Spectrum Computing

Thought this story would be of interest here

https://hackaday.com/2019/03/04/the-pri ... -cupboard/

Not sure about " the decision to ship with an unfinished ROM version meaning that later peripherals had to carry shadow ROMs with updated routines", maybe they mean the Interface 1 to implement Microdrive commands...

The center for computing History has announced the prototype rom is available to download at its site:

http://www.computinghistory.org.uk/det/ ... Prototype/

https://youtu.be/gFg09u6MbMg

Question, what software is ever completely finished?

Sinclair faced a dilemma. Hold up production and release of the computer until the microdrives were finished and working, with no guarantee that all the required code could be included in the ROM. Or release the machine with the ROM that currently worked and provided the all the required code for the basic unexpanded machine.

A no brainier really...

Were there still bugs? Yes. But we don’t know how many were known about by Sinclair at the time. And with the release date approaching fast...

Also, the ROM that was included fully supported the ZX Printer and no ROM code over and above the existing keyboard routines was needed for the ZX Interface 2. It was only the ZX Interface 1 that needed an overlay ROM.

Mark

Just been playing around with the prototype ROM in an emulator, and I've already observed some interesting differences. The MOVE command is printed as "MOVE #", and the CLEAR command seems to be broken. I'll have to try it out some more.

EDIT: Just noticed that for some reason, the CLEAR bug seems to happen in Spectaculator and Fuse, but not Spin.

Ah, that's wonderful news!


Incidentally, I've discovered that if you type "prototype" into the search box at the museum's website, you get reems of really interesting results, including a prototype QL wafer-scale integration device!

http://www.computinghistory.org.uk/det/ ... Expansion/

Looks like there are some minor differences from the final ROM. For example, the prototype ROM has additional channels "N" and "T", where "T" is used for communication over port $fd. Also, it features code for both FORMAT and MOVE, although CAT and ERASE still do nothing. There's even an additional error message (S Device unformatted) that occurs if you try to use channel "N".

Some interesting analysis here

https://spectrumforeveryone.com/technic ... trums-rom/

Interesting article. It's not true however that the rest of the ROM code is identical; I think the author missed some changes. For example, the checks for parentheses at PRINT_4 and IN_ITEM_1 are entirely missing in the prototype ROM. I don't know if these checks affect anything or if they were removed for being redundant.

There may be other differences but I don't recall them now.

I was going to start a new topic about this until I found this one and its relevant here...

Source > https://www.computinghistory.org.uk/det ... Prototype/

Relevant sentence highlighted in blue

... The original plan was for the machine to ship, and an upgraded ROM to be made available later, a plan rendered impossible by the sheer number of machines sold.

To counter this, peripherals had to include a shadow ROM to work alongside the host machine.

The result of all this unsettled period was that Sinclair launched the Spectrum with an unfinished ROM, Nine Tiles continued to work on it until 3 months after launch, but by then too many machines had shipped, so the new code and this prototype were no longer needed...

Question is, just exactly what were they still doing (it would of been good) ? Was it just pushing in the IF1 code perhaps ?

I do wonder if there were some features planned that have not been mentioned too and it was not just (speculation) IF1 or suchlike code being jammed in.

Andy, you may want to read through the topic again

But to sum up, it looks like they were trying to implement some of the commands and functions needed for the microdrive, network and RS232 serial systems.

Mark

1024MAK wrote: ↑Fri Jul 09, 2021 4:38 pm Andy, you may want to read through the topic again

I must be totally honest. I started to compose a new topic and then realised one of my Google searches for the "three months" (as I could not recall where I'd read it at the time) produced a thread/topic here so I then just added to it, without reading it properly

1024MAK wrote: ↑Fri Jul 09, 2021 4:38 pm But to sum up, it looks like they were trying to implement some of the commands and functions needed for the microdrive, network and RS232 serial systems.

Mark

Tis a shame they did not get to be able to do this, "too many sold" does not seem a good reason, then again in penny pinching (maybe?) and perhaps height saving there's not usually a socket. I think and I don't mean this in a bad way towards anyone who was involved at the time a bit of an "an own goal"

As my thoughts to 'upgrades from the manufacturer' , I immediately thought of the early Acorn Proton (aka 'BBC Micro') with the early OS ROM in 2 chips, very few of these survive now as the OS was a bit um 'missing a few things' compared to 1.2, you know what I mean anyway I think I've probably explained it quite badly.

Unfortunately I get the impression that it was organised chaos at Sinclair. This year we only found out that there are actually THREE different lots of ROM code for the ZX81 (previously we were only aware of two). And this does not include the hardware kludge used to overcome the maths bug in the first version of the ROM. So there were two bug fixes…

Only last year did we discover that there was another ZX Spectrum ROM (the one that is discussed in the topic above).

And with the QL, long after Amstrad took over the computers part of Sinclair Research Ltd, we found out that Sinclair was producing QL computers with older versions of the ROM code, even though they had been supplied with newer ROM code updates that had bug fixes weeks earlier!

At least with the QL they did have a programme to replace the ROMs (although we think some, maybe most boards got changed rather than just the ROMs on there own, boards lower than issue 5 are rather uncommon).

Anyway, water under the bridge now…

Mark

1024MAK wrote: ↑Tue Mar 31, 2020 10:19 pm Question, what software is ever completely finished?

DI
HALT