ISIS 1.1 and 2.2

116 views
Skip to first unread message

Bill Beech (NJ7P)

unread,
Apr 28, 2021, 1:21:39 AM4/28/21
to Intel
Guys,

I know there was a heavy discussion on at least the V1.1 code and maybe
on the V2.2 code.  The V1.1 code is for a 16KB machine.  Was this maybe
the Intellec8/80 which was a 16KB machine with the 100-pin Intellec bus?

Judging from the DISKIO code in each, they both target the same machine.

Bill

Vale, Martyn

unread,
Apr 28, 2021, 4:22:07 AM4/28/21
to intel-...@googlegroups.com

Hi Bill,

I think the consensus of opinion was that V1.1 was for the MDS-800 and the premise was that it would allow the Intel marketing people to sell a cheaper 16K Base MDS-800 System.

V1.1 does run on my MDS-800 as demonstrated and V2.2 will as well, the main difference is that V2.2 is much more like the later V3.x variants and although still Single Density only in it's original form you can create a Hybrid version by swapping out ISIS.BIN from say V3.4 this will allow Double Density operation. It's unlikely either V1.1 or V2.2 will run on an MDS II, however if you connected via the serial port this may well work although I'm not in a position to confirm this currently.

From what I've read the Intellec8/80 was paper tape only as supplied by Intel, but there were attempts at various third party disk controllers for it. The code in V1.1 is definitely aimed at the SBC 201 and this wouldn't fit in the Intellec8/80 presumably.

Bye
Martyn.
--
You received this message because you are subscribed to the Google Groups "intel-devsys" group.
To unsubscribe from this group and stop receiving emails from it, send an email to intel-devsys...@googlegroups.com.
To view this discussion on the web visit https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgroups.google.com%2Fd%2Fmsgid%2Fintel-devsys%2F0b6e8c24-b403-74dd-7c1d-2d9b4c1fafeb%2540nj7p.info&data=04%7C01%7C%7C48615db72e554793ec8b08d90a05811c%7C1faf88fea9984c5b93c9210a11d9a5c2%7C0%7C1%7C637551841029289448%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000&sdata=IJ1Vt%2Fe2r6xIJpxBItUz7vqoq1kWo5JEdRfLfevhYLc%3D&reserved=0.

Bill Beech (NJ7P)

unread,
Apr 28, 2021, 4:26:26 PM4/28/21
to intel-...@googlegroups.com
Martyn,

Good point.  I know of no FDC for the Intellec bus. And the initial
MDS-800 was a 16KB machine!

Thanks!

Bill

Bill Beech (NJ7P)

unread,
Apr 28, 2021, 4:47:09 PM4/28/21
to intel-...@googlegroups.com
Is an original .IMD disk image for V1.1 and V2.2 avaIlable?

Thanks!

Bill

Herb Johnson

unread,
Apr 28, 2021, 4:52:53 PM4/28/21
to intel-...@googlegroups.com, Bill Beech (NJ7P)
I'm getting an impression, that "ISIS V1.1" described, is related in
some way to the original ISIS, before "ISIS II". I don't believe the
original ISIS was marketed or developed for the pre-Multibus "Intellec 8
mod 80" system. As Bill points out, Intel did not produce an FDC for the
8080 Intellec. It's possible that machine was used for development.

what I've gathered over time (and from our colleagues here) on ISIS
not-II is on my Web page:

https://www.retrotechnology.com/dri/isis.html#isis

There's first-person interview quotes there about the development of
ISIS "for the upcoming MDS800 Multibus chassis" and as "PL/M code". that
was followed by an assembly-language version (by disassembly of PL/M and
refinement). Those of you staring at the "ISIS V1.1" code can consider
if it's PL/M, assembler, or some mash-up.

It would be good from my view, to confirm that "ISIS 1.1" is the
original "ISIS" before "ISIS II"; or at least derived from it. I'm not
aware of any surviving "ISIS no-II" disks or code. Of course I may be
uninformed.

This is a similar problem to identifying "the Multibus" before Intel
decided to call their new, not-the-Intellec buss architecture, "the
Multibus". I believe Intel engineers designed a new architecture first,
and called it "the Multibus" later.

People forget (or don't know) how ad-hoc all this microprocessor stuff
was in the earliest years. These were not grand product plans, by
Masters of the Universe with Great Vision. It's just what survived and
became products, versus what did not.

Regards, Herb
--
Herbert R. Johnson, New Jersey in the USA
http://www.retrotechnology.com OR .net
preserve, recover, restore 1970's computing
email: hjohnson AT retrotechnology DOT com
or try later herbjohnson AT comcast DOT net

Bill Beech (NJ7P)

unread,
Apr 28, 2021, 5:19:14 PM4/28/21
to Herb Johnson, intel-...@googlegroups.com
Herb,

I agree with what you have gathered as the early history of ISIS-(*).

Good work!

Bill

Bill Beech (NJ7P)

unread,
Apr 28, 2021, 5:38:06 PM4/28/21
to Herb Johnson, intel-...@googlegroups.com
Guys,

There was the MDS Disk Operating System listed with the MDS-800 in the
1975 Data Catalog.  I have never seen the actual product. Might have had
a name change to ISIS-(*) before it was ever distributed?

Bill

On 4/28/2021 1:51 PM, Herb Johnson wrote:

Vale, Martyn

unread,
Apr 28, 2021, 5:51:00 PM4/28/21
to intel-...@googlegroups.com
Hi Bill,

I don't think there are "Originals" of V1.1 or V2.2. Both are in Mark's repository, however the V1.1 is the one Jon recovered from the Museum Collection and the V2.2 one has no ISIS.LAB, so I guess this is a recovery of a copied disk it does work however if you create a recipe file for it and use Mark's mkidsk utility. The V2.2 also includes ICE80 and looks a bit bare to me. The V1.1 looks more complete but has no original Intel Label.
--
You received this message because you are subscribed to the Google Groups "intel-devsys" group.
To unsubscribe from this group and stop receiving emails from it, send an email to intel-devsys...@googlegroups.com.
To view this discussion on the web visit https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgroups.google.com%2Fd%2Fmsgid%2Fintel-devsys%2F8ce23077-4254-22f7-6a21-f36f9e5641ab%2540nj7p.info&data=04%7C01%7C%7Cba824f2032064007fa6508d90a86cba9%7C1faf88fea9984c5b93c9210a11d9a5c2%7C0%7C1%7C637552396327731767%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000&sdata=adTOsp4liYlthknl0UUIWkgPdG1cG37UAVJ4yWIxlaU%3D&reserved=0.

Jon Hales

unread,
Apr 29, 2021, 5:14:17 AM4/29/21
to intel-...@googlegroups.com
Martyn, Bill

On the subject of whether ISIS v1.1 was developed on/for the Intellec 8/80 or for the MDS-800, you may find it interesting to review a section of Intel's 'Microcomputer Fair' in London in May 1975 (I posted a PDF some weeks ago). This section (AP-7; 98-031A, dated 1975) has the heading "Disk Controller Design Uses New Bipolar Microcomputer LSI Components". It goes into some detail of the design using eight 3002 and eight 3601 chips and provides a schematic drawing. It's clearly only part of the story, because the 'controller' has no connectors for the drive. I assume this design may have evolved into the iSBC-201 and iSBC-202.

