Stuff I Wish The XL/XE Had

[peteym5]

Correction I did not have the charts up on my webbrowser.

 

That memory address I was referring to is $BFFD not $BFFC. This has 4 or 5. 5 boots DOS before cartridge, 4 boots cartridge directly. Other values are for diagnostic cartridges. Maybe if $BFFD had the high bit set, it would re-assign "Pin A" as address line 13 and map $4000 to $BFFF to the cartridge. The XL/XE did something with TRIG3 to indicate the cartridge is present. I know someone suggested using TRIG2 and TRIG3 to read a second fire button because the XL/XE only had 2 joystick ports. POTS 4 through 7 could had also been used for something also, but 4 paddles is plenty for the Atari 8-bit.

[flashjazzcat]

Wow... accessing the banked RAM window from one's cart code would then be lots of fun. And what if you want 64KB of ROM, or 128KB? This seems like a solution without a problem, since bank switching code isn't such a big deal, especially with modern assemblers.

[peteym5]

The ideal would be map 32K ROM directly without bank switching. I am not sure when the first bank switching cartridges became available for the Atari 8-bit. I think most of the cartridge types before the XL was made was either 8 or 16K. Atari dropped the 2 cartridge left/right design that the 800 had because either of them could address 16K ROM anyway. Atari could had considered 32K being addressed to the cartridge port at that point and changing the purpose of ROM present pins to an extra address pins. This also would had been a better alternative to the 5200 design also. Atari could had also done the ECI port with the XL instead of the PBI port that essentially just did the same thing.

[fujidude]

 

Atari could had also done the ECI port with the XL instead of the PBI port that essentially just did the same thing.

 

I prefer the cartridge port + PBI to the ECI.

[Bryan]

I prefer the cartridge port + PBI to the ECI.

 

Me too. ECI was a cost-cutting measure.

[holgibo]

I bought my first 800XL in 1984, and I never had a problem with this <Option> thing... or with <Option> and <Start>. Just one look in the manual of the game or on the backcover...

 

For the built-in self test: An early attempt to move the quality gate to the customer, I think. Not very usefull, but it was also a first step to get into the "computerthingie" back in these days. After unpacking I read the into the manual and then I first made a selftest.

Okay, today you think: What a dumb waste of memory, but back in time it was some kind of useful.

 

What I would like to have in a standard 800XL:

1. 80 colums would be great, but in an extra graphicsmode.

2. GR.0 with a charset in 4 colors. This would be really great for utilities, menus and so on.

 

In my opinion the standard graphic capabilities of the A8 are good, the graphic style is one of the things why I love this machine still.

 

3. A faster built-in Basic. --> This Problem I got fixed with a built-in Turbobasic on an eprom, which a can activate by holding down the select functionkey at the boot. Very usefull.

[Keatah]

A "delete the parents" key so I could stay up all night and play games.

[Irgendwer]

Even if this "what if..." is pretty useless, I think little tweaks to the XL hardware could make the machine a lot more powerful with nearly no additional costs:

 

• 16 real color registers. If variable colours are an GTIA architecture problem, introduce at least meaningful static colours to the useless ones in "Graphics 10"
  (like adding a static grey scale palette with 7 buckets)
• Make GRAFPM also /readable/. Why? It would allow PMG-DMA abuse for other purposes - like sacrificing a player for Pokey sample replay. Just load GRAFPM data and store it to the volume register... (player could be out of screen but data transfer would be very easy and CPU friendly, e.g. also for colour values...)
• small memory management banking addition which would allow to select a page to act as zero-page (different pages possible for read/write)
  - would allow more lean and faster programs
• the same thing for the stack and task-switches would benefit too
• 5V on every PBI
  
• repeat the demand for freely selectable colours in GR. 0/8

Edited December 28, 2015 by Irgendwer

[ricortes]

 

Even if this "what if..." is pretty useless, I think little tweaks to the XL hardware could make the machine a lot more powerful with nearly no additional costs:

 

• 16 real color registers. If variable colours are an GTIA architecture problem, introduce at least meaningful static colours to the useless ones in "Graphics 10"

  (like adding a static grey scale palette with 7 buckets)

• Make GRAFPM also /readable/. Why? It would allow PMG-DMA abuse for other purposes - like sacrificing a player for Pokey sample replay. Just load GRAFPM data and store it to the volume register... (player could be out of screen but data transfer would be very easy and CPU friendly, e.g. also for colour values...)
• small memory management banking addition which would allow to select a page to act as zero-page (different pages possible for read/write)

  - would allow more lean and faster programs

• the same thing for the stack and task-switches would benefit too
• 5V on every PBI

• repeat the demand for freely selectable colours in GR. 0/8

 

IMHO: Because of the way ANTIC/GTIA display data, they divide a single byte into both pixel and color data, 16 registers wouldn't be as good of an enhancement as you would expect. You would need 4 bits of each screen byte to select the register so you end up with a colorful low resolution display.

