SMS2 and STELLA

[Tinyfpga]

Sometimes in a community this is because those ideas are wrong!

Do you really think his ideas are erroneous? His stuff seems to work.

Which luckilly I can do unlike information on Stella

I must say that is very droll.

[XorA]

Sometimes in a community this is because those ideas are wrong!

Do you really think his ideas are erroneous? His stuff seems to work.

I cant really tell not having access to Stella or SMS2 details to actually look at. Some of the documentation he wrote seems to suggest he does not believe in IRQs and thinks polling is the best way to handle devices. Not sure that would be so efficient on the 1000+ IRQ devices I use regularly! But there is so little available to really tell!

Which luckilly I can do unlike information on Stella

I must say that is very droll.

It was kind of a hint to link to some actual information about stella, unfortunately the only think I keep turning up is a 2600 emulator, which is cool but not helpful!

[Tinyfpga]

It was kind of a hint to link to some actual information about stella, unfortunately the only think I keep turning up is a 2600 emulator, which is cool but not helpful!

I gathered that it was. Really it should be TT's job to do this. He is, after all, "the enlightened one". There are a number of possibilities that might explain his silence and I hope that one of them is not because Stella is faulty technology.
It seems fine to me but I know nothing and so can't really tell.
Peter Graf seems to like what ever he has seen of it, and as a hobby system it might be perfect. Have you read his
response to TT's post?
I assume a hobby system would not have 1000+ IRQs.

Maybe TT doesn't like the idea that it might end up being used for that purpose.

What does an IRQ do? Aren't there any IRQs in SMS?

[Derek_Stewart]

Hi,

The SMS2 CLI is an Interpreter that has limited functionality, I do not mind CLIss, the Linux BASH Shell is nice to use.

You did answer my point with regards to Stella: make the source available and we can see what you are going on about.

[XorA]

It was kind of a hint to link to some actual information about stella, unfortunately the only think I keep turning up is a 2600 emulator, which is cool but not helpful!

I gathered that it was. Really it should be TT's job to do this. He is, after all, "the enlightened one". There are a number of possibilities that might explain his silence and I hope that one of them is not because Stella is faulty technology.
It seems fine to me but I know nothing and so can't really tell.

Wait you are advocating for all of us to start coding for a system no-one including you has seen?

Peter Graf seems to like what ever he has seen of it, and as a hobby system it might be perfect. Have you read his
response to TT's post?
I assume a hobby system would not have 1000+ IRQs.

Depends if you count your PC or laptop as hobbyist!

Maybe TT doesn't like the idea that it might end up being used for that purpose.

What does an IRQ do? Aren't there any IRQs in SMS?

Interrupt ReQuest!

[tofro]

From the scarce sources I've found, it doesn't sound like Stella would be far from what SMSQ/E is today, minus old QDOS "ballast" (which isn't really that much) which must have been a nuisance to implement in a 100% compatible way (that one I get) to be able to run 99.9% of old QDOS software - Stella was an attempt to drop that. The concepts are more or less the same and I can't really see any breakthrough new inventions. Apparently, it was supposed to target real-time embedded applications.

With the polling I/O concept that's in all of TT's operating systems, it would have had a hard time competing with contemporary hard-real-time OSs (like pSOS or vxWorks) in that segment, however. QDOS (or Stella, for that matter) is a relatively soft real-time OS (because of the 50Hz I/O polling instead of being mainly interrupt-driven). That takes away a lot of complexity and overhead in the I/O subsystem and the scheduler, but does put the system quite a bit behind in terms of responsiveness (which is fine and dandy for a desktop OS, but crucial for a real-time embedded system)

[Tinyfpga]

Wait you are advocating for all of us to start coding for a system no-one including you has seen?

You misunderstand me. When I say, I know nothing, I mean I know nothing about operating systems, not that I know nothing about the existence of Stella. I have said on a number of occasions that it exists in code and documented form and that I have an Atari ST version that works. It is currently stored on a floppy disc and is written in 68000 assembly.

But:- it is not for me to make the code available. The reason for starting this topic is to try and find out why
no one from the QL world has ever expressed much interest in TT's proposals. (For one of them, read "QL Today").

SMS2 was rejected by QLers, and clearly still is, and SMS2 was the second expression of TT's theories.
The third, that is Stella, probably only exists as the working system I have in Atari ST format, and no one today is going to work on system that runs on an Atari ST.

Recently Peter Graf's work on FPGA's that has reawakened my interest in Stella because, as he has suggested on a number of occasions, development of Stella in 68000 assembly and in FPGA is a realistic proposition. This has only become possible recently.

As far as I know he received no answer to his post of March 2017 where he suggested, to Tony Tebby, that FPGA is the way to go.

What to do about this situation? that is what I keep asking myself.

I am certainly not advocating that anyone should do anything, and as you say, on the matter of Stella they can't anyway. I am merely searching for considered opinions in a forum for TT operating system users.

[Tinyfpga]

Postby tofro » Mon Feb 14, 2022 5:56 pm
From the scarce sources I've found

This is the kind of considered opinion that I was looking for.
In one of the documents I posted on this forum (I think), TT claims a superior performance over vxWorks extrapolated, if you can believe it, from timings on Stella in Atari. I may be wrong about this so I will dig up the document.

[Dave]

This is a conversation about what could have been and won't ever be. Besides being an academic exercise, it's also a fruitless one in terms of OS product we could use.

That said, the thread is also highly informative and entertaining.

[Tinyfpga]

Postby Dave » Mon Feb 14, 2022 7:05 pm
This is a conversation about what could have been and won't ever be. Besides being an academic exercise, it's also a fruitless one in terms of OS product we could use.

That said, the thread is also highly informative and entertaining.

I feel you are being unduly pessimistic. After all it is only recently that FPGAs have become powerful and cheap enough
to contain a 68000 "system-on-chip" for hobby use.

Stella in FPGA might take some time, meanwhile we have SMS2 and SMSQE in FPGA to entertain us. (well two of us have SMS2 in FPGA).

Success will not come by giving up before starting.