Figure 1 is a photo of the disk (floppy) controller board, showing the backplane interface. This has an edge connector with 50 contacts per side. The orientation is 'vertical', rather than the 'width greater than height' arrangement for the Multibus.

It seems plausible that a disk controller of the type pictured may have been in use at Intel when the company's Disk Operating System was in early development.

On a related topic:
(From memory - caveat) Gary Kildall attempted to design a disk interface for an unspecified system to which he had access - which might have been an Intellec/8. The record refers to Kildall's friend John Torode having developed a disk interface. We know that CP/M was issued on floppy disks 'in MDS-800 format', which I think means single sided and single density. However, earlier development may have been on a different system with a floppy controller - possibly an Intel design, possibly not. 

Herb's website has material on this episode. However, I think the question of the system on which early development of CP/M took place has not been documented - I would assume it had a controller of some sort and an 8 inch drive.

Regards

Jon

Herb Johnson

unread,
Apr 29, 2021, 1:10:27 PM4/29/21
to intel-...@googlegroups.com, Jon Hales
>> On the subject of whether ISIS v1.1 was developed on/for the Intellec
>> 8/80 or for the MDS-800,

> Intel's 'Microcomputer Fair' in London in May 1975 (I posted a PDF
>> some weeks ago). This section (AP-7; 98-031A, dated 1975) has the

I don't know where that document was posted or archived: I could not
keep up with the flurry of document-archiving discussed weeks ago.
However, this document (intel Application Note AP-7) I found at
"usermanual.wiki" with a Web search of the title. Also, archive.org.

https://archive.org/details/intelap7diskcontrollerdesign/page/1/mode/2up

https://usermanual.wiki/Document/AP7DiskControllerDesignUsesNewBipolarMicrocomputerLSIComponentsJan75.816684720/help

The document is copyrighted 1975. There's a reference to an Sept 5 1974
article in Electronics (a US trade magazine) I should look up.

Point of time reference: Kildall describes his CP/M development work as
sometime in 1974, in the fall of 1974. Kildall's Digital Research
apparently was created around the beginning of 1975. Intel's ISIS dates
in about the same time period.

The document describes the design as quote "Applications Research group
at Intel with the design of a 2310/5440 moving head disk controller (BMDC)"

Wikipedia identifies these drives as "The IBM 2310 Removable Cartridge
Drive was announced in 1964 ... It could store 512,000 16-bit words on
an IBM 2315 cartridge. A single 14-inch (360 mm) oxide-coated aluminum
disk spun in a plastic shell with openings for the read/write arm and
two heads." An IBM 5444 was similar but with one or two drive packs and
different track per inch counts.

So: it's not for a floppy drive, but it's of-a-kind for a floppy drive
controller.

> the 'controller' has no connectors for the
>> drive.

Sure it does. there's two connectors on the board, at the bottom of the
photo and at the top. The bottom connector is 100 pins, as I recall the
Intellec used 100 pin connectors. the top connector is 44 pins, enough
for a floppy interface.

Actually looking at the schematic? I think the square-symbols are the
buss connector, the diamond-symbols are the drive connector (from pin
counts).

> design may have evolved into the iSBC-201 and iSBC-202

(shrug) compare the schematics. Certainly to a first order, it's
plausible that work done for this demonstration controller in 1974-75,
led to the iSBC-201.

I'd be curious to have someone actually LOOK at the schematics in this
document versus the iSBC-201 and draw some conclusions. If I can, if I
have time, I'll look and see if the 100-pin interface has any
relationship at all, to the 8080 Intellec CPU board's 100 pins. My read
today of the document finds no reference.

> It seems plausible that a disk controller of the type pictured may have
>> been in use at Intel when the company's Disk Operating System was in
>> early development.

Not so much.

https://www.retrotechnology.com/dri/isis.html#isis

Developer Ken Bergett's recollections
http://www.rogerarrick.com/osiris/burgett.txt

as interviewed by email by Roger Arrick, and on Roger's Web site. It's a
first-person account, what can I say. But it is 40, 50 years after the
facts.

At the time Ken wrote of, floppies were brand-new. Ken was skeptical
about them, he says. Ken says some tech wire-wrapped a UART based
controller, just to get started. From my own recollections about floppy
controllers of the period, plus a history of S-100 and other floppy
controllers as described on my Web site, that's not an uncommon means.
UARTS were used on microcomputers to drive audio/analog cassette storage.

http://www.retrotechnology.com/dri/first_floppy.html

But Ken doesn't describe the production floppy controller which was
apparently sold with the first "ISIS". His description is/was: quote

After ISIS was released, written in PL/M, consuming 12k of ram and
supporting an assembler and text editor (also written in PL/M), all
fitting into a 32kb ram configuration - end quote

Whether that version in distribution was that PL/M program, or Ken's
disassembly-optimized-assembled version, is not clear. Quote - "
convert[ed] the generated PL/M code back to assembler, which I then
hand-tuned to squeeze 12k bytes into an 8k bag."

None of this, by the way, sounds like the "ISIS v1.1" codes, which I've
had zero time to look at myself. It may be that the ISIS V1.1 is
"really" ISIS-II. (shrug) I dunno.

The mysteries around Intel's first ISIS OS and hardware, apparently
continue. But there's not much mystery around CP/M's development, just a
few details.

>> Herb's website has material on this episode. However, I think the
>> question of the system on which early development of CP/M took place has
>> not been documented - I would assume it had a controller of some sort
>> and an 8 inch drive.

https://www.retrotechnology.com/dri/d_dri_refs.html

You know, I spent several years on exactly. this. topic. , Jon. It was
several years ago, before more documents were available and became
easier to find on the Web. My material was referenced, when Kildall's
first Digital Research offices were memorialized with IEEE recognition,
on the 40th anniversary of CP/M.

Kildall himself, described the development of CP/M, but he did not name
his target development system. He did a lot of work in software
emulation and cross-compilation (PDP-10, Fortran-based PL/M compiler and
assembler). And, he had at times, an Intel Intellec 8 upgraded to an
8080; and an Intel MDS-800 (Multibus). As Intel's software developer,
why would he not?

but Ken's account - described 40 years later, keep in mind - says: quote:

> Gary did manage to get one of the early production MDS 800 systems, with
> the disk controller we built for the ISIS project, trading for something
> Marketing wanted done, and the development of CP/M proceeded in parallel
> with our work on ISIS. Gary and I discussed possible solutions for doing
> file allocation and I/O, and how to manage the disk controller, and such
> like, but each DOS was developed separately.

It's news that Kildall got a disk controller from Intel; I'm skeptical.
But it was not what Kildall ended up using. Kildall says (published
accounts) that Intel declined interest in his CP/M, and went on to
produce an ISIS OS. Kildall did not provide details that I'm aware of.

It's informative that Kildall got an MDS-800 "early". Someone could help
me, by dating when MDS-800 systems were first produced (made available
to customers).

But I'm not aware of any description by or about Kildall, on which Intel
system he ran his first CP/M, to operate on a Shugart floppy drive, to
first physically boot CP/M, in the 1974/early 75 timeframe.

Kildall himself - in public accounts not many years after the events-
said he tried and failed to build a working floppy controller. Kildall
himself, said Dr. John Torode built a controller for Kildall, which worked.

