158 replies to this topic

[atariksi]

Divya16, on Fri Mar 12, 2010 3:21 PM, said:

I don't think people were dumping 5200 for 7800. I kept my 5200 and don't have a 7800, but would have kept both if I had them. Good to hear you have saved a 5200; that's all people can do now is to save what they have and keep the classic system alive.

Too bad some modern gamers missing out on some great stuff.

For some reason, I still prefer the A5200 games over A7800 perhaps it's because they are similar to Atari 400/800, the system I had for 30 years.

[kool kitty89]

atariksi, on Fri Mar 12, 2010 9:03 PM, said:

Another thing about joysticks, Atari 2600, Atari 400/800/XL/XE/XEGS, Atari 7800, Atari ST, C64, C128, Amiga, and quite a few others used the standard DB9 joystick; don't know why A5200 deviated from that given GTIA was quite capable of doing the digital joystick even without PIA's help.

I believe it's because of the added key inputs. With the analog joystick design, there's 1 fire button and 16 keys (including trigger 2), on top of that, they seem to have dedicated 8 pins to provide the keypad and start/select/reset inputs, then 1 pin for each trigger, at least based on this: http://pinouts.ru/In...00_pinout.shtml Again, it seems really odd to use that 4x4 array (and the 4x3 array in the VCS touchpad) rather than simply using 4 digital inputs (with 16 combinations). Granted, with the DE-9 port you'd still be short one pin regardless, or they could have dropped one of the keys to facilitate that. (probably reset)

Quote

Yeah, any graphics adventure like Myst, The Pawn, and King's Quest would require expanded storage/RAM capability that A5200 couldn't achieve.

Never heard of Princess/Mystery House. Which platform are those from?

http://en.wikipedia...._adventure_game I found them here, they're by On-Line games, who later changed to Sierra, and while Mystery House was only on Apple II it seems, Wizard and the Princess is listed as being on Apple II, Apple II Plus, Atari 400/800, Commodore 64, and IBM PC according to Wikipedia's page. (there's also a VIC-20 version posted on youtube)

[atariksi]

kool kitty89, on Sat Mar 13, 2010 3:18 AM, said:

atariksi, on Fri Mar 12, 2010 9:03 PM, said:

Another thing about joysticks, Atari 2600, Atari 400/800/XL/XE/XEGS, Atari 7800, Atari ST, C64, C128, Amiga, and quite a few others used the standard DB9 joystick; don't know why A5200 deviated from that given GTIA was quite capable of doing the digital joystick even without PIA's help.

I believe it's because of the added key inputs. With the analog joystick design, there's 1 fire button and 16 keys (including trigger 2), on top of that, they seem to have dedicated 8 pins to provide the keypad and start/select/reset inputs, then 1 pin for each trigger, at least based on this: http://pinouts.ru/In...00_pinout.shtml Again, it seems really odd to use that 4x4 array (and the 4x3 array in the VCS touchpad) rather than simply using 4 digital inputs (with 16 combinations). Granted, with the DE-9 port you'd still be short one pin regardless, or they could have dropped one of the keys to facilitate that. (probably reset)

Quote

Yeah, any graphics adventure like Myst, The Pawn, and King's Quest would require expanded storage/RAM capability that A5200 couldn't achieve.

Never heard of Princess/Mystery House. Which platform are those from?

http://en.wikipedia...._adventure_game I found them here, they're by On-Line games, who later changed to Sierra, and while Mystery House was only on Apple II it seems, Wizard and the Princess is listed as being on Apple II, Apple II Plus, Atari 400/800, Commodore 64, and IBM PC according to Wikipedia's page. (there's also a VIC-20 version posted on youtube)

You can easily use a keyboard controller rather than the A5200 keypad and use the same DB9 type port. And most A5200 games don't even use most of the keypad. Most common keys are * and # to set level and players and then start/reset/pause. So 5 keys was sufficient and you get more than that with the touchpad or keyboard controller.

[Ransom]

atariksi, on Sat Mar 13, 2010 2:43 PM, said:

You can easily use a keyboard controller rather than the A5200 keypad and use the same DB9 type port. And most A5200 games don't even use most of the keypad. Most common keys are * and # to set level and players and then start/reset/pause. So 5 keys was sufficient and you get more than that with the touchpad or keyboard controller.

Good thoughts. And of course you don't really need # and * to set the level and players; joystick movements would do that fine (as on the 7800, among others). The start button begins the game if at the select screen, or pauses while in the game, as with the NES among others. That eliminates the pause button. At the end of the game, the final screen remains until the Start button is pressed (so you can see any final message of the game or whatever). When Start is pressed, you go back to the title/select screen, where you can either change the settings or hit Start again to play.

So you end up with a joystick, a start button, and two fire buttons (let's face it, by that point a second fire button was used in so many arcade games that not having it would have made things awkward).

But without a keypad, you'd have trouble playing Star Raiders (a very hot game at the time). It's much more awkward to push buttons on a second controller than on the one in your hands. Of course, if Atari could have invented the PS2 controller in 1982, we'd have been all set.

Perhaps they could have been a little more sophisticated and allowed for multiple controllers, auto-detected on power up. A standard controller with digital stick, two action buttons, and Start. A "pro" controller with analog stick, two action buttons, Start, and a keypad (for the games that benefit from analog control, and perhaps with selectable self-centering like with the Apple ][ controllers of the time). And of course the awesome 5200 trak-ball as a third controller option.

As far as shape...boy, if Atari had invented ergonomic controllers, they would have saved us all a lot of pain up through the early 90s at least!

[Divya16]

Ransom, on Sat Mar 13, 2010 4:32 PM, said:

atariksi, on Sat Mar 13, 2010 2:43 PM, said:

You can easily use a keyboard controller rather than the A5200 keypad and use the same DB9 type port. And most A5200 games don't even use most of the keypad. Most common keys are * and # to set level and players and then start/reset/pause. So 5 keys was sufficient and you get more than that with the touchpad or keyboard controller.

Good thoughts. And of course you don't really need # and * to set the level and players; joystick movements would do that fine (as on the 7800, among others). The start button begins the game if at the select screen, or pauses while in the game, as with the NES among others. That eliminates the pause button. At the end of the game, the final screen remains until the Start button is pressed (so you can see any final message of the game or whatever). When Start is pressed, you go back to the title/select screen, where you can either change the settings or hit Start again to play.

So you end up with a joystick, a start button, and two fire buttons (let's face it, by that point a second fire button was used in so many arcade games that not having it would have made things awkward).

