Things the Amiga didn't get right from Day 1

[chb]

Originally Posted by Thomas Richter

The problem did not appear with OCS. It appeared with ECS where registers were added behind the color registers. But then again, those registers could have been considered private, and it wasn't the application to poke them, but rather the operating system, so anything could have been added later. Actually, AAA was designed to be only ECS compatible, and any extended registers would have been elsewhere.

AFAIK the hackish way the additional color registers were added is necessitated by the limited RGA bus address space (it only supports up to 256 16-bit registers). I guess it would have been quite complex (read: costly) to change that.

[chiark]

The Amiga colour model was going to be Hue, Saturation and Luminance/Value (HSV or HSL depending on your preference) according to some interviews for Commodore: The Amiga Years. HAM would make much more sense if you hold two of HSV and modify just one, compared to RGB...

It was hacked to RGB and left in, and I'm very, very glad it was. Incidentally, I've been trying without much success to find out more about the HSL model proposed as the base in the amiga: it would still have used bitmaps, and indexed colour, however dff180 (and 1 ) would be storing HSV, not RGB... Maybe that makes sense of OS1.x colours ;-)

[roondar]

I read somewhere* that orange/white/blue/black is actually a very good colour combination (though you and I might disagree ). It supposedly had to do with two things, the first was that these colours are relatively good fits to composite video which allows for the image to be sharper than many other combinations. The second was that the colours are a good fit to our eyes as well, meaning they cause less eye strain than black letters on a white background.

The latter was (allegedly) also why so many of the old PC coding environments (and stuff like Word Perfect when displayed on a colour monitor) had a blue background.

So there may have been method to the madness there

*) It was rather long ago, so I can't recall exactly when & where

[Jope]

Originally Posted by chiark

Maybe that makes sense of OS1.x colours ;-)

The story about those is that they are legible on the cheapest NTSC TV set via composite/RF. :-)

[idrougge]

Originally Posted by Thomas Richter

But the workbench is multithreaded. You mean the "loading part"? Or the "copy part"? Program-wise, this is trivial enough, but it would have meant more ROM space, and that was already tight.

So it's not multithreaded where it actually counts.

What made cooperative multitasking so bad was exactly this; You could pretend you were multitasking until you actually needed to multitask.

[idrougge]

Originally Posted by sandruzzo

Remember that Amiga was born as Computer, not a pure consolle, so A Little ram for only 68k would have done a lot of difference. Think about saving registers into stack...

A memory layout like that (chip memory + a small (16~32 kB) amount of high speed CPU-only memory) only works well in a console setting, though, not on a computer.

In a single-task game setting, allocating the small block of fast memory for fast operations works fine.

In a multitasking OS, you immediately reach contention of the limited fast memory and every task started afterwards will be relegated to slower memory. It would also mean that programs would be written in a way that would be too concerned with memory types instead of treating all memory as equal.

[idrougge]

Originally Posted by McTrinsic

The ST had the Software you mention due to the MIDI interface that was available as standard.

The ST also has an advantage in its cheap monochrome high resolution screen with a steady 72 Hz refresh rate. Perfect for a sequencer and without any annoying 15 kHz squeal of a (more expensive) 1084.

[roondar]

Originally Posted by idrougge

So it's not multithreaded where it actually counts.

What made cooperative multitasking so bad was exactly this; You could pretend you were multitasking until you actually needed to multitask.

The Amiga OS is perfectly multitasking/multithreading as is. I can load a program while all my other programs still run and respond. The WB itself might not be, but that is ultimately pretty much irrelevant for all but starting new programs.

There is no "pretend" multitasking on the Amiga.

[idrougge]

Originally Posted by drHirudo

This happened in the Atari land with the ST (The Jackintosh). Both Macs and Atari's used mostly off the shelf parts - one of them succeeded, the other miserably failed.

I wonder where this misconception first came from — I believed it myself for many years until I actually looked inside an ST.

The Atari ST has a custom chipset consisting of GLUE, MMU, Shifter and an optional Blitter.

What is not custom in the ST is the sound chip (off the shelf Yamaha YM2149) and chips which Commodore would have been better off not making themselves, such as the Motorola 68901 multi function peripheral

Notably, the X68000, which has a "custom" reputation, also used the 68901 and an off-the-shelf Yamaha sound chip.

So the ST certainly had a custom chipset — there were no off-shelf parts you could use to build an ST in 1985 — it just was not as advanced as the Amiga's.

[idrougge]

Originally Posted by roondar

The Amiga OS is perfectly multitasking/multithreading as is. I can load a program while all my other programs still run and respond. The WB itself might not be, but that is ultimately pretty much irrelevant for all but starting new programs.

There is no "pretend" multitasking on the Amiga.

I never wrote that the Amiga used cooperative multitasking, did I?