And i document the microcomputer history of Dr John Torode and his
Digital Systems (subsequently Digital Microsystems) in exquisite detail
on my Web site, down to the schematics of the controllers and S-100
interfaces his company produced - derivatives of the microprogrammed
controller he made for Kildall.

I even describe Torode's PhD advisor Dr. Ted Kehl who developed
micro-programmable controllers in the early 1970's.

http://www.retrotechnology.com/dri/torode_origins.html

John Torode himself, *told me personally* it was his controller, not
Kildall's, that first ran CP/M; and it was like the controllers Torode
produced immediately thereafter:

https://www.retrotechnology.com/dri/d_dri_refs.html#torode_work
2008 discussion with Dr. John Torode

And I *own* some Digital Systems floppy controllers:

https://www.retrotechnology.com/restore/digsys.html

There's a point to this apparent rambling. All the early Digital System
floppy controllers, have a simple parallel-like interface. DS sold the
controller and described a separate interface design for some arbitary
buss, such as the "Altair bus". But they soon sold an Altair bus
interface card (which was later named "the S-100 bus").

So: a Torode floppy controller, could well have been built by Torode,
for either the Intellec "bus" or a Multibus. Since the Torode products
could be operated by a 2MHz clocked 8080 (the MITS Altair 8800); either
the Intellec 8/80 or the MDS-800 could have operated Torode's hand-built
controller for Kildall.

If my Web pages on the subject seem a little rough and disjointed, it's
because it was accumulated work of several years, done over several
years ago before more documents were online. No one was particularly
interested at the time. Fortunately, that disinterest changed, and now
Kildall's legacy is more likely to be preserved. And now, for reasons I
don't have to detail, the legacy of early ISIS development is becoming
preserved. and, the two are intertwined with the legacy of floppy disk
drive and controller development; and the association of the principals
for CP/M and for ISIS through Intel. Thus my interest in both.

Regards, Herb Johnson
> <mailto:intel-devsys%2Bunsu...@googlegroups.com>.
> <https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgroups.google.com%2Fd%2Fmsgid%2Fintel-devsys%2F8ce23077-4254-22f7-6a21-f36f9e5641ab%2540nj7p.info&amp;data=04%7C01%7C%7Cba824f2032064007fa6508d90a86cba9%7C1faf88fea9984c5b93c9210a11d9a5c2%7C0%7C1%7C637552396327731767%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000&amp;sdata=adTOsp4liYlthknl0UUIWkgPdG1cG37UAVJ4yWIxlaU%3D&amp;reserved=0>.
>
> --
> You received this message because you are subscribed to the Google
> Groups "intel-devsys" group.
> To unsubscribe from this group and stop receiving emails from it,
> send an email to intel-devsys...@googlegroups.com
> <mailto:intel-devsys%2Bunsu...@googlegroups.com>.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/intel-devsys/DB8PR01MB56593D8B2ACD41996CA125FCC1409%40DB8PR01MB5659.eurprd01.prod.exchangelabs.com
> <https://groups.google.com/d/msgid/intel-devsys/DB8PR01MB56593D8B2ACD41996CA125FCC1409%40DB8PR01MB5659.eurprd01.prod.exchangelabs.com>.
>
> --
> You received this message because you are subscribed to the Google
> Groups "intel-devsys" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to intel-devsys...@googlegroups.com
> <mailto:intel-devsys...@googlegroups.com>.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/intel-devsys/CANPXPJmhtToH4ygUQhur8sQKqFvgXwS%3DZ1gyz2EKWc%3DFhXPjxQ%40mail.gmail.com
> <https://groups.google.com/d/msgid/intel-devsys/CANPXPJmhtToH4ygUQhur8sQKqFvgXwS%3DZ1gyz2EKWc%3DFhXPjxQ%40mail.gmail.com?utm_medium=email&utm_source=footer>.

mark.p...@btinternet.com

unread,
Apr 29, 2021, 1:37:08 PM4/29/21
to intel-...@googlegroups.com

Roger Arrick

unread,
Apr 29, 2021, 1:44:18 PM4/29/21
to intel-...@googlegroups.com
> People forget (or don't know) how ad-hoc all this microprocessor stuff
> was in the earliest years. These were not grand product plans, by
> Masters of the Universe with Great Vision. It's just what survived and
> became products, versus what did not.
-Herb



This is SO true Herb. There was no master plan, the endpoint could not be seen.
It sorta happened TO US in a way.

As someone who developed long term product lines from scratch, I can tell you
that product lines take on a life of their own. The market pulls in one
direction or another, vendors, parts-availability, finances, employees,
buildings, assembly lines all do the same.

You can often see this best in incoherent product names and part numbers.

As the future becomes clearer, we struggle to keep things on a preconceived
path. A smart courageous company will be willing to sacrifice their current
cash cows for future unknowns.

Interesting discussion.

_______________________________________________
Roger Arrick
Ro...@Arrick.com
Tyler, TX

mark.p...@btinternet.com

unread,
Apr 29, 2021, 2:01:55 PM4/29/21
to intel-...@googlegroups.com
Herb
The ISIS 1.1 I decompiled shows that it is hand optimised PL/M and fits in 7168 bytes.
Unlike full implementations of ISIS it does not support SEEK.
This is consistent with the comments in Ken Burgett's interview
Mark
To view this discussion on the web visit https://groups.google.com/d/msgid/intel-devsys/004d01d73d1e%24467a1c80%24d36e5580%24%40btinternet.com.

Roger Arrick

unread,
Apr 29, 2021, 2:05:16 PM4/29/21
to intel-...@googlegroups.com
Bill Beech (NJ7P) wrote:
> There was the MDS Disk Operating System listed with the MDS-800 in the
> 1975 Data Catalog. I have never seen the actual product. Might have had
> a name change to ISIS-(*) before it was ever distributed?



I looked for MDS800 in here and don't see it,

https://ia800408.us.archive.org/0/items/bitsavers_inteldataBtalog_31960676/1975_Intel_Data_Catalog.pdf

Herb Johnson

unread,
Apr 29, 2021, 2:30:40 PM4/29/21
to intel-...@googlegroups.com, Roger Arrick
Thank you, Roger. One reason I preserve the past, is that it often
gets repeated. Some say "it rhymes". The examples of the Apple Mac 128k
(and the Lisa) with cash-cow Apple II, and the IBM-PC are informative.
They were informed by the Apple II, S-100 products and CP/M, and by
Intel as we discuss here.

- regards, Herb

Herb Johnson

unread,
Apr 29, 2021, 3:04:30 PM4/29/21
to intel-...@googlegroups.com
> -----Original Message-----
> From: Behalf Of Herb Johnson
> Sent: 29 April 2021 18:09
>
> Developer Ken Bergett's recollections
> http://www.rogerarrick.com/osiris/burgett.txt
>
> But Ken doesn't describe the production floppy controller which was
> apparently sold with the first "ISIS". His description is/was: quote
>
> After ISIS was released, written in PL/M, consuming 12k of ram and
> supporting an assembler and text editor (also written in PL/M), all
> fitting into a 32kb ram configuration - end quote
>
> Whether that version in distribution was that PL/M program, or Ken's
> disassembly-optimized-assembled version, is not clear. Quote - "
> convert[ed] the generated PL/M code back to assembler, which I then
> hand-tuned to squeeze 12k bytes into an 8k bag."
>
> None of this, by the way, sounds like the "ISIS v1.1" codes, which I've
> had zero time to look at myself. It may be that the ISIS V1.1 is
> "really" ISIS-II. (shrug) I dunno.