But without a keypad, you'd have trouble playing Star Raiders (a very hot game at the time). It's much more awkward to push buttons on a second controller than on the one in your hands. Of course, if Atari could have invented the PS2 controller in 1982, we'd have been all set.

Perhaps they could have been a little more sophisticated and allowed for multiple controllers, auto-detected on power up. A standard controller with digital stick, two action buttons, and Start. A "pro" controller with analog stick, two action buttons, Start, and a keypad (for the games that benefit from analog control, and perhaps with selectable self-centering like with the Apple ][ controllers of the time). And of course the awesome 5200 trak-ball as a third controller option.

As far as shape...boy, if Atari had invented ergonomic controllers, they would have saved us all a lot of pain up through the early 90s at least!

There's some auto-detection going on on the A5200 since it can tell Trackball vs. regular 5200 joysticks.

Here's the update of score:

How Atari Could Have Saved The 5200: 105 posts
Do you think the 5200 could have been saved? 105 posts

It's dead even-- this is going to be a photo finish. Regarding "ditching the 5200" topic, that seems to have lost enthusiasm-- more people prefer saving it.

[kool kitty89]

[quote name='Ransom' date='Sat Mar 13, 2010 2:32 PM' timestamp='1268519529' post='1963480']Perhaps they could have been a little more sophisticated and allowed for multiple controllers, auto-detected on power up. A standard controller with digital stick, two action buttons, and Start. A "pro" controller with analog stick, two action buttons, Start, and a keypad (for the games that benefit from analog control, and perhaps with selectable self-centering like with the Apple ][ controllers of the time). And of course the awesome 5200 trak-ball as a third controller option.

[quote]As far as shape...boy, if Atari had invented ergonomic controllers, they would have saved us all a lot of pain up through the early 90s at least![/quote]
The original form factor would be OK, something like the 3200/Sylvia mock-up and some other prototype controllers along with the 5200. (albeit the pre-5200 ones had an integrated paddle and I'd assume, digital joystick) Basically just a have the 5200 style form (perhaps slightly smaller), similar joystick placement (digital or analog), no keypad, and one button per side with hard plastic like the 7800, 2800, and paddles, and with a stiff, tough boot around the stick. (preferably along with springs if analog)
The coleco controllers were a bit like that too, but had the poorly placed, small, recessed buttons, and knobby joystick. (fine as a thumbstick, but not as a joystick, while the 5200 works in either fashion)

As for a controller using 2-buttons, a digital joystick plus a couple keys (start/select) is the Atari joyports are still limiting there. For a few keys (not for in-game action, but fine for start/pause/select) you could easily use unused combinations of the 4 directional lines (only 9 of 16 combinations used by the stick), so you get up to 7 key inputs that way, but for the added button (for in-game action), that's tougher. You'd have to make it able to work simultaneously with the joystick inputs and button-1, so not like the start/select/etc keys, and at that point you've exhausted the 5 digital lines so the other option is a pot line. (unless, of course, you remove one of the pot lines entirely in favor of a digital one, but then you los the possibility of 2 analog axes per port and paddle compatibility)
With such a configuration, you could indeed have the integrated paddle as well, like the 2700, 2800, and 3200 prototype seem to have.

Without PIA (or RIOT), it makes things tougher too, especially for multiple digital joystick to be used. (as atariski mentioned, you could get 1 digital joystick using the pot lines). I suppose you could have the joystick inputs read as keys with POKEY (treating the 8 direction as 8 different keys), but I think that would be a good deal inferior to PIA/RIOT/GTIA digital inputs. (not sure how it would compare to analog pot polling though)

[atariksi]

[quote name='kool kitty89' date='Sat Mar 13, 2010 8:07 PM' timestamp='1268528875' post='1963559']
[quote name='Ransom' date='Sat Mar 13, 2010 2:32 PM' timestamp='1268519529' post='1963480']Perhaps they could have been a little more sophisticated and allowed for multiple controllers, auto-detected on power up. A standard controller with digital stick, two action buttons, and Start. A "pro" controller with analog stick, two action buttons, Start, and a keypad (for the games that benefit from analog control, and perhaps with selectable self-centering like with the Apple ][ controllers of the time). And of course the awesome 5200 trak-ball as a third controller option.

[quote]As far as shape...boy, if Atari had invented ergonomic controllers, they would have saved us all a lot of pain up through the early 90s at least![/quote]
The original form factor would be OK, something like the 3200/Sylvia mock-up and some other prototype controllers along with the 5200. (albeit the pre-5200 ones had an integrated paddle and I'd assume, digital joystick) Basically just a have the 5200 style form (perhaps slightly smaller), similar joystick placement (digital or analog), no keypad, and one button per side with hard plastic like the 7800, 2800, and paddles, and with a stiff, tough boot around the stick. (preferably along with springs if analog)
The coleco controllers were a bit like that too, but had the poorly placed, small, recessed buttons, and knobby joystick. (fine as a thumbstick, but not as a joystick, while the 5200 works in either fashion)

As for a controller using 2-buttons, a digital joystick plus a couple keys (start/select) is the Atari joyports are still limiting there. For a few keys (not for in-game action, but fine for start/pause/select) you could easily use unused combinations of the 4 directional lines (only 9 of 16 combinations used by the stick), so you get up to 7 key inputs that way, but for the added button (for in-game action), that's tougher. You'd have to make it able to work simultaneously with the joystick inputs and button-1, so not like the start/select/etc keys, and at that point you've exhausted the 5 digital lines so the other option is a pot line. (unless, of course, you remove one of the pot lines entirely in favor of a digital one, but then you los the possibility of 2 analog axes per port and paddle compatibility)
With such a configuration, you could indeed have the integrated paddle as well, like the 2700, 2800, and 3200 prototype seem to have.

Without PIA (or RIOT), it makes things tougher too, especially for multiple digital joystick to be used. (as atariski mentioned, you could get 1 digital joystick using the pot lines). I suppose you could have the joystick inputs read as keys with POKEY (treating the 8 direction as 8 different keys), but I think that would be a good deal inferior to PIA/RIOT/GTIA digital inputs. (not sure how it would compare to analog pot polling though)
[/quote]

It's a bad idea to overload keystrokes on the joystick's 4 digital lines. That was a bad move on C64. You want to allow joystick to be independent of keystrokes. Also, there's mouse and other devices using joystick ports so 9 states isn't the only combinations.

I said using select lines and trigger lines for digital joystick; pot lines are slower to read.

[Ransom]

Sounds like we'll have to toss the analog joystick to get the second button then. That still leaves the +5V and one pot line so you can have one paddle on the port for paddle games.

