pjw wrote: Stella, now that is something different.
Yes it really is, and it did leave the drawing board and I have it running on an Atari ST, so it must exist. I imagine that Qurus, worthier than me, would also have copies.
I will suggest an alternative view on the SMS2/Stella scenario. Alternative as in alternative to pjw and RalpR's understanding as expressed earlier.
As far as I am concerned, the main, actually only significant, difference between SMS2 and SMSQE is the absence of a BASIC interpreter as a CLI. All the BASIC programs I write in SMSQE run perfectly on SMS2. The interpreter was introduced into SMS2 to turn it into SMSQE at the request of Jochen Mertz and Miracle systems because they both thought that the QL user would not be able to understand a system with a different (better in my opinion) CLI.
It is true to say that SMSQE has been enhanced with GD2 and other useful drivers but I cannot imagine it would be difficult to add them to SMS2.
The only time I use the SMSQE interpreter is to run a BOOT program. I find the equivalent SMS2 STARTUP file easier to understand. It does exactly the same thing and uses similar instructions. I have never found another use for the SBASIC interpreter.
Running an SBASIC program on my HD SMSQE setup always scrambles my screen. Executing a _bas program is slightly less useful than executing a _obj program, so I don't do it.
SMS2 was not built to be an alternative QDOS for QL users. It was designed to be an easy to use system that embodied TT's OS theories.
What I find strange is that JM, an SMS2 vendor, thought that the true paleoQtologists needed an enhanced OS, downgraded to look like a QL. I know that he ran an extensive suit of interpreted programs to run his business and wanted the benefits of SMS2 without having to compile his programs, but why sabotage an advanced OS just to satisfy his particular problem.
One can tell from some of TT's plaintive missives, posted elsewhere on this forum, that he was not exactly enamoured with the use of an interpreter as a user interface.
SBASIC is embedded in SMS2 (well it is on my version) via a QD-parser-Qlib IDE. It is this "SMS2 IDE" that I use in SMSQE, together with various other programs, to learn how to write SBASIC applications. If an SMS type OS was ever to escape the confines of the QL world, to be used as a modern hobby platform, the last thing it should look like on startup, is a system designed in the 1980s.
I think that an enhanced SMS2 (ie with GD2 etc. added) in something like a Qzero would be of great interest to the general computer hobbyist.
RalfR writes that SMS2 "needs quit a lot of additional toolkits" (to make it useful?). The PEROM SMS2 that I use (in an Atari PC emulator) contains everything worthwhile including THINGS and an expandable SBASIC, whereas SMSQE in its "naked" form, looks like a basic QL. To make SMSQE usable I, do indeed, need to add lots of stuff via a BOOT program. In my case it is a variant of an SMS2 STARTUP file.
Stella is his opus magnum. TT was, I think, convinced that QDOS and SMS2 proved beyond doubt that his OS theories had merit, but that they were both significantly flawed. At least twice he proposed building a new system, without flaws, from scratch.
One proposal was made approximately 6 months after SMS2 was finished. He suggested developing something (as in something completely different) with a select band of Qgurus. It came to nothing, partly I imagine, because there was nothing to run it on. Another proposal was to build a "multi processor Lego like system" with an OS to match but again using what processors.
Finally he had the opportunity, whilst developing SMS2-in-PEROM to create a brand new OS on the only off-the-shelf hardware he had at his disposal. That hardware was the Atari ST and, as far as I know, the ST is the only hardware where it still resides to date.
One of his problems may have been "on what do I run Stella". This is almost the exact opposite to the problem faced by HP and their billion dollar attempt at building something new titled "The Machine". At least TT had a novel (see
http://www.hpe) "Memory-Driven" OS.
Without the billions regularly used to develop more fashionable systems, TT was confined to using low cost 680xx based computers and we all know what happened to them.
The reason I keep "banging on" about FPGAs (the clue is in my user name) is that for the first time since the creation of Stella an OS can be developed without being dependent on conventional hardware. For those with the skills, as Peter Graf has demonstrated, creating hardware in FPGA is no more difficult than creating its companion OS.
If the Qurus are really looking for a challenge then Stella or SMS2, as a stepping stone, in FPGA is the way to escape the confines of history. SMSQE, although an eminently usable system, that I like a lot, was a sop to JM, and possibly others unwilling to move forwards.