> On 4/29/2021 2:01 PM, mark.pm.ogden via intel-devsys wrote:
>> Herb
>> The ISIS 1.1 I decompiled shows that it is hand optimised PL/M and fits in 7168 bytes.
>> Unlike full implementations of ISIS it does not support SEEK.
>> This is consistent with the comments in Ken Burgett's interview
>> Mark
>>

Thanks Mark for working this detail. So let's review the trail in the
Burgett interview:

> We decided early on that we would write the entire system
> as a single PL/M program ...

and there's specific details in the paragraph. Next about development is

> The final ISIS memory footprint was 12 kb, which made running an 8k
> assembler in a 16k ram system an impossibility. Marketing had a
> melt-down. After a great display of angst, Marketing finally agreed to
> ship a 32k RAM minimum ISIS dual-floppy system. For this, they demanded
> that I produce an assembly language version of the OS,

> The 16k RAM requirement
> stayed, but the single drive idea was dropped. ...
>
> After ISIS was released, written in PL/M, consuming 12k of ram and
> supporting an assembler and text editor (also written in PL/M), all
> fitting into a 32kb ram configuration, I did produce a hand-built
> assembly language version....
> [with] a dis-assembler to convert the generated PL/M
> code back to assembler, which I then hand-tuned to squeeze 12k bytes
> into an 8k bag. I had to drop a random access ‘seek’ feature from the
> file system... We sold two of those systems.

Note the sequence: 1) ISIS was released, written in PL/M; 2) a
hand-built assembly language version was produced by dis/re/assembly. 3)
"we sold two of those systems", the squeezed version.

So maybe, ISIS V1.0 or no version, was the PL/M program? and ISIS V1.1
which is now in-our-hands, is the "squeezed" 8K assembly version?

Reading along Ken's account:

> ISIS did very well in the market and had a stunning 0 bug reports over
> the first year. [Remark about Kildall and "parallel development of CP/M"]

Remarks about Kidall and CP/M put the time frame as 1974 into 1975.

> That done, it was time to make the PL/M compiler run on ISIS.

Additional comments name these features: object module definition,
assembler to support that, ICE support. Ken was Intel's "Release Chief
to make sure all the parts got done on time and integrated". So he
should know.

> A good friend of mine, Bruce, got the job of developing the Object
> Module Format (OMF) for the 8080, and the upcoming 8086.

I call out the OMF, because I know Bill Beech has worked hard on nailing
down the variations on the OMF. He can probably date the first versions
of that software product. The OMF and object modules were "how ISIS
handled load modules".

--------------------

So now we have an operational definition of ISIS-II. It supported a
native PL/M, an object-module based assembler, and a loader. If our ISIS
V2.2 has those features, then it's the start of ISIS-II. Since ISIS V1.1
doesn't have those features, it's a candidate for ISIS not-II;and in
particular the reassembled version of ISIS not-II.

That's my read of the document.

Regards, Herbn

mark.p...@btinternet.com

unread,
Apr 29, 2021, 3:34:57 PM4/29/21
to intel-...@googlegroups.com
Herb
Based on some of the object files in some of the earlier intel .lib files, there were versions of PL/M which created .OBJ files prior to the known native version 3.0. These may have been cross compilers or internal only ones. The V1.0 PL/M compiler generated inline code for utility functions. These were moved to PLM80.LIB by v3.0. Additionally, I am aware of some variants of the V3.1 compiler in terms of code generation, which suggests Intel had internal versions of the later compiler as well.
The Fortran based cross-compiler v2.0 and v4.0 generated located binary files and not .OBJ files. They also did not support STRUCTURE or AT and there are some differences for DATA declarations.

As to ISIS I vs. ISIS II, OMF was used for ISIS II but not ISIS 1, although ISIS.BIN remained in the old ISIS I Binary format, until ISIS III. Early versions of ISIS II contained utilities to convert from BIN to OBJ format.

Note there may have been other ISIS I 1.x and ISIS II 2.x and 3.x variants, possibly never released publicly.
What is interesting is the disk version naming convention suggests that ISIS II V3.4 was the 4th public version of ISIS
I am assuming ISIS 1.1 and ISIS 2.2 are two of them, although ISIS 1.1 may have been prior to the versioning model.
The version of 2.2 I have, came from a backup disk, so I do not know the true reference. This disk also had PL/M v3.0

Mark
--
You received this message because you are subscribed to the Google Groups "intel-devsys" group.
To unsubscribe from this group and stop receiving emails from it, send an email to intel-devsys...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/intel-devsys/2faeabaa-dc6f-69ac-c31b-441a3404c4d4%40retrotechnology.com.

Bill Beech (NJ7P)

unread,
Apr 29, 2021, 4:30:07 PM4/29/21
to intel-...@googlegroups.com

Jon,

On 4/29/2021 2:14 AM, Jon Hales wrote:
Martyn, Bill

On the subject of whether ISIS v1.1 was developed on/for the Intellec 8/80 or for the MDS-800, you may find it interesting to review a section of Intel's 'Microcomputer Fair' in London in May 1975 (I posted a PDF some weeks ago). This section (AP-7; 98-031A, dated 1975) has the heading "Disk Controller Design Uses New Bipolar Microcomputer LSI Components". It goes into some detail of the design using eight 3002 and eight 3601 chips and provides a schematic drawing. It's clearly only part of the story, because the 'controller' has no connectors for the drive. I assume this design may have evolved into the iSBC-201 and iSBC-202.
I have studied this document.  I support your assumption.


Figure 1 is a photo of the disk (floppy) controller board, showing the backplane interface. This has an edge connector with 50 contacts per side. The orientation is 'vertical', rather than the 'width greater than height' arrangement for the Multibus.
That is the connector used in the Intellec chassis that the Intellec 8/80 used.


It seems plausible that a disk controller of the type pictured may have been in use at Intel when the company's Disk Operating System was in early development.
Very likely!


On a related topic:
(From memory - caveat) Gary Kildall attempted to design a disk interface for an unspecified system to which he had access - which might have been an Intellec/8. The record refers to Kildall's friend John Torode having developed a disk interface. We know that CP/M was issued on floppy disks 'in MDS-800 format', which I think means single sided and single density. However, earlier development may have been on a different system with a floppy controller - possibly an Intel design, possibly not. 

Herb's website has material on this episode. However, I think the question of the system on which early development of CP/M took place has not been documented - I would assume it had a controller of some sort and an 8 inch drive.
I recall a picture of Kildall's office with an MDS-800 and 8-inch disk drives. Herb would know better!

Regards

Jon
Bill

Bill Beech (NJ7P)

unread,
Apr 29, 2021, 4:38:46 PM4/29/21
to intel-...@googlegroups.com
Herb,

Good input!

Maybe this became the iSBC 206?

Bill

On 4/29/2021 10:09 AM, Herb Johnson wrote:
>>> On the subject of whether ISIS v1.1 was developed on/for the
>>> Intellec 8/80 or for the MDS-800,
>
>>  Intel's 'Microcomputer Fair' in London in May 1975 (I posted a PDF
>>> some weeks ago). This section (AP-7; 98-031A, dated 1975) has the
>
> I don't know where that document  was posted or archived: I could not
> keep up with the flurry of document-archiving discussed weeks ago.
> However, this document (intel Application Note AP-7) I found at
> "usermanual.wiki" with a Web search of the title. Also, archive.org.
I has been on my site with a full size 98XXX number for years.