 

Much better was Bob's dual ANTIC/GTIA system where the second set of LSI chips were able to toss an addition 8k of RAM data to screen in each frame. I mention this because it was available to be done anytime since the original 800/400 were done but not taken advantage of. I mean it is OK to speculate on stuff that wasn't able to be done because of cost of development but the added detail of tossing 16k of memory to screen was available from 1979 with the original chips. Personally I wish Atari had done this over development of the 1400/1450 computers.

Taken to its illogical conclusion<sic> there is probably no reason we could not have had quad ANTICs by the time the XLs came out. They would have exceeded most other contemporary game systems displays by a wide margin.

[Keatah]

Most of the things being discussed in this thread were not done because of cost. What a bummer..

[bbking67]

I'm sure the engineers wanted some improvements and would have liked to make various improvements, but business decisions usually superseded these changes... but even the proposed additions in the 1400XL/1450XLD mostly were the wrong things in my mind (though at the time sounded completely awesome to me).

[Irgendwer]

IMHO: Because of the way ANTIC/GTIA display data, they divide a single byte into both pixel and color data, 16 registers wouldn't be as good of an enhancement as you would expect. You would need 4 bits of each screen byte to select the register so you end up with a colorful low resolution display.

 

Now we have both, the low resolution and even there not 16 colours but only 9. 16 would definitely help, especially when mixing modes like in RIP/JAG. 16 registers would also be useful when mixing PMGs by three free definable colours and not only "ORing" two...

Edited December 29, 2015 by Irgendwer

[FifthPlayer]

 

Even if this "what if..." is pretty useless, I think little tweaks to the XL hardware could make the machine a lot more powerful with nearly no additional costs:

 

• 16 real color registers. If variable colours are an GTIA architecture problem, introduce at least meaningful static colours to the useless ones in "Graphics 10"

  (like adding a static grey scale palette with 7 buckets)

 

Even an "extra half-brite" or "complement" mode that provided 8 additional pseudo-registers, based off the existing color registers, could have been useful. Similar to what the A8 designers went on to implement in the Denise chip for Amiga.

Edited December 29, 2015 by FifthPlayer

[Keatah]

That goddamned amiga sucked away talent and resources and distracted atari from becoming all that much better. No wonder why I hate that ecosphere.

Edited December 29, 2015 by Keatah

[kenjennings]

Keatah: Atari did that on its own by ignoring the engineer's ideas. So, a lot of hardware and software folks packed their bags and set out to make a brave new world. That Atari talent is the same group that made the Amiga. Many here recognize the Amiga as the logical evolution from the Atari 8-bits.

 

Atari succeeded on the momentum of those initial, great engineers. But I think the Atari 800 was the last good decision by Atari management and it looks to me to have been pure stinking luck. Atari was a slow motion train wreck most of the time until its death in 1984.

 

What if Atari management had better listening skills?

 

The engineers knew the 2600 had a limited lifespan and the Atari 400 should have been packaged as a VCS-style super system to replace the VCS instead of confusing the home computer line. That would leave a distinctive computer line with the Atari 800 "personal graphics workstation." :-)

 

What if Atari let the engineers make the 68000-based system that they wanted? There could have been an updated, power-users' Atari graphics workstation using the 68000 released in 82 while Commodore's offering was the C64 videogame system. No contest. Commodore declares bankruptcy in 1983 and Jack Tramiel is later seen homeless, eating in a soup kitchen.

 

Since a fictional, competent Atari would have been happy to support the engineers' ideas an Amiga-class system would not have been delayed by engineers leaving the company, finding investors, and struggling to run a company to develop their dreams. So, an Amiga (or better) could have been available at the same time as the Macintosh. RIP Apple. If it had run unix then Atari would have owned the world.

 

 

We now depart from our retro fantasies and return to the leftover quart of eggnog.

[FifthPlayer]

What if Atari let the engineers make the 68000-based system that they wanted? There could have been an updated, power-users' Atari graphics workstation using the 68000 released in 82 while Commodore offering was the C64 videogame system.

 

Submitted as evidence for this statement: a talk by Joe Decuir and Ron Nicholson describing the development of the Amiga hardware. Decuir shows the very first block diagram (scribbled in his engineering notebook) of what the Amiga hardware architecture would look like.....a block diagram drawn while he was at Atari.

 

[almightytodd]

Um, how about Flash Jazz Cat's 8-bit GUI in ROM? Duh!

[kenjennings]

 

Decuir shows the very first block diagram (scribbled in his engineering notebook) of what the Amiga hardware architecture would look like.....a block diagram drawn while he was at Atari.

 

For anyone looking, that hand-drawn block diagram occurs at 22 minutes in the video. (He says that diagram was done in 1979.)

[Lynxpro]

Keatah: Atari did that on its own by ignoring the engineer's ideas. So, a lot of hardware and software folks packed their bags and set out to make a brave new world. That Atari talent is the same group that made the Amiga. Many here recognize the Amiga as the logical evolution from the Atari 8-bits.

 

Atari succeeded on the momentum of those initial, great engineers. But I think the Atari 800 was the last good decision by Atari management and it looks to me to have been pure stinking luck. Atari was a slow motion train wreck most of the time until its death in 1984.

 

What if Atari management had better listening skills?

