

Can you give an example?

[roondar]

Sure, most versions of Amiga OS won't run on a 68060 without the OS being patched.

Quite a lot of OS friendly software won't run on a 68040 or 68060 without patches or extra libraries either (and some still fails even then). There's a ton of software that won't run on a valid Amiga OS system that comes without the custom chips, such as the Dracovision. There's a number of OS legal programs that won't run if you have certain network cards installed. Some software doesn't like certain HDD controller cards or other expansions.

Apparently, some software doesn't run on Vampire cards (though I can't personally vouch for this one way or the other). Some software doesn't run correctly if you have certain accelerator cards, but will run on others. The list of compatibility issues is almost endless.

And yes, I'm only talking about OS-legal software here - not HW banging games and demos.

[Dunny]

Nobody needs or must support anything.

If the game requires an A500 with ECS 1MB then that's what it needs.
If it needs an A1200 with 030 and 4MB Fast RAM then that's what it needs.

It's not up to the developer to make sure their code runs on anything but that which they developed it on, or that which they desire it to run on. End of.

[kremiso]

i remember the disappointment of the first A1200 users, when they realized that many ocs games did not work at all on the new hw, at the time.
But today is imo different, due the tons of different configurations...
also, the complete playtesting/bugfixing would be a true pita

I also say that should be a developer free choice, it could be an ocs only title or a let's say 'Psygnosis Wipeout' like one

My vote would be for 'go to other platforms, grab the runned away big titles, and make them run again on papa system'

[Minuous]

If something like a DraCo or Vampire is incompatible with a real Amiga the fault is with those, surely.

[roondar]

Originally Posted by Minuous

If something like a DraCo or Vampire is incompatible with a real Amiga the fault is with those, surely.

Maybe, but where do you draw the line? I also mentioned the 68060, which has and still causes issues. Is that not a real Amiga?

The point I'm making is just that compatibility is not really simple, given enough variations to support. And there's many, many variations in Amiga land. It's actually a real-world problem as well, where just like on the Amiga, all sorts of extremely weird issues can crop up due to tiny (or not so tiny) HW differences between supposedly compatible bits of HW.