Herb Johnson

unread,
Apr 29, 2021, 6:06:43 PM4/29/21
to intel-...@googlegroups.com, Bill Beech (NJ7P)
On 4/29/2021 4:37 PM, Bill Beech (NJ7P) wrote:
> Herb,
>
> Good input!
> [various references to Intel App Note AP-7]
> Maybe this became the iSBC 206?
>
> Bill

Someone can compare the AP-7 controller schematic, to the schematics for
the iSBC 206, iSBC-201 and iSBC-202. Those who worked closely with the
iSBC controllers would be better judges than I. Of course in a general
way, it's a predecessor. The world of floppy and hard-disk controllers
before LSI chips is a curious bit of design history.

regards, Herb

Herb Johnson

unread,
Apr 29, 2021, 7:10:22 PM4/29/21
to intel-...@googlegroups.com
I have remarks on Mark's considerations, and also a "hit" on Kildall's
CP/M computer. Sorry for the length, I show my work as I go. - Herb

On 4/29/2021 3:34 PM, mark.pm.ogden via intel-devsys wrote:
> Herb
> Based on some of the object files in some of the earlier intel .lib files, there were versions of PL/M which created .OBJ files prior to the known native version 3.0. These may have been cross compilers or internal only ones. The V1.0 PL/M compiler generated inline code for utility functions. These were moved to PLM80.LIB by v3.0. Additionally, I am aware of some variants of the V3.1 compiler in terms of code generation, which suggests Intel had internal versions of the later compiler as well.
> The Fortran based cross-compiler v2.0 and v4.0 generated located binary files and not .OBJ files. They also did not support STRUCTURE or AT and there are some differences for DATA declarations.
>
> As to ISIS I vs. ISIS II, OMF was used for ISIS II but not ISIS 1, although ISIS.BIN remained in the old ISIS I Binary format, until ISIS III. Early versions of ISIS II contained utilities to convert from BIN to OBJ format.
>
> Note there may have been other ISIS I 1.x and ISIS II 2.x and 3.x variants, possibly never released publicly.
> What is interesting is the disk version naming convention suggests that ISIS II V3.4 was the 4th public version of ISIS
> I am assuming ISIS 1.1 and ISIS 2.2 are two of them, although ISIS 1.1 may have been prior to the versioning model.
> The version of 2.2 I have, came from a backup disk, so I do not know the true reference. This disk also had PL/M v3.0
>
> Mark

We are zeroing in on the taxonomy of these ISIS versions and their
related tools. We can only examine, the codes we have in hand of course.
Otherwise we have accounts, and whatever documents we can uncover.

If ISIS.BIN is loaded by some boot loader, it's plausible to have it in
absolute format - boot loaders are small and simple by circumstance.

ISIS V1.1 in-hand, disassembled, may inform us about its likely PL/M
source, possibly the version of the PL/M compiler as Mark suggests.
Since it's likely code already reduced by hand in assembly, that task is
harder. Put another way: ISIS before ISIS-II, could ONLY have been
produced by a PL/M cross-compiler; that would ONLY be the FORTRAN
programs that ran on various mini and mainframes. And we know the date -
1974 - because that work was concurrent with Kildall's CP/M work in 1974.

=======================================================

Some google work: "intel 8080 isis software" (and my Web search profile
at Google) was not further productive about ISIS. But it did resolve a
concurrent question, about Kildall's development system for CP/M.

My Web search, found the CHM's account of the development of CP/M.

https://computerhistory.org/blog/early-digital-research-cpm-source-code/

It quotes Kildall's Jan 1980 article in Dr Dobbs Journal, one familiar
to me. But it mentions something I'd forgotten. I go to the original
article:

https://archive.org/details/dr_dobbs_journal_vol_05_201803

"The History of CP/M" by Gary A Kildall. [following emphasis mine]

Kildall discusses his floppy-controller interface failures as follows.
"After several attempts at constructing a [Shugart floppy controller]
interface *to my Intellec-8*...." Meanwhile he continued CP/M
development "designed to support a resident PL/M complier". That was
part of the proposition he made to Intel - not an OS, but a resident
PL/M 8080 compiler, written in PL/M and then compiled to 8080 code as "a
next step" from the FORTRAN PL/M compiler Kildall provided to Intel.

Kildall continues: "The timesharing version of PL/M, along with the
Interp [8080] simulator, allowed me to develop and checkout the various
file operations to the level of primitive disk I/O."

"Shortly thereafter, *in the fall of 1974*, John Torode became
interested in the project." That puts a time-stamp on this CP/M work.
"Our first controller ... often performed ... just as requested. For
agonizing minutes, we loaded the CP/M machine code through the paper
tape reader into *the Intellec-8* memory. To our amazement, the disk
system went through its initialization and printed the CP/M prompt at
the Teletype."

So there it is. A 1980 account, of a 1974-5 event, first hand. Kildall
was running an Intellec 8, of course with an 8080 CPU board, and
Torode's completed floppy controller; for the first run of CP/M on
actual hardware.

As Kildall recounts, Torode's Digital Systems sold the first CP/M
commercial licenses in 1975, to Lawrence Livermore Labs and to Omron of
America for a terminal. Over about a year, Kildall's spare-time was
spent in developing the program tools associated with CP/M. And in 1976,
Kildall ported CP/M to IMSAI, at IMSAI - producing CP/M 1.3 AKA IMDOS.
From other sources, CP/M 1.4 was the first marketed version by
Kildall's Digital Research, followed by 2.0 (noted in the Jan 1980
article) and 2.2 (minor fixes to 2.0) and beyond.

Sorry I missed this clear set of references to Kildall's CP/M computer
model. I worked this out many years ago and did not revisit it very
often since. I should read my own fine Web pages.

Regards, Herb Johnson

Bill Beech (NJ7P)

unread,
Apr 29, 2021, 7:27:07 PM4/29/21
to intel-...@googlegroups.com
Sorry! It was the 1976 Data Book - Page 10-8

Roger Arrick

unread,
Apr 29, 2021, 9:32:35 PM4/29/21
to intel-...@googlegroups.com
On page 10-10 it starts referring to MDS-DOS as ISIS

Page 552 on

http://www.bitsavers.org/components/intel/_dataBooks/1976_Intel_Data_Catalog.pdf

rA

Herb Johnson

unread,
Apr 29, 2021, 11:53:55 PM4/29/21
to intel-...@googlegroups.com, Roger Arrick
>>> Bill Beech (NJ7P) wrote:
>>>> There was the MDS Disk Operating System listed with the MDS-800 in the
>>>> [1976] Data Catalog.  I have never seen the actual product. Might have
>>>> had a name change to ISIS-(*) before it was ever distributed?

> On 4/29/2021 9:32 PM, Roger Arrick wrote:
>> On page 10-10 it starts referring to MDS-DOS as ISIS
>>
>> Page 552 on
>>
>> http://www.bitsavers.org/components/intel/_dataBooks/1976_Intel_Data_Catalog.pdf
>> rA
>>

Sure enough, The MDS-DOS (Disk Operating System of software and
hardware) for the MDS-800, is supported by two "modules" on the MDS-800:
"the channel board" with Intel's 3000-series chipset, and "the interface
board". The word "Multibus" does not appear in this catalog! Only "the
INTELLEC MDS bus".