[kool kitty89]

atariksi, on Sat Mar 13, 2010 8:10 PM, said:

It's a bad idea to overload keystrokes on the joystick's 4 digital lines. That was a bad move on C64. You want to allow joystick to be independent of keystrokes. Also, there's mouse and other devices using joystick ports so 9 states isn't the only combinations.

I said using select lines and trigger lines for digital joystick; pot lines are slower to read.

Yes, but you can't even get 2 full digital joysticks that way. Or you might if you use the 4 select lines and 4 trigger inputs for the directional inputs for 2 joysticks and had the buttons read as keys and/or though the pot lines. (1 button as a key, one though the pot line) 2 joysticks would be fine, and that's what they ended up cutting it down to anyway. (ver few 4-player games) Plus, any paddle based games could still be 4-player like on the VCS. Or you could keep ports 3 and 4 as well and have them as analog+key only. (cutting back to 2 ports might be simplest though)

What are GTIA's select lines currently used for on the 5200?

I do thing having at least 2 fire buttons is necessary though. (to be compatible with the VCS pinout, only 1 of the button inputs could use a pot line, the other would need to be the same pin as the original VCS joystick -as the 7800 did)

Still, what I meant by the key comment, was that I was wondering if reading the directional inputs as keys might be preferable (speed wise) to polling pots of analog controllers.

Ransom, on Sat Mar 13, 2010 8:41 PM, said:

Sounds like we'll have to toss the analog joystick to get the second button then. That still leaves the +5V and one pot line so you can have one paddle on the port for paddle games.

The button limitation is only a problem if you want digital sticks. With the analog joysticks you have 5 digital lines available; dedicating 2 for buttons, the remaining 3 would still allow for 7 keys/function buttons (ie start, select, pause, etc), or you could use them independently for 5 separate buttons. Without PIA or RIOT, you only get 8 digital lines from GTIA (4 select, 4 trigger), so without resorting to POKEY (for keys) you could have a total of 8 buttons using those. (so 2 per controller on a 4-port system, or 4 per controller if you opted for a 2-port system). Seeing as only 2 of the buttons per controller would be action buttons anyway, using 2 GTIA inputs and any additional ones as keys could work fine.

However, going for 2-ports only, or 2 main ports, and using 4 GTIA inputs per controller could facilitate use of digital joysticks too, as I address in my response to Atariksi. So, have the standard directional inputs wired to GTIA for ports 1+2, have button one on its normal digital line (but ready as a key by POKEY) and any additional inputs would be through the pot lines (as simple as 2 de-factor digital inputs, or more than that with varying resistances for different inputs -the former being the cleanest option). In this case, you could keep port 3 and 4 for keys and analog inputs only. (possibly for use with keypads, 4-player analog joystick games, or 8 player paddle games)

Edited by kool kitty89, Sun Mar 14, 2010 2:02 AM.

[Kripto]

An optional, large, arcade-style joystick in a heavy mount would have made the system seriously wanted by every kid around.

The non-self-centering analog stick was a deeply stupid idea but they looked damn cool, so I don't think their inclusion reduced sales.

[Ransom]

kool kitty89, on Sun Mar 14, 2010 1:56 AM, said:

The button limitation is only a problem if you want digital sticks.

Certainly I'd be satisfied with analog sticks, and I think most would if they had a self-centering option a la Apple.

[kool kitty89]

A stiffer rubber boot does help somewhat without proper centering springs too, late releases did use such. (very stiff initially, though they get looser as they wear)

Even with springs, the boot should have been tougher than the initial ones they used; those seem to have had a tendency to shred rather quickly. (with springs, not necessary as stiff a boot as the later 5200 controllers, but tough enough not to shred)

I think one thing that might have helped for analog control in general would have been to make the sticks somewhat shorter throw. (more limited range of motion)

And, of course, there could be a simple 9-state analog set-up with resistors to approximate an 8-way digital joystick. (though it would be slower to read as Atariksi has been mentioning -not sure how important that really is for most games though) The configuration should work just as well as the Wico controllers and such, (which I believe use that very mechanism)

Again, sticking to the stock controller ports, you could manage that fine without PIA, just not the standard digital joystick. (unless it was read in an alternate manner) Properly configured, it could be set-up to work with the standard paddles without an adapter, and the VCS keypads too. (if you used the same pins for the fire buttons as paddle buttons 1 and 2, those could work similarly as well -just without the start/select buttons on the paddles)


That of course is in such a case where direct 2600 compatibility is not considered; or at least native compatibility. (ie a separate module available, to match coleco -a front or side mounted expansion port would help for that)



However, given that relatively few games work best with an analog joystick (star wars being a prime example), it really isn't that preferable to use them standard as such rather than as accessories. Even without PIA, you could use the resistor+pot line set-up for an approximated digital joystick (only 8 directions possible, no actual potentiometers, just 9 different resistance values). The same form factor could be used of course, just the the joystick rather stiff, like the VCS ones, in a simpler pseudo-digital joystick. (which is what the old Gravis gamepad uses, and the Wico 5200 joystick) Plus there's GTIA's digital lines, of course, at least useful for 2 ports. (though with the resistor set-up, games could be made to easily work with both analog and pseudo-digital controllers)

Obviously some trackball games could be done better with an analog joystick than a digital one, but they could also be done worse, plus a dedicated trackball controller would be best. I'd prefer speed sensitive analog control to fixed position analog control for soem games though - think missile command uses the latter for the 5200, though the former would probably be best for a game like star wars. (reticule is centered when the stick is and moves with the stick rather liek with paddle games -opposed to the speed sensitive approach, which is like digital cursor movement, but at variable speeds, like moving a mouse -hence good for track-ball games)

Same would go for paddle based games, could be done well with an analog stick, but best with real paddles. It could be good for driving games though, perhaps better than the driving controllers.
The vectrex is a pretty good example of how reasonable contemporary analog control could work though, with many games being oriented towards 4/8 direction control. (though some do take greater advantage of the analog control)




With the original controllers, the simplest changes would to have been to use a tougher boot from the start and use 1 large hard capped plastic button per side. (having the start/pause keys more raised would have helped too, probably not reset -to avoid accidental use)
Omitting the keypad and going with the standard DE-9 port would have been pretty simple. (it would only mean slight modifications to games requiring keys for options on the 8-bit computers to use a simple menu instead, or a separate keypad for some other games needing it) They could have facilitated the separate keypad by allowing it to lock-on to the main controller for more convenient use. (perhaps using the keyboard controller's lock-together mechanism, or a new keypad better suited to fitting onto the main controller as an accessory)
Replacing the ports with a simple resistor set-up for 9 values would also have been simple, eliminating the centering and wide range of motion issues. (plus, a couple resistors set-up to provide the necessary values and simple wiring might even be cheaper than the potentiometer set-up)

Actual spring centering might have been a bit problematic in the set-up they used, I seem to recall that they had problems fitting a spring mechanism in there properly with initial attempts.

