AGA Guide in PDF or similar format

[Leffmann]

Originally Posted by Photon

3.0 and 3.1 "N"DK, are they linked somewhere Leffman?

I've attached the 3.1 NDK below.

Originally Posted by Photon

If it is like I suspected then, there's only FMODE, palette switcher reg, and the new screenmode (and scroll/sprite mode...) bits. Huh.

There's a bit more, check the AGA.guide, but do you suspect there's something in there that's not documented? I don't know of any application, game or demo that uses an actual feature that is not known otherwise. WinUAE emulates everything out there, and Toni has the same HW documents everyone else has had for the last 15 years.

Originally Posted by Photon

Leffman, I think a combination of no new HRM released, first appearance of a homegrown HW doc, few new features could breed suspicions of mysteries. Can you personally guarantee no other features were added than those described in the AGA.guide?

Most of the AGA.guide, if that's what you mean by homegrown HW doc, is copied verbatim from Commodore's Lisa AA spec.

And yeah I guess there could be the odd side effect in the hardware that nobody knows of, but those can't really be compared to actual and implemented features.

[Photon]

Originally Posted by Leffmann

I've attached the 3.1 NDK below.

Originally Posted by Leffmann

There's a bit more, check the AGA.guide, but do you suspect there's something in there that's not documented? I don't know of any application, game or demo that uses an actual feature that is not known otherwise. WinUAE emulates everything out there, and Toni has the same HW documents everyone else has had for the last 15 years.

No, as I wrote I suspected that there are only those few display registers. But actual HW != HW docs, and if those HW docs consist of 1 doc and 1 TFA doc based on that 1 doc, then we only know of the registers they chose to include in that doc. That might correspond 1:1 to the docs, and it could just as easily not. We can rule out changes in chips that are identical to OCS/ECS (ie. separate chips, same package, same revision) but otherwise not. With 1 source for the docs, only those who have actually 'poked around where they shouldn't' (in some devised contexts) know if we have the whole truth.

If only the display chip has changed, the only possible extra bits or registers would be display ones and so might not be worth the trouble. It's perhaps unlikely that they would have put in, say, some CD32 affiliated registers in preparation for the console.

Originally Posted by Leffmann

Most of the AGA.guide, if that's what you mean by homegrown HW doc, is copied verbatim from Commodore's Lisa AA spec.

No I meant the AGA tutorial HTML thingy linked above (There was also some ascii file I saw at a mate's at some point).

Someone have the actual spec, and could upload it?

Originally Posted by Leffmann

And yeah I guess there could be the odd side effect in the hardware that nobody knows of, but those can't really be compared to actual and implemented features.

Wikipedia sums them up to what I suspected, and I do understand why you question my questioning. But it's nevertheless true that we can't claim to know until we've done what the C64 instruction hackers, MSX palette-switchers etc have done. This is of course true for ECS and OCS as well.

[TheDarkCoder]

Originally Posted by Lonewolf10

errr... how could they be looking forward? They had little or no cash, and although Dave Haynie (and others?) developed new hardware (AAA and Hombre) once done they could do nothing with it.





Ok, I think


Regards,
Lonewolf10

I still have Amiga magazines of the time the AGA machines were released. The magazines says that in the presentation of the new machines, C= publicly stated that they were planning a new chipset (I think they were referring to AAA) which would be offered in2 versions a lo-end and an hi-ed one. The hi -end would have used special VRAM to achieve better performance. These new chipsets were not supposed to be backward compatible with OCS-ECS-AGA.
That is what they said, what amount of real work already existed is not clear, Dave Haynie said many details.

Originally Posted by Photon

No, as I wrote I suspected that there are only those few display registers. But actual HW != HW docs, and if those HW docs consist of 1 doc and 1 TFA doc based on that 1 doc, then we only know of the registers they chose to include in that doc. That might correspond 1:1 to the docs, and it could just as easily not. We can rule out changes in chips that are identical to OCS/ECS (ie. separate chips, same package, same revision) but otherwise not. With 1 source for the docs, only those who have actually 'poked around where they shouldn't' (in some devised contexts) know if we have the whole truth.

If only the display chip has changed, the only possible extra bits or registers would be display ones and so might not be worth the trouble. It's perhaps unlikely that they would have put in, say, some CD32 affiliated registers in preparation for the console.


No I meant the AGA tutorial HTML thingy linked above (There was also some ascii file I saw at a mate's at some point).

Someone have the actual spec, and could upload it?

Wikipedia sums them up to what I suspected, and I do understand why you question my questioning. But it's nevertheless true that we can't claim to know until we've done what the C64 instruction hackers, MSX palette-switchers etc have done. This is of course true for ECS and OCS as well.

Theoretically you are right, it may be. And indeed Toni has discovered several undocumented features. However, I don't think there are magical register/features hidden in AGA. What would be the purpose of keeping them secrets also to the professional game develpoers?
Also, Dave Hanye and several others ex-C= engineers described many detail of what happened inside C=. They would have disclosed such secrets after C= bankrupts.

[Photon]

Another assumes I suspect It's not about trust, it's about knowledge. Certainly you must agree there wasn't the same openness and third-party book writing etc surrounding the release of AGA as it was for OCS?

You say secrets, I say there are lots of - simply reasons. There's inadvertent omissions, project-killing budgets, competence got fired, middle management decisions, planned revisions or addendums that are never released for whatever reason. Certainly any statements to the effect of preventing hardware tricks don't reassure, and neither does the timing of the bankruptcy.

But DarkCoder, you're right that it's theoretical and the probability of finding major features is low. Still, there was no HRM revision released and until we get the docs the game developers got, we don't know we have them. They could be the same as the AGA.guide, and they could be something else.

But the basic point was that even if we have them, regardless of whether they are same or not, and regardless of whether they describe only the already known registers or not, it's true for every platform that HW != HW docs.


Anyway.

I like to write and sometimes overdo it. I'm still interested in undocumented features, but we can't come much further in the discussion until someone pokes around where he shouldn't.

I'm sure the AGA.Guide is useful for h0ffman, I'll try to convert it to HTML and can host it if you want.

[SparkyNZ]

Originally Posted by Jherek Carnelia

I've put a pdf of the AGA registers in the zone. It may be useful; I also have a pdf of How To Code 7, if you would like that.

Hi. Can you please provide a link to the AGA document (PDF) you refer to. I've had a look and can't find it in the grandis FTP site.

Thanks.