And the operating software is described as "INTEL SYSTEM IMPLEMENTATION
SUPERVISOR (ISIS)". Included programs are the ISIS text editor, the ISIS
8080 Macro Assembler. The ISIS supervisor program manipulates the files,
provides diskette formatting, and debugging with the Intellec MDS
Monitor program.

So "DOS" is the entire package of hardware and software. "ISIS" is the
diskette operating "supervisor" for the rest of the software.

And the ISIS details described for the MDS-DOS system described, matches
the ISIS described by Ken Bergett's account; and as described for the
ISIS V1.1 by Mark Ogden et al.

The floppy controller hardware for the MDS-DOS product was NOT
explicitly described in Ken's account. Ken described an early hardware
attempt, using a UART-encoding floppy controller scheme.

But Jon Hales found Intel's 1975 AP-7 App Note for a Intel 3000
microcontroller based floppy controller. And the same technology is
described in the MDS-DOS document for the two-board floppy controller.
And, the date is about right.

By visual inspection of the tiny photos of the MDS-DOS channel and
interface boards, I see they visually match these:

http://joe.classiccmp.org/mds-800/fd-chan.jpg
http://joe.classiccmp.org/mds-800/fdc-cdc.jpg

http://www.rogerarrick.com/mds888/ various

What is later called the iSBC 202 (double density) board set.

I could not find an iSBC-201 (211, 212) board (single density) set
image. Bill's site has

http://www.nj7p.org/Manuals/PDFs/Intel/9800349B.pdf

Appendix A has board-layout drawings. And they look the same as the 202
boards. I think we established that the single and M2FM controller cards
are pretty much the same except for the microcodes, ditto for the two
interface boards. But we know the first ISIS and the first MDS-DOS was
single density (right?).

And by the way? The connector at the top of either board? It's *100
pins*, the same pin count as on the AP-7 controller card. The P2
connector is apparently the cross-connect between the boards. So *now*
the mission is: see if the AP-7 BDMC 100 pins correlate to the
iSBC-201/202 100 pin cross-connector.

So we almost have our ducks in a row, on the first ISIS and the first
hardware that ran it in distribution. Everyone contributed, kept
everyone honest, and there's still some things to chase down. I'm
calling it a night.

Regards, Herb

mark.p...@btinternet.com

unread,
Apr 30, 2021, 4:03:58 AM4/30/21
to intel-...@googlegroups.com
All
The control structure parts of ISIS 1.1 that haven't been hand optimised are consistent with being compiled with the FORTRAN cross-compiler, although the compiler mvi l,xx and inr l/h usage has been eliminated, probably because this would have been difficult to manage by hand. Other apps and ISIS.T0 were definitely cross compiled.

What is interesting is that the copyright notice is for 1975, 1976, which hints at an earlier 1975 version.
For reference AP-7 was printed in June 1975 which aligns with these dates

Note ISIS II 2.2 was also cross compiled and has the same copyright dates, suggesting possibly parallel development with the low memory footprint variant. It also suggests a potential earlier version or possibly that ISIS II was based on the same source code as ISIS I, without the copyright being updated.

For reference ISIS II 3.4 is copyright 1978, this has DD support.



Mark




Eric Smith

unread,
Apr 30, 2021, 4:24:14 AM4/30/21
to intel-devsys
On Thu, Apr 29, 2021 at 9:53 PM Herb Johnson <hjoh...@retrotechnology.com> wrote:
Appendix A has board-layout drawings. And they look the same as the 202
boards. I think we established that the single and M2FM controller cards
are pretty much the same except for the microcodes,

Yes. The PROMs are the only difference in the channel boards between the single- and double-density controllers.
(However, the SBC-206 channel board for the CDC Hawk hard disk (MDS-740) is significantly different, even though it also uses the 3001 and 3002 bit slice components.)

Note that on the channel board, the edge connector on the top of the board (opposite from the backplane connectors) is test points for factory test. The pinout of that doesn't match anything else in the universe, and in normal use that connector should not be connected to anything.
 
ditto for the two interface boards.

Unfortunately not. The single-density and double-density (M2FM) interface boards are TOTALLY different. They don't even look similar, except for the edge connectors.

On the interface boards, the edge connector on the top of the board connects to the floppy drive(s). Even MORE unfortunately, the pinouts are DIFFERENT between the single-density and double-density interface boards, so the same cabling can not be used.

Jon Hales

unread,
Apr 30, 2021, 4:36:20 AM4/30/21
to intel-...@googlegroups.com
Eric

Thank you for your comments on the floppy controllers.

Some months ago, I started to trace the test points on the iSBC-202 boards, and made some progress. When I realised that the boards I have represent several PWA and PWB numbers, I lost the will to complete the task.  I don't recall finding test points indicated on the schematic drawings - but I was noting on a copy of the schematics the tracing I had done.

What I have in mind to do now ('now' because I have just collected several systems I had loaned to someone else) is attach a logic analyser to the test points and see what comes out of a 'good' board. In order to compare that with several non-working boards.

Are you aware of documentation of the test points?

Best regards

Jon

--
You received this message because you are subscribed to the Google Groups "intel-devsys" group.
To unsubscribe from this group and stop receiving emails from it, send an email to intel-devsys...@googlegroups.com.

Eric Smith

unread,
Apr 30, 2021, 5:29:51 AM4/30/21
to intel-devsys
On Fri, Apr 30, 2021 at 2:36 AM Jon Hales <jonh...@gmail.com> wrote:
Some months ago, I started to trace the test points on the iSBC-202 boards, and made some progress. When I realised that the boards I have represent several PWA and PWB numbers, I lost the will to complete the task.  I don't recall finding test points indicated on the schematic drawings - but I was noting on a copy of the schematics the tracing I had done.
[...]
Are you aware of documentation of the test points?

The test connector on the channel board is P1, and is shown as connector pins on the schematic.
 
The test connector can be used, among other things, to disable the firmware PROMs and drive the microcode from the test fixture instead.

Jon Hales

unread,
Apr 30, 2021, 5:32:43 AM4/30/21
to intel-...@googlegroups.com
Herb

Thank you for your sleuthing regarding the point that Gary Kildall carried out early development of CP/M on an Intellec 8 (most likely an 8/80).

The UK Computer Conservation Society has a journal named 'Resurrection'. In the following issue:


[That's the Spring 1993 edition] there's an article by Robin Shirley about the origins of the 'Personal Computer' with the text:

 "In 1973, Gary Kildall, a young software consultant at Intel, was fed up with trying to develop the PL/M programming language for microprocessor development systems on paper tape using an ASR 33 teletype, and so begged an ex-10,000-hour-test floppy drive with worn out bearings from the marketing manager at Shugart Associates, a few miles up the road. 

However his attempts at interfacing proved abortive, and it was not until late 1974 that a colleague, John Torode, took an interest in his problem and completed a wire-wrap controller to interface the drive to Gary's Intellec-8 development system. 