Edited by kool kitty89, Sun Mar 14, 2010 5:40 PM.

[atariksi]

kool kitty89, on Sun Mar 14, 2010 1:56 AM, said:

atariksi, on Sat Mar 13, 2010 8:10 PM, said:

It's a bad idea to overload keystrokes on the joystick's 4 digital lines. That was a bad move on C64. You want to allow joystick to be independent of keystrokes. Also, there's mouse and other devices using joystick ports so 9 states isn't the only combinations.

I said using select lines and trigger lines for digital joystick; pot lines are slower to read.

Yes, but you can't even get 2 full digital joysticks that way. Or you might if you use the 4 select lines and 4 trigger inputs for the directional inputs for 2 joysticks and had the buttons read as keys and/or though the pot lines. (1 button as a key, one though the pot line) 2 joysticks would be fine, and that's what they ended up cutting it down to anyway. (ver few 4-player games) Plus, any paddle based games could still be 4-player like on the VCS. Or you could keep ports 3 and 4 as well and have them as analog+key only. (cutting back to 2 ports might be simplest though)

What are GTIA's select lines currently used for on the 5200?

I do thing having at least 2 fire buttons is necessary though. (to be compatible with the VCS pinout, only 1 of the button inputs could use a pot line, the other would need to be the same pin as the original VCS joystick -as the 7800 did)

Still, what I meant by the key comment, was that I was wondering if reading the directional inputs as keys might be preferable (speed wise) to polling pots of analog controllers.

Ransom, on Sat Mar 13, 2010 8:41 PM, said:

Sounds like we'll have to toss the analog joystick to get the second button then. That still leaves the +5V and one pot line so you can have one paddle on the port for paddle games.

The button limitation is only a problem if you want digital sticks. With the analog joysticks you have 5 digital lines available; dedicating 2 for buttons, the remaining 3 would still allow for 7 keys/function buttons (ie start, select, pause, etc), or you could use them independently for 5 separate buttons. Without PIA or RIOT, you only get 8 digital lines from GTIA (4 select, 4 trigger), so without resorting to POKEY (for keys) you could have a total of 8 buttons using those. (so 2 per controller on a 4-port system, or 4 per controller if you opted for a 2-port system). Seeing as only 2 of the buttons per controller would be action buttons anyway, using 2 GTIA inputs and any additional ones as keys could work fine.

However, going for 2-ports only, or 2 main ports, and using 4 GTIA inputs per controller could facilitate use of digital joysticks too, as I address in my response to Atariksi. So, have the standard directional inputs wired to GTIA for ports 1+2, have button one on its normal digital line (but ready as a key by POKEY) and any additional inputs would be through the pot lines (as simple as 2 de-factor digital inputs, or more than that with varying resistances for different inputs -the former being the cleanest option). In this case, you could keep port 3 and 4 for keys and analog inputs only. (possibly for use with keypads, 4-player analog joystick games, or 8 player paddle games)

Keys are faster than POTs but still both are slower than digital lines like trigger and PIA directional inputs. There's a fast pot scan mode as well on A5200 and A800 that can read POTs at 7.5Khz (rather than 60Hz) (similar for PAL). KB is 15Khz. PIA is speed of processor (LDA 54016-- 4 cycles).

Currently, on A5200 they use the select lines of CONSOL to select which port gets use of the keypad-- only one keypad can be active at one time as they share the same POKEY for decoding the keys. Sel3 is used for internal speaker on A400/A800, Sel0..2 for START/SELECT/OPTION. Another advantage of SEL0..3 is that are I/O not just input just like PIA digital lines. Trigger lines are just input.

[kool kitty89]

atariksi, on Mon Mar 15, 2010 12:15 PM, said:

Keys are faster than POTs but still both are slower than digital lines like trigger and PIA directional inputs. There's a fast pot scan mode as well on A5200 and A800 that can read POTs at 7.5Khz (rather than 60Hz) (similar for PAL). KB is 15Khz. PIA is speed of processor (LDA 54016-- 4 cycles).

Currently, on A5200 they use the select lines of CONSOL to select which port gets use of the keypad-- only one keypad can be active at one time as they share the same POKEY for decoding the keys. Sel3 is used for internal speaker on A400/A800, Sel0..2 for START/SELECT/OPTION. Another advantage of SEL0..3 is that are I/O not just input just like PIA digital lines. Trigger lines are just input.

OK, but the question is how to best distribute the available I/O functionality among the controllers. If you wanted 4 controller ports, all equally compatible (in terms of the controllers used) and read similarly, then you couldn't have GTIA reading all the joystick inputs. In this case, without PIA, you're limited to 1 trigger input per controller, 1 select for key enabling, and the rest read as pots or keys in various possible configurations. So, 1 digital line on each port read by GTIA, the other 4 by POKEY, and the pots by POKEY, of course. --With what you provided, it seems like the fast pot scanning might be faster than keys if the pots are polled simultaneously. (rather than the keys which must be read concurrently -so 2x as fast for pot scanning when all ports are used, no speed advantage for 2 controllers though, and a disadvantage for pot scanning with a single controller -for fast scanning mode)

If, however, you limited the number of ports, or the purpose of some ports, things would change. So, with only 2 controller ports you could have all 8 GTIA digital lines dedicated to the joysticks of the 2 controllers, then read the buttons through the POT lines (if you added multiple resistors you could have more buttons as well, so possibly 2 fire buttons plus a few function buttons -start/select/pause/etc), and you wouldn't have the select lines for key enabling. You could have additional analog ports, and 1 port capable of using keys.

