What's with the history revisionism on (only) Pouet? :D
category: general [glöplog]
how many compos are there that are really new with those crazy fresh names like parties with their 2k+whatever for established types? the last 20+ compo types that got added were sizecoding related and won't be there without lovebyte happening. and then there is jaguar demo and 2 or 3 others that simply were not covered. the rest covers all current compo types thanks to general naming like "combined bla bla bla" or "platform demo" or some slash. sure there is some juggling going on sometimes but it is not like a crime to select "bla bla intro" instead of "bla bla 1k" if that does not clash with another compo at the party because it is clear from the prod type what is going on.
maybe the "problem" mentioned in the initial post would ne none (or nobody would care) if "amiga demo" would have been chosen as compo type which does not interfere with anything on the mentioned party. quote: "it's in fact simple."
maybe the "problem" mentioned in the initial post would ne none (or nobody would care) if "amiga demo" would have been chosen as compo type which does not interfere with anything on the mentioned party. quote: "it's in fact simple."
Pouet after fulfilling all publics wishes ;)
Hugs @ gargaj
You should see the emails.
Here's my attempt at a small summary on the subject, in order to help with educated decisions. Corrections welcome.
The distinction between OCS and ECS is largely irrelevant in a demoscene context, but the "classic Amiga" as a demo platform has no encompassing definition on Pouet, of which the OCS chipset is just one component - and in delineation not so much with ECS, but AGA. As a platform it would comprise of these minimum requirements:
- 68000 CPU
- 512k chipmem
- 512k any memory: can be chip, fast, slow - of which slow needs to be tested against, as it combines the disadvantages of the other two.
- Kickstart 1.2
- OCS Agnus and Denise chips
Under this definition OCS demos, old and new, would run on an Amiga 500 /w trapdoor memory expansion, and can (and new ones normally do, notwithstanding timing or OS interaction issues) run on all other Amigas as they have been sold in stores - with perhaps the rare exception of some A1000.
ECS is distributed over two chips. Mixed combinations are common in the field. In a demo context this isn't a big problem, as accidentally depending on ECS usually doesn't crash a demo, it just makes it look worse.
- ECS Agnus can blit larger regions, which is of very little relevance in demo practice.
- OCS Agnus is either PAL or NTSC. ECS can switch between PAL and NTSC. This is rarely relevant, a demo where this matters is e.g. Maga uncut by Zymosis, which on ECS runs in NTSC, and still works on OCS.
- ECS Denise is more flexible with regard to display windows and can show more sprites under certain conditions. This is relevant in demo practice, as people sometimes aren't aware of OCS sprite/overscan limitations. See for example Abstractica by Nah-Kolor, which suffers from sprite issues on OCS.
As Amigas have maintained a great deal of downward comptibility, you can usually get most productions to run on newer and more powerful machines. Problems usually arise from interaction with the OS and lacking synchronization on faster CPUs, but rarely from chipset differences. An example for accidental reliance on a newer OS version would be February by Insane.
The only known instance of an actual chipset incompatibility are some 7 bitplane modes under OCS, which cannot look the same on AGA without a distinct code path - but still this doesn't crash a production. An example of such an actual AGA/OCS incompatibility would be Imacennial by Dekadence, which denies running on AGA, as it knows it would look horrible, but it works totally fine on ECS.
Summary: A classic Amiga platform tag would be more helpful than a chipset delineation tag. I outlined a minimally intrusive solution above.
The distinction between OCS and ECS is largely irrelevant in a demoscene context, but the "classic Amiga" as a demo platform has no encompassing definition on Pouet, of which the OCS chipset is just one component - and in delineation not so much with ECS, but AGA. As a platform it would comprise of these minimum requirements:
- 68000 CPU
- 512k chipmem
- 512k any memory: can be chip, fast, slow - of which slow needs to be tested against, as it combines the disadvantages of the other two.
- Kickstart 1.2
- OCS Agnus and Denise chips
Under this definition OCS demos, old and new, would run on an Amiga 500 /w trapdoor memory expansion, and can (and new ones normally do, notwithstanding timing or OS interaction issues) run on all other Amigas as they have been sold in stores - with perhaps the rare exception of some A1000.
ECS is distributed over two chips. Mixed combinations are common in the field. In a demo context this isn't a big problem, as accidentally depending on ECS usually doesn't crash a demo, it just makes it look worse.
- ECS Agnus can blit larger regions, which is of very little relevance in demo practice.
- OCS Agnus is either PAL or NTSC. ECS can switch between PAL and NTSC. This is rarely relevant, a demo where this matters is e.g. Maga uncut by Zymosis, which on ECS runs in NTSC, and still works on OCS.
- ECS Denise is more flexible with regard to display windows and can show more sprites under certain conditions. This is relevant in demo practice, as people sometimes aren't aware of OCS sprite/overscan limitations. See for example Abstractica by Nah-Kolor, which suffers from sprite issues on OCS.
As Amigas have maintained a great deal of downward comptibility, you can usually get most productions to run on newer and more powerful machines. Problems usually arise from interaction with the OS and lacking synchronization on faster CPUs, but rarely from chipset differences. An example for accidental reliance on a newer OS version would be February by Insane.
The only known instance of an actual chipset incompatibility are some 7 bitplane modes under OCS, which cannot look the same on AGA without a distinct code path - but still this doesn't crash a production. An example of such an actual AGA/OCS incompatibility would be Imacennial by Dekadence, which denies running on AGA, as it knows it would look horrible, but it works totally fine on ECS.