Meanwhile Gary had put together a primitive disc operating system for the drive, and in due course (according to Gary) the paper tape was loaded and, to their amazement, the drive went through its initialisation and printed out the system prompt on the first try (legend doesn't record whether it also did so on the second try). [End of quote] 

Now to the point of this message. 


The home page of the website states: "We are one of the most extensive web archives of vintage computers."

The following text is part of the commentary about an Intellec 8 with photos indicating the author(s) have one (Serial No 0188, Type 884, 115v and 60Hz):

"Intel Intellec
Long before the Altair was a gleam in Ed Roberts eye, Intel was quietly producing one of first microcomputers, and in fact, the first complete useable microcomputer, the Intel Intellec. The Intellec was developed by Intel for industry to help in the prototyping of computer systems using their microprocessors. The Intellec contains front panel programming, paper-tape or teletypewriter I/O, 16 slots, PL/M compiler, 16K RAM for a price of $2395.  Intel did not sell a lot of these computers as they didn't know what they had and did very little marketing, as such, few were made.

The Intellec-8 was used by Ed Roberts to help develop the Altair, and if you look closely at the front panel (and at the whole computer) you will noticed many similarities. The Altair front panel has just about all the data/address/control functions and LED's as the Intellec. Arrange in almost the same way. (The Altair used toggle switches instead of the push button switches on the Intellec). The Altair has the same type of motherboard, oriented in the same position as the Intellec. The Altair is slightly larger then the Intellec, due mostly by the slightly larger boards.

The Intellec is a much more robust system, however, more solidly made and reliable. Of course, the Intellec did cost about 5 times the price of the Altair. A nifty hinged hood makes for easy access of the internal boards.

As a part time programming consultant for Intel in 1972, Gary Kildall received an Intel Intellec-8 as partial payment for his services. Fascinated by computers, Gary created the first microcomputer language, PL/M on the Intellec in 1972, when Basic was just a gleam in Bill Gates eyes. PL/M was a simplified version of the mainframe language, PL/I, but significantly more sophisticated then the first BASIC introduced with the Altair.

Later, Gary realized that a cheap storage medium would be essential to making the personal computer practical. So he wrote another program, using the Intellec 8, that interfaced directly with the new floppy disk drives being produced at that time, he called this program CP/M (1974). Shortly afterwards he started his new company, Digital Research. CP/M became the standard programming language for personal computers until IBM announced their PC in 1983.

[End of quote - I think some details in the text above could do with re-wording for clarity/accuracy]

Given your (Herb's) extensive information about S-100 and the history of development of this specification, do you have any information to corroborate the assertion about the extent to which the design of the Altair 8800 was 'influenced' by that of the Intellec/8?

Ed Roberts was talking to Intel about purchasing considerable numbers of the 8080 - so they presumably had a good relationship. I have a memory of seeing a comment that Intel had loaned an Intellec 8/80 to Ed Roberts, but it must have been in a different article.

Note: the search {"ed roberts" altair intellec} produced links to retrotechnology.com - of course.

Postscript:
The practical implication - if this link turns out to be corroborated - is that the Centre for Computing History should place one or both of its Intellec 8/80s alongside the very prominent display of Altair 8800, 8800b and Imsai 8080 machines.

Best regards

Jon

--
You received this message because you are subscribed to the Google Groups "intel-devsys" group.
To unsubscribe from this group and stop receiving emails from it, send an email to intel-devsys...@googlegroups.com.

Herb Johnson

unread,
Apr 30, 2021, 1:22:25 PM4/30/21
to intel-...@googlegroups.com, Jon Hales
On 4/30/2021 5:32 AM, Jon Hales wrote:
> Herb
>
> Thank you for your sleuthing regarding the point that Gary Kildall
> carried out early development of CP/M on an Intellec 8 (most likely an
> 8/80).

Well, I forgot I had already referenced Kildall's 1980 account.

> The UK Computer Conservation Society has a journal named 'Resurrection'.
> [That's the Spring 1993 edition] there's an article by Robin Shirley
> about the origins of the 'Personal Computer' with the text:

Jon, the 1993 text quoted is so much like Kildall's own account in Jan
1980, I am suspicious that Robin Shirley is representing the published
account and is not an independent account. And in fact, her article
references that publication. And her article is a general retrospective
of 15 plus years of microprocessor computing development. So I'd be
inclined to conclude, that her account is a secondary source, and I'd
take a bet that I have identified the primary source.

> Now to the point of this message.
>
> On the Italian website:
> https://www.1000bit.it/storia/vari/intellec/intel_intellec_e.asp
>
> The home page of the website states: "We are one of the most extensive
> web archives of vintage computers."
>
> The following text is part of the commentary about an Intellec 8 with
> photos indicating the author(s) have one (Serial No 0188, Type 884, 115v
> and 60Hz):

> Given your (Herb's) extensive information about S-100 and the history of
> development of this specification, do you have any information to
> corroborate the assertion about the extent to which the design of the
> Altair 8800 was 'influenced' by that of the Intellec/8?
>

Jon, I don't know from memory - I wrote about this stuff years ago.
That's why it's on my Web site, to answer questions like yours.

The article has some speculations, as to whether the Intel Intellec
product, inspired the design of the MITS Altair 8800. There's a claim
quote "The Intellec-8 was used by Ed Roberts to help develop the
Altair". What's the basis of that person's claim? It's not explicitly
stated.

I've also read, as I recall in my head, that Ed Roberts (or rather his
company) owned a Data General NOVA computer. That computer too has a
toggley-blinky front panel. It had a bus. I recall - but can't chase
down right now - some claim that the NOVA inspired the Altair
front-panel design. Factually, there were a few decades of minicomputers
with toggle/lights front panels, before and during the MITS Altair
development.

The article is interesting and of course presents details about the
owner's Intellec 8 and some history about it. But it's a secondary
source, and most likely was written decades after the events in question.

Now to ask Google about my Web site:" mits altair roberts intellec
site:retrotechnology.com"

six hits (your hits will vary), First hit is:

https://retrotechnology.com/herbs_stuff/s_origins.html

My speculations about the MITS Altair and S-100. There's a link to an Ed
Roberts interview apparently from 1995.

http://www.virtualaltair.com/virtualaltair.com/mits0011.asp

Roberts claims his design goals were quote "in principal could do
anything that a general purpose minicomputer of the time could do." end
qoute. That makes my case, that Roberts was following minicomputers of
the 1975 era.

and here's the killer quote from the interview:

> We had a Nova 2 by Data General in the office that we sold time share on, and as a matter of fact that was how we got into building a little terminal. The front panel on an Altair essentially models every switch that was on the Nova 2.

(drops keyboard) (picks up keyboard)

Roberts goes on to say, a front-panel computer today, "would look just
like an Altair, or Nova, PDP-10 or any of those front panel computers."
And he's right to the extent he's defined a front-panel computer,
essentially of the same period.

Thanks, Jon for this question. If this is not on my Web pages, it should
be. I may add it to the s_origins.html Web page.

> Ed Roberts was talking to Intel about purchasing considerable numbers of
> the 8080 - so they presumably had a good relationship. I have a memory
> of seeing a comment that Intel had loaned an Intellec 8/80 to Ed
> Roberts, but it must have been in a different article.