Given that they did end up switching to 2-ports later on (on both the 5200 and computers), with very few games allowing >4 simultaneous players using joysticks (or >4 simultaneous paddles), that doesn't seem to be too big of a deal. Also, I don't think any games required multiple keypads and joysticks to be used simultaneously (correct me if I'm wrong). A few games might have required 2 keypads, but I think any that required a joystick with a keypad were single player only. So, one could have 2 controller ports using DE-9 connectors with the standard pinout, with the directional lines connected to GTIA's digital inputs but you wouldn't be able to use the standard 8-bit joysticks then, as you're still missing the 5th digital input which can't be read as a key due to the select lines being used already. (you could easily use proprietary joystick using the analog lines for buttons, but the joysticks and keypads for the VCS wouldn't be compatible -paddles would be fully compatible though)
POEKEY's key inputs could either be run to the expansion port or to a dedicated keypad/keyboard port, as above. (12-key keypads should work fine with the normal controller ports though, just not the ones of the VCS due to the lack of the 5th digital line)

Any other options for 2 controller ports that allow full compatibility with the VCS/8-bit joysticks would again require use of the select line for enabling keys, so again, limiting things. Or could the key inputs could be distributed without select lines? I assume there is an 8-bit path (8 key input lines) for the key inputs (255 possible KEYs), so could one split that to 2x 4-bit paths? (or further divided) In the case of 4 GTIA inputs for 2 controllers, you only 1 one more available pin, so a 1-bit path per controller would be the only possible. (the trigger button line in the VCS pinout) If this is indeed possible, then you could easily have a 2-port system with full peripheral compatibility with the VCS as well as additional enhanced controllers, just as the VCS has the potential for, and 7800 did expand upon to some extent. (using pot lines for additional buttons, unused combinations on the digital lines for start/pause/etc, analog joystick, keypads, etc)

[atariksi]

kool kitty89, on Mon Mar 15, 2010 5:59 PM, said:

atariksi, on Mon Mar 15, 2010 12:15 PM, said:

Keys are faster than POTs but still both are slower than digital lines like trigger and PIA directional inputs. There's a fast pot scan mode as well on A5200 and A800 that can read POTs at 7.5Khz (rather than 60Hz) (similar for PAL). KB is 15Khz. PIA is speed of processor (LDA 54016-- 4 cycles).

Currently, on A5200 they use the select lines of CONSOL to select which port gets use of the keypad-- only one keypad can be active at one time as they share the same POKEY for decoding the keys. Sel3 is used for internal speaker on A400/A800, Sel0..2 for START/SELECT/OPTION. Another advantage of SEL0..3 is that are I/O not just input just like PIA digital lines. Trigger lines are just input.

OK, but the question is how to best distribute the available I/O functionality among the controllers. If you wanted 4 controller ports, all equally compatible (in terms of the controllers used) and read similarly, then you couldn't have GTIA reading all the joystick inputs. In this case, without PIA, you're limited to 1 trigger input per controller, 1 select for key enabling, and the rest read as pots or keys in various possible configurations. So, 1 digital line on each port read by GTIA, the other 4 by POKEY, and the pots by POKEY, of course. --With what you provided, it seems like the fast pot scanning might be faster than keys if the pots are polled simultaneously. (rather than the keys which must be read concurrently -so 2x as fast for pot scanning when all ports are used, no speed advantage for 2 controllers though, and a disadvantage for pot scanning with a single controller -for fast scanning mode)

If, however, you limited the number of ports, or the purpose of some ports, things would change. So, with only 2 controller ports you could have all 8 GTIA digital lines dedicated to the joysticks of the 2 controllers, then read the buttons through the POT lines (if you added multiple resistors you could have more buttons as well, so possibly 2 fire buttons plus a few function buttons -start/select/pause/etc), and you wouldn't have the select lines for key enabling. You could have additional analog ports, and 1 port capable of using keys.

Given that they did end up switching to 2-ports later on (on both the 5200 and computers), with very few games allowing >4 simultaneous players using joysticks (or >4 simultaneous paddles), that doesn't seem to be too big of a deal. Also, I don't think any games required multiple keypads and joysticks to be used simultaneously (correct me if I'm wrong). A few games might have required 2 keypads, but I think any that required a joystick with a keypad were single player only. So, one could have 2 controller ports using DE-9 connectors with the standard pinout, with the directional lines connected to GTIA's digital inputs but you wouldn't be able to use the standard 8-bit joysticks then, as you're still missing the 5th digital input which can't be read as a key due to the select lines being used already. (you could easily use proprietary joystick using the analog lines for buttons, but the joysticks and keypads for the VCS wouldn't be compatible -paddles would be fully compatible though)
POEKEY's key inputs could either be run to the expansion port or to a dedicated keypad/keyboard port, as above. (12-key keypads should work fine with the normal controller ports though, just not the ones of the VCS due to the lack of the 5th digital line)

Any other options for 2 controller ports that allow full compatibility with the VCS/8-bit joysticks would again require use of the select line for enabling keys, so again, limiting things. Or could the key inputs could be distributed without select lines? I assume there is an 8-bit path (8 key input lines) for the key inputs (255 possible KEYs), so could one split that to 2x 4-bit paths? (or further divided) In the case of 4 GTIA inputs for 2 controllers, you only 1 one more available pin, so a 1-bit path per controller would be the only possible. (the trigger button line in the VCS pinout) If this is indeed possible, then you could easily have a 2-port system with full peripheral compatibility with the VCS as well as additional enhanced controllers, just as the VCS has the potential for, and 7800 did expand upon to some extent. (using pot lines for additional buttons, unused combinations on the digital lines for start/pause/etc, analog joystick, keypads, etc)

It's a waste lines to select keyboard for each port. Best to use two ports and no select lines. First port uses 4 trigger lines for directions and second port uses 4 select lines for directions. Buttons can be fast pot scan based and/or keyboard based. I.e., if you use keys on one, use pots on the other port. So no select lines are needed. I don't even see much use for keyboard on 2nd, 3rd and 4th ports anyway. Normally, keyboard is used on port #1 only anyways.

[kool kitty89]

atariksi, on Tue Mar 16, 2010 4:39 PM, said:

It's a waste lines to select keyboard for each port. Best to use two ports and no select lines. First port uses 4 trigger lines for directions and second port uses 4 select lines for directions. Buttons can be fast pot scan based and/or keyboard based. I.e., if you use keys on one, use pots on the other port. So no select lines are needed. I don't even see much use for keyboard on 2nd, 3rd and 4th ports anyway. Normally, keyboard is used on port #1 only anyways.

OK, but what I was saying is that if you wanted to have ports that were pin compatible with the VCS/8-bit/7800's then you'd need to have 5 digital lines per port as well as the 2 pot lines.