[roondar]

Originally Posted by idrougge

I never wrote that the Amiga used cooperative multitasking, did I?

I do apologize, but that is kind of how I read it. You seemed to imply that the Amiga lacked multithreading where it "actually counts" and seemingly compared that situation to co-operative multitasking.

If you meant something different, forget what I wrote

[idrougge]

The Workbench lacks multi-threading where it counts, just as the classic Macintosh lacks multitasking where it counts.

[Old_Bob]

Originally Posted by Thomas Richter

CBM was, always, a cheap company. Including a joystick would have made the system more expensive. It was rather unclear to them, however, how to sell the machine. The original A1000 package was more that of a business machine. Mouse and keyboard included, joystick not included.

Is there any reason why they didn't give us a three button mouse? Given that every Amiga ever sold came with a mouse in the box, I can't imagine that tiny change increasing the cost much. It would surely add no more than a cent, or two, to price of a tank mouse. As well as being potentially incredibly useful.

Originally Posted by Thomas Richter

The problem did not appear with OCS. It appeared with ECS where registers were added behind the color registers. But then again, those registers could have been considered private, and it wasn't the application to poke them, but rather the operating system, so anything could have been added later. Actually, AAA was designed to be only ECS compatible, and any extended registers would have been elsewhere.

Is there any technical reason preventing the addition of these new registers a bit further up in the memory map, while still leaving room for, say... another 32 colour registers to go with our six bitplanes? Or room for all 256, to be utilised sometime in the future?

Originally Posted by Thomas Richter

As said, this would have complicated the design as then Agnus has also to allocate DMA slots for the faster CPU, hence would have required a new Agnus design. However, CBM never "redesigned" much of the system. The development costs for such a system would be much higher than that of the "budget" A500 system which only collected some gates in advanced custom chips, without extending their capabilities.

Remind me how many revisions of Agnus there actually are?!

If they're already committed to a redesign for the A500 and A2000 models then this is a great time to update it. 68HC000 parts would surely be cheap enough by 1987/88, perhaps not quite down to jelly bean levels, but enough not to affect the price of a £500 computer? Although, the price of necessarily faster ICs for Chip RAM, admittedly would have. Adding Fast RAM, too, would be ideal, of course at additional expense. If this new Agnus chip could have a 14mhz blitter inside it, that would have been fantastic.

That's a machine that would leave absolutely no doubt, in anybody's mind, about which home computer/game system was top dog out of all the options available at the time.

None of this sounds to me to be technically much of a leap over the original Amiga. Merely updating an existing design to operate a bit faster, as opposed to creating a whole new beast from scratch. Surely all it would take is the will to do so among the management at C=, along with money and time?

B

[Thomas Richter]

Originally Posted by idrougge

The Workbench lacks multi-threading where it counts, just as the classic Macintosh lacks multitasking where it counts.

Sorry, i fail to understand this. Where does *it* count? Loading? Why? There was only one disk drive, thus you cannot make it "load faster" by two tasks busy on the drive. This creates just a lot of "disk grunting".

Things did change when harddrives became available.

I beileve it counts a lot that you can start another program while another one is still running. The Mac Finder could not do that. You need to quit the running program to return to the Finder. MultiFinder allowed that, but only later on (but in a kludgy way).

[roondar]

Originally Posted by idrougge

The Workbench lacks multi-threading where it counts, just as the classic Macintosh lacks multitasking where it counts.

So to clarify: you're not talking about the OS, just the Workbench (i.e. GUI)?

Because in that case I agree with you and I also kind of disagree. As in: yes, I agree the Workbench (but not OS) does not have multithreading. And I'd also agree that now (in 2020), that seems like a serious oversight.

But back in 1985?

Well, no... I kind of disagree that it misses out on this feature, because it didn't actually matter all that much (IMHO). WB was advanced enough when it first came out and still did better than competing options at the time for many tasks. Mainly thanks to offering multitasking and moving some of the basic actions into separate programs. This way, at least you could copy/format disks etc, while doing other stuff using WB.

There also was a usable workaround - if you really somehow needed to copy files (or something like that) while keeping WB available, the OS offered the CLI (which generally is a much better way of dealing with file management anyway), which would happily do the tasks while you went on using the WB.

But to be clear: you are correct that WB did not offer multithreading and it would be nice if it did. I just personally don't feel it was all that big a deal back then.

Originally Posted by Old_Bob

None of this sounds to me to be technically much of a leap over the original Amiga. Merely updating an existing design to operate a bit faster, as opposed to creating a whole new beast from scratch.

Commodore never made the Blitter/chipset faster, even when they did ECS and AGA. Atari actually had the same problem with it's old chipset (chucking most of it out when moving to the Falcon - the Blitter was still in there from the STE, but AFAIK didn't become faster but rather slower).

