Assembler Editor and Dos I, was there an earlier release?

[tschak909]

I find myself wondering something very interesting...

 

Assembler Editor was one of the two cartridges written by Shepardson Microsystems (SMI) during the late 1978 push to get system software out the door.

 

And the original Atari DOS I FMS had been formally released on 9/24/79 (September 24, 1979)

 

It uses a different pair of binary header segment bytes: hex 84 09

(I'd love to know why...would love to see the source code!)

 

But assembler (ASM) emits a standard segment pattern of FF FF for each *= encountered in the source.

 

And in fact, there is an appendix in every version of the Assembler/Editor manual that I can find that contains instructions for patching DOS I's Binary Load and Save routines to emit FF FF, making it capable of loading and saving DOS II binary files, and thus being compatible with the output of ASM.

 

So there's a bit of an anachronism here: Was Assembler Editor patched quietly at some point to emit DOS II binary files?

 

Or was Assembler/Editor officially released much later than what is typically thought?

 

-Thom

 

 

[Rybags]

AsmEd actually creates lots of headers and short segments from memory (a bit annoying actually), something like 128 or 256 bytes in length?

 

Earlier version - not sure.  Possibly they held off on it's release though you'd have to wonder then what tool was available (Atari's AMac, though wasn't that later?)

 

Do we know when DOS 2.0S came out?  Really, DOS 1 was totally not worth using compared to it.  The only reason I can think is if it had a smaller memory footprint (which I sort of doubt)

[Faicuai]

19 hours ago, tschak909 said:

I find myself wondering something very interesting...

 

Assembler Editor was one of the two cartridges written by Shepardson Microsystems (SMI) during the late 1978 push to get system software out the door.

 

And the original Atari DOS I FMS had been formally released on 9/24/79 (September 24, 1979)

 

It uses a different pair of binary header segment bytes: hex 84 09

(I'd love to know why...would love to see the source code!)

 

But assembler (ASM) emits a standard segment pattern of FF FF for each *= encountered in the source.

 

And in fact, there is an appendix in every version of the Assembler/Editor manual that I can find that contains instructions for patching DOS I's Binary Load and Save routines to emit FF FF, making it capable of loading and saving DOS II binary files, and thus being compatible with the output of ASM.

 

So there's a bit of an anachronism here: Was Assembler Editor patched quietly at some point to emit DOS II binary files?

 

Or was Assembler/Editor officially released much later than what is typically thought?

 

-Thom

 

 

From what you describe, I can only conclude that they already knew in advance what was coming.

 

To confirm (or not) the above, we will need to plot DOS-I, ASM Editor and DOS-II (temporal) development timelines, and see where they intercept.

 

The fact that ASM-Editor already generated $FF:$FF header and DOS-I retro-patch is documented says a lot.

Edited April 10, 2021 by Faicuai

[Rybags]

84 09 wouldn't be here or there as a file header but you'd have potential havoc for subsequent sections since the FF,FF is optional there.

 

In the real world you never get FFFF as a segment start address but 0984 is in low memory and perfectly valid to use as a start address, though not possible if in use as a header sequence.

Have to wonder what the reasoning was for using 84,09.

[+slx]

Not an answer, just a puzzle piece, but according to the Atari BASIC Source Book Atari BASIC and DOS were assembled on an IMP-16-based system at Shepardson Microsystems, so there wasn't necessarily an Atari 800-based assembler before Atari Assembler/Editor. Before farming out to SMI, Atari probably used the same minicomputer-based cross-assembler systems used for VCS development.

[tschak909]

The IMP-16 was used at SMI as their primary assembler. Paul (or Bill, or Kathleen...) would write out their code onto code sheets, and hand them to Mike Peters to feed as cards, outputting a paper tape, which would then be read into the target system (Paul used a punch card reader hooked to a board that Woz had built, in an Apple ][). 

 

The FMS was initially brought up on the Apple ][ (using even a ported version of the CIO in the OS ROM), and debugged there until it could self host on the Atari (they had to write a simple emulator that hooked CIO up to RWTS and simulated Atari 128 byte sectors).

 

Atari themselves initially used a cross-assembler from Microtec on a PDP-11 initially, downloading code directly to a specially modified Atari (which added a 6850 ACIA, and two high speed serial connections in the unused left front section of the Atari 800 casing)

 

I do remember an early assembler from Eastern House (it pre-dates MAE even, and was loaded from tape) circa 1980...

 

Most of the third party tools that most people used for software development that we think of for hosted Atari 800 development arrived at the beginning of 1981, MAE, Synassembler, and Atari finally released AMAC in the spring of 1981. They had a version of it on their newly acquired MV8000 which they called CAMAC before AMAC was released.

 

-Thom

 

Edited April 10, 2021 by tschak909

[ColleenLover]

WOW-- Seriously, sometimes the depth of knowledge of some of our members is truly astounding.

 

Thank you.

 

[Faicuai]

On 4/10/2021 at 12:37 PM, tschak909 said:

The FMS was initially brought up on the Apple ][ (using even a ported version of the CIO in the OS ROM), and debugged there until it could self host on the Atari (they had to write a simple emulator that hooked CIO up to RWTS and simulated Atari 128 byte sectors).

RESPECT.

 

All I can say (plus amazing, minute-detail anecdotes on your end...) 

 

??

[tschak909]

6 minutes ago, Faicuai said:

RESPECT.

 

All I can say (plus amazing, minute-detail anecdotes on your end...) 

 

??

Paul Laughton and Bill Wilkinson went into great detail on their workflow over the last couple of decades. There's enough detail to weave the pieces together with very few gaps.