 

The engineers knew the 2600 had a limited lifespan and the Atari 400 should have been packaged as a VCS-style super system to replace the VCS instead of confusing the home computer line. That would leave a distinctive computer line with the Atari 800 "personal graphics workstation." :-)

 

What if Atari let the engineers make the 68000-based system that they wanted? There could have been an updated, power-users' Atari graphics workstation using the 68000 released in 82 while Commodore's offering was the C64 videogame system. No contest. Commodore declares bankruptcy in 1983 and Jack Tramiel is later seen homeless, eating in a soup kitchen.

 

Since a fictional, competent Atari would have been happy to support the engineers' ideas an Amiga-class system would not have been delayed by engineers leaving the company, finding investors, and struggling to run a company to develop their dreams. So, an Amiga (or better) could have been available at the same time as the Macintosh. RIP Apple. If it had run unix then Atari would have owned the world.

 

 

We now depart from our retro fantasies and return to the leftover quart of eggnog.

 

 

But Atari Inc Advanced Research did design single and dual 68000 based systems that also used their SnowCap GUI running atop BSD. And that was without any input from the folks who left to create the Amiga.

[Lynxpro]

IMHO: Because of the way ANTIC/GTIA display data, they divide a single byte into both pixel and color data, 16 registers wouldn't be as good of an enhancement as you would expect. You would need 4 bits of each screen byte to select the register so you end up with a colorful low resolution display.

 

Much better was Bob's dual ANTIC/GTIA system where the second set of LSI chips were able to toss an addition 8k of RAM data to screen in each frame. I mention this because it was available to be done anytime since the original 800/400 were done but not taken advantage of. I mean it is OK to speculate on stuff that wasn't able to be done because of cost of development but the added detail of tossing 16k of memory to screen was available from 1979 with the original chips. Personally I wish Atari had done this over development of the 1400/1450 computers.

Taken to its illogical conclusion<sic> there is probably no reason we could not have had quad ANTICs by the time the XLs came out. They would have exceeded most other contemporary game systems displays by a wide margin.

 

Didn't the 1400XL/1450XLD have the feature that allowed the ANTIC to grab an extra 6K of RAM? Perhaps the FREDDIE chip allowed that to happen.

 

I'm not finding a resource that compares the STIA [super TIA] - designed for the proposed 2600 replacement system Sylvia - with the GTIA. The same goes for comparing that system's FRANTIC chip with the ANTIC… It seems the Home Computer Division got the idea to use the Votrax speech chip in the 1400XL/1450XLD from the Sylvia designs...

 

It's annoying to think of all of the custom silicon Atari Inc created and never used, or never used in consumer equipment. Not finishing the AMY so it would run easily. Not using Dual & Quad POKEYs [or the PORKEY], etc.

[Bryan]

A "delete the parents" key so I could stay up all night and play games.

What? The Smith & Wesson add-on?

[FifthPlayer]

 

 

But Atari Inc Advanced Research did design single and dual 68000 based systems that also used their SnowCap GUI running atop BSD. And that was without any input from the folks who left to create the Amiga.

 

 

Right, and those systems never saw the light of day. That's why the Amiga folks left, rather than try to build the Amiga at Atari. Atari management at the time was more focused on milking the 2600 than building the next-generation systems the engineers wanted to make (and knew had to be made, in order for Atari to remain competitive).

Edited December 31, 2015 by FifthPlayer

[peteym5]

I am not sure if the 65c02 was available when they were putting together the XL line in 1982. Adding those opcodes probably boost the capabilities of the machine and maintaining backward compatibility unless something used undocumented 6502 opcodes. I do not find many of those undocumented codes too useful. ML code be more compact.

[JamesD]

I am not sure if the 65c02 was available when they were putting together the XL line in 1982. Adding those opcodes probably boost the capabilities of the machine and maintaining backward compatibility unless something used undocumented 6502 opcodes. I do not find many of those undocumented codes too useful. ML code be more compact.

I think the 65C02 came out in 1983. I think the Enhanced Apple IIe was the first machine to use it. Even the first Apple IIe didn't have it.

[kiwilove]

Although it was the Wild West frontier of consumer videogame console/computer systems etc - there is the failure of recognising that history repeats - that the first wave was the 2600 - which was highly successful, and while the 5200 was not as successful - it was that significant upgrade from the 2600, though it was based on the Atari 400/800 computer lineage - the 7800 needed to be as significant an upgrade - which it wasn't due to too much penny pinching and lack of imagination in it's design? The Jaguar did not live up to it's hooha advertising - although Raiden was a good sign of it's sprite/scrolling/playfield capabilities but failed in the 3D area when compared to the PlayStation.

That a different videogame company captured the mass market - with the Atari 2600, Nintendo NES, Sega Genesis/Nintendo SNES, Sony PlayStation, Sony PlayStation 2 / Microsoft XBOX --- shows how technical superiority with very strong leading software (killer games) - and with mass market sales - the ability to deliver at a reasonable price. Though games prices have not really come down, nor new videogame consoles when they first appear.

 

I believe that the Atari 400/800/5200 hardware can still surprise us with it's software - if you are willing to step out and experiment with it. Perhaps even the 7800 too? But it's a harder beast to develop for?

 

Harvey