Now, I'm no expert on hardware by any stretch of the imagination, but to me that indicates that it may be harder than we think to increase these clockrates (even the suggested AGA+ still had the chipset run at the same 3.5ish MHz).

It would've been great to see though!

[Thomas Richter]

Originally Posted by Old_Bob

Is there any reason why they didn't give us a three button mouse?

Technical? No, the system could handle it hardware wise. It was probably an interface design issue. I rarely need the middle mouse button on systems that offer one. The overall idea - left button for selection and clicking, right for the menu - is quite consistent. What is the middle mouse button for? Text pasting is the only function I use it for today.

Originally Posted by Old_Bob

Given that every Amiga ever sold came with a mouse in the box, I can't imagine that tiny change increasing the cost much. It would surely add no more than a cent, or two, to price of a tank mouse. As well as being potentially incredibly useful.

Useful for *what* back then?

Originally Posted by Old_Bob

Is there any technical reason preventing the addition of these new registers a bit further up in the memory map, while still leaving room for, say... another 32 colour registers to go with our six bitplanes? Or room for all 256, to be utilised sometime in the future?

There is room for them in OCS, if you want to say so. Just that this room was then allocated for other purposes.

Originally Posted by Old_Bob

Remind me how many revisions of Agnus there actually are?!

Feature wise, the only substantial change happened with AGA. Fat agnus just integrated glue logic without changing anything - ok, one extra bit for memory refresh. Second generation added one additional bit for each DMA engine and moved some hard decisions about the video timing into registers. Third generation brought a lot of changes.


Now think of what you ask for: This would require a new DMA slot allocation scheme and a new bus design. Approximately the same change made on the bus and the DMA engine as made for AGA. That is quite a task.

Originally Posted by Old_Bob

None of this sounds to me to be technically much of a leap over the original Amiga.

But it's quite a development cost attached to it. For a computer system the company just bought, with unclear return of investment, I understand very well why that change was not made.

[Old_Bob]

Originally Posted by Thomas Richter

Useful for *what* back then?

Useful for whatever some programmer, out there in the world, decided when writing some piece of software. Chucking grenades in Cannon Fodder, without having to hold down both buttons seems like a good idea.

Originally Posted by Thomas Richter

This would require a new DMA slot allocation scheme and a new bus design. Approximately the same change made on the bus and the DMA engine as made for AGA. That is quite a task.

Of course. But that's exactly what all the hardware designers and engineers working at C= back then got paid to do. What else are they going to do all day? Which segues nicely to...

Originally Posted by Thomas Richter

But it's quite a development cost attached to it. For a computer system the company just bought, with unclear return of investment, I understand very well why that change was not made.

As ever, it comes down to the management at C= not being willing to make the investment. In my opinion, this was an error of judgement. Especially when we consider the things they did end up wasting tons of cash developing.

B

[idrougge]

Originally Posted by Thomas Richter

Sorry, i fail to understand this. Where does *it* count? Loading? Why? There was only one disk drive, thus you cannot make it "load faster" by two tasks busy on the drive. This creates just a lot of "disk grunting".

Things did change when harddrives became available.

I beileve it counts a lot that you can start another program while another one is still running. The Mac Finder could not do that. You need to quit the running program to return to the Finder. MultiFinder allowed that, but only later on (but in a kludgy way).

Sorry, I am not referring to when the respective system was introduced, but to their entire lifetime, including Mac System 7 & 8.

[idrougge]

Originally Posted by roondar

So to clarify: you're not talking about the OS, just the Workbench (i.e. GUI)?

Yes, just the Workbench. We all know AmigaOS has preemptive multitasking, as we were taught to repeat before we even knew what it meant.

Originally Posted by roondar

Because in that case I agree with you and I also kind of disagree. As in: yes, I agree the Workbench (but not OS) does not have multithreading. And I'd also agree that now (in 2020), that seems like a serious oversight.

But back in 1985?

Agreed, but note that this wasn't brought up by me, and that my original comment was in response to Thomas Richter's claim that it was in fact multithreaded — partially.

The fact that multitasking is claimed proudly on the packaging while the main interface does not multitask internally is also a bit embarrassing for Amiga users.

[idrougge]

Originally Posted by Old_Bob

Is there any reason why they didn't give us a three button mouse? Given that every Amiga ever sold came with a mouse in the box, I can't imagine that tiny change increasing the cost much. It would surely add no more than a cent, or two, to price of a tank mouse. As well as being potentially incredibly useful.

It would probably cost more than that. You need to add not only an extra button, but also an extra wire to the cable.

A standard two-button Amiga mouse has 8 wires, and mouse cables need to be both thin, flexible and durable, since you move it around all the time – much more so than a joystick cable. Sourcing such a cable with 9 wires that is both thin, flexible and durable can be quite difficult.