Now, looked up the pinout and I was mistaken in my assumption of POKEY having 8 key input lines, rather there ate 6 lines, still more than necessary for what I was asking. Again, would it work to have those lines divided between the controller ports? In the specific case of 2 ports using 4 GTIA lines each, there could only be one key input line used per controller anyway. So the pinout would be identical to the VCS and 8-bit, but instead of RIOT/PIA reading the directional lines, those go to GTIA, and instead of TIA/CTIA/GTIA reading the trigger lines, that falls to POKEY. (and additional buttons would be done with the analog lines plus additional combinations fromt he digital lines for keys and non-action buttons)

With compatible ports and the VCS keypads, you still be reading the keys with the analog and digital lines combined, so a bit sloppy. (otherwise you could add a new, proprietary keypad)


If you don't care about using compatible joystick ports, than that opens up other possibilities anyway, like sticking with the DA-15 ports the 5200 used, with added signals on them, etc.

[atariksi]

kool kitty89, on Tue Mar 16, 2010 6:27 PM, said:

atariksi, on Tue Mar 16, 2010 4:39 PM, said:

It's a waste lines to select keyboard for each port. Best to use two ports and no select lines. First port uses 4 trigger lines for directions and second port uses 4 select lines for directions. Buttons can be fast pot scan based and/or keyboard based. I.e., if you use keys on one, use pots on the other port. So no select lines are needed. I don't even see much use for keyboard on 2nd, 3rd and 4th ports anyway. Normally, keyboard is used on port #1 only anyways.

OK, but what I was saying is that if you wanted to have ports that were pin compatible with the VCS/8-bit/7800's then you'd need to have 5 digital lines per port as well as the 2 pot lines.

Now, looked up the pinout and I was mistaken in my assumption of POKEY having 8 key input lines, rather there ate 6 lines, still more than necessary for what I was asking. Again, would it work to have those lines divided between the controller ports? In the specific case of 2 ports using 4 GTIA lines each, there could only be one key input line used per controller anyway. So the pinout would be identical to the VCS and 8-bit, but instead of RIOT/PIA reading the directional lines, those go to GTIA, and instead of TIA/CTIA/GTIA reading the trigger lines, that falls to POKEY. (and additional buttons would be done with the analog lines plus additional combinations fromt he digital lines for keys and non-action buttons)

With compatible ports and the VCS keypads, you still be reading the keys with the analog and digital lines combined, so a bit sloppy. (otherwise you could add a new, proprietary keypad)


If you don't care about using compatible joystick ports, than that opens up other possibilities anyway, like sticking with the DA-15 ports the 5200 used, with added signals on them, etc.

Well, we do want DB9 compatible joysticks not some new arbitrary pinouts like Odyssey2 machines did. Odyssey2 used DB9 connectors but their pinouts were different from Atari joysticks. So in this case, we have 8 pot lines and we can assign 3 per DB9 port; i.e., have trigger also as a pot input so you have 3 paddles per DB9 port. Keyboard input can be a separate DB15 port like they have on XEGS machines. So a perfectly compatible DB9 pair of ports can be had on A5200 without PIA if they had thought about it. And nothing sloppy about the keyboard as it works great on XEGS for the few games that need it. Two pot lines are left still for START/PAUSE so that keyboard can be avoided even further.

[kool kitty89]

atariksi, on Wed Mar 17, 2010 6:44 PM, said:

Well, we do want DB9 compatible joysticks not some new arbitrary pinouts like Odyssey2 machines did. Odyssey2 used DB9 connectors but their pinouts were different from Atari joysticks. So in this case, we have 8 pot lines and we can assign 3 per DB9 port; i.e., have trigger also as a pot input so you have 3 paddles per DB9 port. Keyboard input can be a separate DB15 port like they have on XEGS machines. So a perfectly compatible DB9 pair of ports can be had on A5200 without PIA if they had thought about it. And nothing sloppy about the keyboard as it works great on XEGS for the few games that need it. Two pot lines are left still for START/PAUSE so that keyboard can be avoided even further.

Would you have 5V pulled through a resistor on the 3rd pot line internal to the machine in that case? (such that the joysticks would be interchangeable in the same way the VCS, 8-bit, and 7800 ones were)

Also, is it not possible to split up the POKEY key input lines as I suggested? (ie not use all 6 together)

Edited by kool kitty89, Thu Mar 18, 2010 1:02 AM.

[atariksi]

kool kitty89, on Thu Mar 18, 2010 12:57 AM, said:

atariksi, on Wed Mar 17, 2010 6:44 PM, said:

Well, we do want DB9 compatible joysticks not some new arbitrary pinouts like Odyssey2 machines did. Odyssey2 used DB9 connectors but their pinouts were different from Atari joysticks. So in this case, we have 8 pot lines and we can assign 3 per DB9 port; i.e., have trigger also as a pot input so you have 3 paddles per DB9 port. Keyboard input can be a separate DB15 port like they have on XEGS machines. So a perfectly compatible DB9 pair of ports can be had on A5200 without PIA if they had thought about it. And nothing sloppy about the keyboard as it works great on XEGS for the few games that need it. Two pot lines are left still for START/PAUSE so that keyboard can be avoided even further.

Would you have 5V pulled through a resistor on the 3rd pot line internal to the machine in that case? (such that the joysticks would be interchangeable in the same way the VCS, 8-bit, and 7800 ones were)

Also, is it not possible to split up the POKEY key input lines as I suggested? (ie not use all 6 together)

Yeah, you would need a +5V pull-up resistor on the pot line used for trigger to make it compatible with standard DB9 joysticks. Also, the SEL0..SEL3 lines on GTIA can be used for output as well so a keyboard controller/touchpad can also be hooked up just as you would on A800/A2600.

POKEY keyboard has a strobe line along with the 6-inputs so it won't be a straightforward method to use some lines as buttons.

[ledzep]

CV Gus, on Fri Mar 12, 2010 2:24 PM, said:

In a way, the title here is a bit misleading: it should actually read "Could the 5200 have been saved from Atari?"

Many people over the years, because it's been 26 years (oh, man...), are under the impression that the 5200 was ruined by the marketer-inspired crash of 1984. Maybe not here, for obvious reasons, but elsewhere, well...

The one thing that must be kept in mind at all times is that no outside factor doomed the 5200.

Atari itself pulled the plug on the 5200 months before the crash of 1984. Crash or no crash, Tramiel or no Tramiel, NES or no NES, it was over for the 5200. Atari dumped it for their 7800 project. The crash simply stalled that part, nothing more.

While there is no doubt that the CV played a role in that decision- it may have been a combination of CV and 2600.

The 2600 was the absolute #1 console for years. It received massive support from all directions, and was the undisputed king of the home consoles. Period. Nothing even really approached it.

So Atari had perhaps gotten used to being #1?

I kind of agree with that viewpoint - it was more Atari's fault than anything else.