 

-Thom

 

[tschak909]

and yes, CIO was actually ported to the Apple ][, it was part of CP/A, the original name for OS/A+. There was indeed an Apple version, and the earliest manuals for OS/A+ mention the differences between the Apple and Atari versions. 

 

It was rather bizarre because you could send commands to devices, not just by traditional physical slot calls to option ROMs, but through the hooks reserved for sending commands to the DOS from BASIC (the CTRL-D hook), by specifying device names, e.g. P: for printer, D1: for first disk, etc.

 

But this version of CP/A has yet to be found. It's probably in a pile of disks at Bill Wilkinson's house.

 

(the SMI / OSS people really liked CIO, a lot, and Bill Wilkinson mentioned in the Beneath Atari DOS manuals that his ideal setup would be an AIM/65 with Apple ][ slots, and the CIO for I/O) 

 

-Thom

[+kheller2]

59 minutes ago, tschak909 said:

Paul Laughton and Bill Wilkinson went into great detail on their workflow over the last couple of decades. There's enough detail to weave the pieces together with very few gaps.

 

-Thom

 

Any pointers to these details that one could read up on?

[ClausB]

[+hunmanik]

I don't believe DOS II came out until summer 1981, but the DOS II FMS was completed 19-Aug-80 according to the source code.

 

I had narrowed the Assembler Editor ship date to Summer or Fall 1980.  Probably more like fall 1980 based on this contingency on DOS II already being completed that you point out.

[tschak909]

wonder if anyone has the source code to DOS I ?

 

-Thom

 

[Atari_Ace]

On 4/9/2021 at 5:47 PM, Rybags said:

AsmEd actually creates lots of headers and short segments from memory (a bit annoying actually), something like 128 or 256 bytes in length?

I thought that was AMAC which did that, limiting segments to 256 bytes.

[+Stephen]

1 hour ago, tschak909 said:

wonder if anyone has the source code to DOS I ?

 

-Thom

 

All I have is this manual.

Atari DOS 1.0 Manual.zip

[+slx]

17 hours ago, ClausB said:

Interesting how it doesn’t actually say ‚for Apple‘ in the ad, except for the logo on top. 

[+Nezgar]

On 4/9/2021 at 6:47 PM, Rybags said:

AsmEd actually creates lots of headers and short segments from memory (a bit annoying actually), something like 128 or 256 bytes in length?

Yeah, 128 bytes - you can hear this when running such a program from a stock/unbuffered 1050 or Rev C 810 - since DOS 2.x only has 125 bytes of data per sector, each 128 byte segment takes 2 sector reads, then a pause to process the next segment meanwhile blowing a disk rev, then 2 sectors again, and repeat. The Atari 1050 diagnostics program does this and it drove me nuts so I ran it through the streamline utility to reduce it to like 4 segments... Boots much faster now, especially on a disk reformatted by a US doubler, or similar drive that uses the "Chicago" interleave. I also prepended Doug Wokoun's Screen Accelerator Wedge for fun too. Blazin!

[ClausB]

On 4/11/2021 at 3:31 PM, ClausB said:

OSS Apple manuals here:

http://www.trailingedge.com/apple2/

[tschak909]

4 hours ago, ClausB said:

OSS Apple manuals here:

http://www.trailingedge.com/apple2/

Yup, somebody bought a copy. Where the $#(@(@# is it?!

 

-Thom

 

[thorfdbg]

On 4/10/2021 at 2:47 AM, Rybags said:

Really, DOS 1 was totally not worth using compared to it.  The only reason I can think is if it had a smaller memory footprint (which I sort of doubt)

No, it has a larger memory footprint. The reason being is that the "DUP Menu" was part of DOS.SYS, i.e. DOS I does not have a DUP.SYS, and a MEM.SAV as we know it from DOS 2.0S. Thus, indeed, a total waste.

 

[luckybuck]

There were 2 different types of the Atari Assembler cartridge:

 

https://atariwiki.org/wiki/Wiki.jsp?page=Atari Assembler Editor

 

Have asked Kathleen and she agreed, so this makes the carts type A and B 'official' after all the years.

Did not check what exactly has been changed, but maybe, this is of some kind of help?

[DavidMil]

Man are we talking about a long time ago!  When I try to remember that far back all I see is cobwebs...  I did get out my old, 'ASSEMBLER EDITOR

USER'S MANUAL' (which has a copyright date of 1980 Rev. 1), and it has almost two full pages of saving and loading files from a cassette tape and

about 2/3 of a page talking about saving and loading DOS 1 files.

 

David 

[tschak909]

Just tested, the revision A still outputs DOS II binary files.

 

So am really starting to think that poor DOS 1 users were without a functional assembler from Atari for a good 6-9 months.

 

-Thom

 

[luckybuck]

Hi Thom, Yes, there werde some guys in Eastern Germany, who missed just any assembler and therefore inserted the code via POKE....

 

Anyway, there were 2 stickers on the cart, if, then just the cart with the sticker: 'EDITOR ASSEMBLER' could be the candidate. Atari Geezer is our Yoda of the carts. If he can say for sure, that his 'EDITOR ASSEMBLER' cart is the A type cart, then indeed, no assembler for DOS I. Carol Shaw told me, that they had the Microtec cross assembler, when they compiled code in the beginning of ATARI. – For me of great interest is: who build AMAC? There is a source from Charles, it came from APX, but this is only true for MEDIT. Further,  no source code or trace for EASMD, just the disks. But from OSS, too.