Summary: A classic Amiga platform tag would be more helpful than a chipset delineation tag. I outlined a minimally intrusive solution above.
OCS/ECS is a spectrum for sure. Where does even the original pre-OCS (what does the O stand for again...) Denise without EHB mode fit in?
From a technical standpoint there are also demos (especially older ones) that require the original chipset and won't run on ECS. Same is true for older games as well. The last post on this thread summarizes it best.
Yes, but the chipset normally wasn't the cause - but assumptions about memory layout and indifference towards OS vectors and CPU caches.
The beauty of a changed definition - one taking into account the platform as a whole - would be that it unifies and allows for comparison of all productions from the oldest to the newest, under exclusion of few productions in a "mixed Amiga" bag. Giving the "neither AGA nor OCS" productions a home could be an incentive in itself - I know people who were entertaining the idea of making A500/A600 demos with more memory and HD.
Such productions would get a home and be pitted against e.g. https://www.pouet.net/prod.php?which=8262, which is incomparable to an A500 demo with its >700k chipmem requirement. Also it doesn't do this production a big favour to run it on an unaccelerated A500+ or A600, as it benefits tremendously from a faster CPU.
The beauty of a changed definition - one taking into account the platform as a whole - would be that it unifies and allows for comparison of all productions from the oldest to the newest, under exclusion of few productions in a "mixed Amiga" bag. Giving the "neither AGA nor OCS" productions a home could be an incentive in itself - I know people who were entertaining the idea of making A500/A600 demos with more memory and HD.
Such productions would get a home and be pitted against e.g. https://www.pouet.net/prod.php?which=8262, which is incomparable to an A500 demo with its >700k chipmem requirement. Also it doesn't do this production a big favour to run it on an unaccelerated A500+ or A600, as it benefits tremendously from a faster CPU.
Quote:
Giving the "neither AGA nor OCS" productions a home could be an incentive in itself
Are there really enough demos for expanded OCS/ECS machines to warrant a separate category? There's no category for unexpanded A1200 either, probably for the same reason.
A few dozen for OCS at most, it also depends on how pedantic you are about productions that were intended for OCS A500 but ended up working only on ECS, more RAM or newer OS versions, i.e. never received a fix.
I can't speak for AGA, it's a platform with different peculiarities, my impression is that AGA productions are more evenly distributed between accelerated and non-accelerated, but there are only a quarter as many entries as for OCS. Subdivision might be a consideration.
I can't speak for AGA, it's a platform with different peculiarities, my impression is that AGA productions are more evenly distributed between accelerated and non-accelerated, but there are only a quarter as many entries as for OCS. Subdivision might be a consideration.
AGA is, in my opinion, even more confusing than OCS/ECS. Is the prod for vanilla A1200 - and are pure fastmem expansions "vanilla" or "accelerated" (2x speed)? Is fastmem perhaps optional, but sort of expected? Is it buttery smooth on 030/50 MHz but will crash and burn on 040 and up (A4000)? Does it need an FPU? Does it need 060? If yes, is 50, 66 or 75 MHz recommended? Or maybe 030/50 but it's sorta mostly fine on 020/28? Is it a trackmo or is a hard drive required? How much fastmem does it need, if any? Anywhere beteen 4 and 64 megs have been "the standard" at some point. The list goes on. :D
(For the record, I think the current categories are fine.)
(For the record, I think the current categories are fine.)
The current categories are indeed fine.
The categories are just that; categories. They are not full hardware compatibility specifiers. People need to understand that they'll have to read the readme-files to get a better idea if a prod will likely work on their system.
We also don't have categories like "Windows, 4GB RAM, NVIDIA GPU". And even if we did stuff like that, demos can (and will) find ways to require something more that isn't encoded in the categories, creating needlessly complicated fractals of categories.
What might not be quite ideal is the hard-coded list of compos, as PS pointed out. Maybe each party should be able to specify compos in free-text. But the current solution has worked for us for over 20 years now. So it can't be that bad either.
The categories are just that; categories. They are not full hardware compatibility specifiers. People need to understand that they'll have to read the readme-files to get a better idea if a prod will likely work on their system.
We also don't have categories like "Windows, 4GB RAM, NVIDIA GPU". And even if we did stuff like that, demos can (and will) find ways to require something more that isn't encoded in the categories, creating needlessly complicated fractals of categories.
What might not be quite ideal is the hard-coded list of compos, as PS pointed out. Maybe each party should be able to specify compos in free-text. But the current solution has worked for us for over 20 years now. So it can't be that bad either.
How about people add an info file in their releases stating the hardware requirements of the prod, and people watching the prod actually read this info file and judge the demo accordingly? Crazy shit, I know...
or just do what normal people do and watch the youtube
*ducks*
*ducks*
There are two differences, about 17.000 Amiga productions already do make up a platform like the C64. And incompatibilities with the remaining ones are most of the times unintended, so wouldn't show up in an info file either.
@farfar, that's precisely the issue here. :-)
@farfar, that's precisely the issue here. :-)
absolutely tonedeaf...
"OCS/ECS is a spectrum for sure"
I don't think that is the only spectrum that is being evidenced in this topic.
I don't think that is the only spectrum that is being evidenced in this topic.
I run all amiga demos on my amiga, a computer I definitely own
Quote:
I run all amiga demos on my amiga, a computer I definitely own
I can confirm this. It is the same computer he uses when tracking music for amiga demos.
I often run Amiga demos on okkie's Amiga!
Just popped in to see if "it's time to introduce tags" had been mentioned yet
as in "trigger warning! might include amiga!"-tags?
bumping boing-balls may hurt you