People complain about the controllers. But that seems to be more hindsight than anything. Consider how when you encountered a new arcade game you just dealt with the controller layout before you and started playing (worse if you were left-handed like me). Same thing with the 5200 joysticks. Back then I owned a 2600, my friend had the 5200 and an Intellivision. We hated the Intellivision controllers, specifically that idiotic disc. Using them would pull the skin back from under our thumbnails because we pressed down on it so hard trying not to die. Not fun when eating Fritos (salt). Yet we played that thing too because of its superior baseball game and also games like Sea Battle and Utopia. And it's been pointed out that the Vectrex also had an analog controller and that system had some fun, great games. The key there, I think, is that most of its best games were? Yes, arcade ports. Very important back then when arcades were everywhere. And they found it easy to abandon the stick for buttons on many games.

People complain about the size of the 5200 console. The size? Who cares how big it is? You don't have to hold it while you're playing, you don't have to mount it on the TV screen, either. Seriously, complete this sentence - "If the Atari 5200 were half the size it would be a more fun game system because..." I can't think of any way that the console's size affects game play, the amount of games available for it, nothing.

I'm not sure how a lack of a lockout for 3rd party games matters. There's much talk about how a 5200 that was essentially a reskinned Atari 400 would have been better. Well, there would be no stopping 3rd party coders from writing games for that, would there? Either in Assembly or Basic, anyone could make a game and put it on a floppy disc or tape cassette and sell it, right? Or just print it out in an issue of Analog. I would think that open season on a console's cartridge port would be better in the sense that it would become the industry standard (sort of what the 2600 accomplished). All consoles had their garbage games so no issue there. Hardware/cart licensing = free money for Atari, that would be the bigger consideration, I can see how that would have mattered in terms of control and profits.

I agree that the pack-in game was a poor choice. But I think that's part of the solution to this question. Because Atari was neck deep in 2600 support that took away from concentrating on new 5200 games. If they'd done the smart thing and slowly but surely moved past the 2600 a year or two earlier then the 5200 would have had a bigger/better games selection waiting for it when it debuted. I mean wasn't Asteroids supposedly numbered 5201? As in first? Certainly that would been a better way to introduce the 5200 to the world. And I'm a fan of Super Breakout but still. I think also the 2600 multiple game variations approach should have been continued with the 5200 catalog. But I'm also no fan of "cute" games. I hate Pac-Man. Hate it. Same goes for Donkey Kong and Q-Bert and the rest. About the only cute games I like playing are Centipede (coded by a woman with girls in mind, yes, but you are killing everything, that's my kind of girl, pal) and Dig Dug. I'm sure I'm in the minority there but I grew up with arcade games that required that I waste everything that moved. Tempest, my friend, Asteroids. Missile Command, not Fruit Command.

Really what needed to happen was two things. First, stop supporting the 2600 like it was still new when it was actually old and primitive. Make new games for the 5200 first, show that it was the new #1. The 2600 had had its moment in the sun, for a long time. But the Colecovision and Intellision showed where the future was headed. Second, prevent the Big Crash. Part of that would have been abandoning the 2600 so no one would care to create a glut of terrible games made for it but, specifically, find the coder for the 2600 version of Pac-Man and put him in a coma before he started that monstrosity. That was the beginning of Atari's profits panic, the assumed millions upon millions of dollars that never materialized because of how horrible Pac-Man was. And the E.T. coder, 12-foot-hole time. But without the Crash and with the 5200 having full or major portion support, that would have gotten it rolling for twice the amount of time, easy. Asteroids and Xevious and the rest would have had time to be finished. That also would have given the Vectrex more time to grow. Competition is good. Imagine if Atari could have worked out a licensing deal with them for releasing Atari's vector games? Free money! More interest in Atari's own arcade ports, too.

And release a digital joystick and a paddle controller. Come on, already. That's not major, just obvious.

[jetset]

ledzep, on Fri Mar 19, 2010 3:18 AM, said:

People complain about the controllers. But that seems to be more hindsight than anything. Consider how when you encountered a new arcade game you just dealt with the controller layout before you and started playing (worse if you were left-handed like me). Same thing with the 5200 joysticks.

I've said that many times. I knew of nobody who complained about the controllers back in the day (save the fire buttons wearing out), and of all the gaming mags that reviewed the 5200, there were few, if any bad words about the joysticks being bad because they did not self center. I'm so tired of hearing how bad the 5200 controllers are, and I have no doubt these people are going by what they read on the internet. Sure they're not perfect. Frogger, Gorf...there's several games that are just borderline unplayable. The vast majority are just fine if you're halfway decent at playing videogames.

[Ransom]

jetset, on Fri Mar 19, 2010 5:06 AM, said:

I've said that many times. I knew of nobody who complained about the controllers back in the day (save the fire buttons wearing out), and of all the gaming mags that reviewed the 5200, there were few, if any bad words about the joysticks being bad because they did not self center. I'm so tired of hearing how bad the 5200 controllers are, and I have no doubt these people are going by what they read on the internet. Sure they're not perfect. Frogger, Gorf...there's several games that are just borderline unplayable. The vast majority are just fine if you're halfway decent at playing videogames.

..and most of the unplayable games could have been better designed and would have been fine as well.

[Curt Vendel]

and I will continue to reinterate - when compared to the Colecovision controllers, the Atari 5200 controllers were superior in tactile feel, control and comfort.

The short knob-like handle of the CV joystick is painful to use for even moderate periods of gameplay, the fire buttons on every controller I've used are way to strong/stiff and lead to fatigue and well can all see now in retrospect, upon opening both controllers, how cheaply and poorly the CV controllers were made and how they don't stand up to the test of time nearly as well as the 5200 controllers.

I would say the sore spot specifically on the 5200 controllers are the fire buttons, they should've been plastic, not rubber and they should've been like the 7800 controllers, an A button on one side and a B button on the other side.



Curt

Ransom, on Fri Mar 19, 2010 7:19 AM, said:

jetset, on Fri Mar 19, 2010 5:06 AM, said:

I've said that many times. I knew of nobody who complained about the controllers back in the day (save the fire buttons wearing out), and of all the gaming mags that reviewed the 5200, there were few, if any bad words about the joysticks being bad because they did not self center. I'm so tired of hearing how bad the 5200 controllers are, and I have no doubt these people are going by what they read on the internet. Sure they're not perfect. Frogger, Gorf...there's several games that are just borderline unplayable. The vast majority are just fine if you're halfway decent at playing videogames.

..and most of the unplayable games could have been better designed and would have been fine as well.

[kool kitty89]

Curt Vendel, on Fri Mar 19, 2010 10:03 AM, said:

and I will continue to reinterate - when compared to the Colecovision controllers, the Atari 5200 controllers were superior in tactile feel, control and comfort.

The short knob-like handle of the CV joystick is painful to use for even moderate periods of gameplay, the fire buttons on every controller I've used are way to strong/stiff and lead to fatigue and well can all see now in retrospect, upon opening both controllers, how cheaply and poorly the CV controllers were made and how they don't stand up to the test of time nearly as well as the 5200 controllers.

What about using the knob with your thumb rather than a "handel"? The problem with that, of course, would be accessing the buttons simultaneously. (same with the 5200 stick used with the thumb as well though) With the 5200's short stick, it's best in both respects though, comfortable as a small joystick or thumbstick (albeit a bit long throw), so I guess the overall point still applies.
Had the colecovision used a normal joystick stalk more like the 5200 or perhaps coleco's own gemini controllers, it might have been a different story. (the recessed fire buttons are also a disadvantage, at least they were reliable though)

Curt Vendel, on Fri Mar 19, 2010 10:03 AM, said:

if any bad words about the joysticks being bad because they did not self center. I'm so tired of hearing how bad the 5200 controllers are, and I have no doubt these people are going by what they read on the internet.

For games designed for 4 or 8 directions, it's not so much the lack of self centering, but the lack of self centering combined with the long throw of the joystick, such that it's hard to make precise movements. The same thing happens with analog joysticks on other platforms, things like the Vectrex and most modern consoles don't suffer from it as much though due to the small, short throw (and self centering) analog thumbsticks used. But still, a real digital joystick (or especially d-pad -short of an arcade quality microswitch joystick) is best for 4/8 direction games and you will indeed find a lot of "hardcore" gamers complaing about such on modern consoles. (that the d-pads are poor if present, and analog sticks a bit sloppy compared to a proper d-pad)

The other thing is that a large portion of games for the system weren't originally designed for analog control, and many were direct ports from the 8-bits, so the analog joysticks would rarely be at an advantage. In a few cases, they were quite useful though, some trackball based games used it fairly well, as did some paddle based ones (albeit weaker than real paddles), but the few games to really benefit would be ones meant for analog joysticks (or similar) in the original format, like the 1983 Star Wars arcade game. (though, if it had been a simple port from the 8-bit, the analog controllers wouldn't have been taken advantage of either)

Edited by kool kitty89, Fri Mar 19, 2010 8:37 PM.

[kool kitty89]

ledzep, on Fri Mar 19, 2010 3:18 AM, said:

And it's been pointed out that the Vectrex also had an analog controller and that system had some fun, great games. The key there, I think, is that most of its best games were? Yes, arcade ports. Very important back then when arcades were everywhere. And they found it easy to abandon the stick for buttons on many games.

The vectrex has a very different controller, it's almost a gamepad, though a bit big. (more like a scaled down arcade control board) But the button layout was simple, and reliable, and the joystick was small, short throw and self-centering.
Even with conrollers with broken centering springs, many games are still playable, but there are a few which really need a properly centering controller for precise movements.

Quote

People complain about the size of the 5200 console. The size? Who cares how big it is? You don't have to hold it while you're playing, you don't have to mount it on the TV screen, either. Seriously, complete this sentence - "If the Atari 5200 were half the size it would be a more fun game system because..." I can't think of any way that the console's size affects game play, the amount of games available for it, nothing.

You can't play it if you have no room to set it up in your entertainment system, on top of, or in front of your TV. It is a significant issue, and unlike a bulky VCR, you need access to the top for cartridges, so you cant put it on a shelf with limited clearance. On top of that, it make it less cost effective (materials cost, packaging, warehouse space, shelf space for retailers, etc).

Quote

I'm not sure how a lack of a lockout for 3rd party games matters. There's much talk about how a 5200 that was essentially a reskinned Atari 400 would have been better. Well, there would be no stopping 3rd party coders from writing games for that, would there?

Exactly, I always try to maintain both of those points. However, since the 5200 didn't implement any lockout, it doesn't much matter in that case. But from a practical standpoint, the most optimized quick conversion for a console based on the 8-bits would have been to keep the hardware the same, memory map the same (except perhaps more flat mapped cart space -but combining both carts would give 16 kB at least), remove SIO and change the cartrige port. (no worring about 2600 compatibility either, except perhaps facilitating an add-on for such -like a front or side mounted expansion port) Plus add a lockout mechanism, probably a checksum set-up like with the 7800. (not as finicky or expensive as Nintendo's route)
Since Atari kept saturating the 2600, the lockout still didn't matter that much either. (witht he 2600 being completely open to 3rd parties) They'd have need to start shifting resources and slowing things down with he 2600 more.

Quote

Tempest, my friend, Asteroids. Missile Command, not Fruit Command.

What would have really been nice is a Missile Command+ release, like the XL eventually got, with the proper arcade-accurate 3 missile launchers.

Quote

Really what needed to happen was two things. First, stop supporting the 2600 like it was still new when it was actually old and primitive. Make new games for the 5200 first, show that it was the new #1. The 2600 had had its moment in the sun, for a long time. But the Colecovision and Intellision showed where the future was headed. Second, prevent the Big Crash. Part of that would have been abandoning the 2600 so no one would care to create a glut of terrible games made for it but, specifically, find the coder for the 2600 version of Pac-Man and put him in a coma before he started that monstrosity.

Yes, the crash is tied mostly to Atari/Warner management problems, overinflation of the market, and oversaturation, clining to the 2600 too long is part of all that too. The thing was PAC Man was a bad screw-up, it's a double whammy with the overproduction of carts and poor quality though. (if it was really good quality, perhaps they'd have sold the lot) But things like ET and PAC man are just symptoms of the internal problems Atari/Warner had. Bunches of poor quality 3rd party titles probably were one of the lesser problems though. And, of course, Commodore's price war in the home computer market really acted as a catylist and heaved everything over the edge.
It might not have been that bad for Atari had they managed to compete with the 8-bit line during the crash (with computer games remaining popular, if not increasing), but they screwed things up with that already. In fact, that's part of the argument for the directly compatible 8-bit game system; it would have beein in a good position in spite of the crash, as it was, in fact, a home computer.
Coleco might have had a chance at that too, but totally botched the Adam.

Quote

And release a digital joystick and a paddle controller. Come on, already. That's not major, just obvious.

It shouldn't have been tough for the standard pack-in controller to be "digital" (8-way) either, though any games specifically catering to the analog control (in a paddle like position sense, not in the speed sensitive sense) would be problematic for such a controller. (but no different than requiring special controllers for some VCS games, like the driving or keyboard controllers)