The deal that made the MITS Altair a hit, to my recollections, was this.
In 1975, INtel was, factually, selling 8080 chips for $360 each. yes,
three hundred, sixty, US dollars. Why? they were worried about yield in
their first runs of the processor; they were worried about supply and
demand; they were afraid there was no demand. They wanted to recover
some costs. I don't have quotes for these propositions, but the fact of
the price is established.

Then: Ed Robert's MITS announces the MITS Altair kit, in Jan 1975 - for
$400. It included the 8080 processor. OK? Simple math prevailed: buy a
computer kit for $400 and shipping, or buy a 8080 processor for $360.
What would you choose?

HOw did Roberts achieve that price? He got Intel to provide a wholesale
price to him; at a time presumably when Intel got production up to
speed. I don't have exact quotes or references on this. I think one can
look up such considerations with the keywords I've provided.

And... my doing so (removing my Web site from the search) found:

https://www.landley.net/history/mirror/cpm/article18047.htm

Apparently a memorial piece written after Ed Robert's passing, it says

> The Altair 8800 saved the company. Ed. Roberts had brokered a deal with Intel to buy Intel 8080 chips in bulk for $75/chip (normally they were $360/chip). The cheap CPUs allowed the Altair 8800 to retail for $439 ($621 assembled) at the time when Intel's Intellec-8 Microprocessor Development System, another Intel 8080 based system, sold for $10,000.

I don't know about the $10K price point for an Intellec 8 without disk
drives in 1975. Someone can look that up from preserved Intel documents,
possibly old Computerworld press releases. Wikipedia quotes that price
in its Altair 8800 piece, and cites "Fire in the Valley" for the $75
MITS/Intel 8080 price.

So I don't have more time, right now, to research whether Ed Roberts AKA
MITS owned or accessed an Intel Intellec 8 or 80. But 1) he was likely
aware of it, 2) MITS couldn't afford a several-thousand dollar computer
in 1975, 3) Ed Roberts cites the DG NOVA and every other minicomputer as
inspiration for his front-panel design, bus structure, etc.


> Postscript:
> The practical implication - if this link turns out to be corroborated -
> is that the Centre for Computing History should place one or both of its
> Intellec 8/80s alongside the very prominent display of Altair 8800,
> 8800b and Imsai 8080 machines.
>
> Best regards, Jon

Jon, the Centre for Computing History certainly gets MY vote, for
displaying the Intellec computers with other vintage computers. I'm sure
others in this list would agree. Here's my general case for your
proposition.

The argument for doing so, in my opinion, is that these are the systems
used by Intel in development of software and hardware products, used by
industry and scientists in the era. Intel of course, produced one of the
dominant microprocessor lines of the era, dominant in early personal
computing, and dominant almost to the present day.

Intel Intellec systems were also in use, by developers of hardware and
software for what became "personal computing", which started as
microcomputers bought by small businesses, scientists with smaller
budgets, and small-budget software and hardware developers who created
"personal computing".

It's my view, that looking backwards today, there's a view that
industrial and scientific computing is completely divorced from personal
computers. But that divorce only happened, when personal computing
became an established, self-supporting enterprise. When did that divorce
happen? Perhaps with the 1981 IBM PC? Maybe with the 1984 Apple
Macintosh? Maybe with the slew of home video-gaming computers in the
1980's? Point being, there was no such divorce in the mid 1970's, as
there *was no personal computing* at that time. Intel's work was part of
the background which created personal computing.

Jon: it's up to the CCH, to decide their own narratives and their own
display policies. That's my experience with both museums and with
collectors. They have a point of view and certain objectives, from which
they decide what is of interest to them and what is not, and how to rank
and compare-and-contrast such things. I don't pretend to be any
different. I hope my points and interests have merits, and I try to use
the strongest evidence I can when I express those.

Regards, Herb

Jon Hales

unread,
Apr 30, 2021, 3:57:29 PM4/30/21
to Herb Johnson, intel-...@googlegroups.com
Herb

Thank you for your expert commentary - it's useful to know about the circumstances at that period of computing history.

There was a pleasing 'convergence' between Altair and Intel in the production of a series of the Intel UK and Europe price lists in the late 1970s. One of the Altair 8800b systems at CCH was donated by the person whose Public Relations company had produced Intel's price lists using that same Altair 8800b (to be exact, the Altair was used to generate the data file from which the price list was prepared by a typesetting outfit). Because of the priority for a high degree of accuracy, the Altair was equipped with voice output. Every item that was entered at the terminal keyboard was read back to the operator. Each component listed had three price levels according to quantity.

A comment:
If Intel's UK office considered that they had the capacity to prepare the price list using in-house resources, they would perhaps not have outsourced the preparation of the data file.

On the relative cost of Intellec and Altair:
The $400 Altair kit included a real 8080, but it was implied the user could achieve what they bought the kit for with the 256 bytes of RAM provided.

[Perhaps Intel made some profit on the additional memory buyers then needed to acquire ... ?]

The figure of $2,395 stated on the Italian site for the Intellec 8 might have bought the chassis and some boards. The owner needed an ASR Teletype or Intel paper tape reader and other peripherals. Hence, the cost of a useful system could perhaps be several times that figure.

The 1978 Intel prices in UK didn't list the Intellec 8 system or components. The 1978 price for the MDS-800 provided 16Kb in the standard spec. "Standard software includes a ROM resident monitor, RAM resident 8080/8085 assembler, and text editor".

In 1978, ISIS II was mentioned with Series II systems.

Best regards

Jon

Bill Beech (NJ7P)

unread,
Apr 30, 2021, 8:39:26 PM4/30/21
to intel-...@googlegroups.com
Herb,

Good data!  We are closing in on it all.  So CPM was first a SD system
on an MDS800 then moved onto the S-100 bus?

Bill

On 4/29/2021 8:52 PM, Herb Johnson wrote:
>>>> Bill Beech (NJ7P) wrote:
>>>>> There was the MDS Disk Operating System listed with the MDS-800 in
>>>>> the
>>>>> [1976] Data Catalog.  I have never seen the actual product.
>>>>> Mighthave had a name change to ISIS-(*) before it was ever

Bill Beech (NJ7P)

unread,
Apr 30, 2021, 9:00:24 PM4/30/21
to intel-...@googlegroups.com

Jon,

On 4/30/2021 2:32 AM, Jon Hales wrote:
Herb

Thank you for your sleuthing regarding the point that Gary Kildall carried out early development of CP/M on an Intellec 8 (most likely an 8/80).
He probably had an Intellec 8/8 as well, since I believe he developed a PLM for that machine, first.


The UK Computer Conservation Society has a journal named 'Resurrection'. In the following issue:


[That's the Spring 1993 edition] there's an article by Robin Shirley about the origins of the 'Personal Computer' with the text:

 "In 1973, Gary Kildall, a young software consultant at Intel, was fed up with trying to develop the PL/M programming language for microprocessor development systems on paper tape using an ASR 33 teletype, and so begged an ex-10,000-hour-test floppy drive with worn out bearings from the marketing manager at Shugart Associates, a few miles up the road. 

However his attempts at interfacing proved abortive, and it was not until late 1974 that a colleague, John Torode, took an interest in his problem and completed a wire-wrap controller to interface the drive to Gary's Intellec-8 development system. 

Meanwhile Gary had put together a primitive disc operating system for the drive, and in due course (according to Gary) the paper tape was loaded and, to their amazement, the drive went through its initialisation and printed out the system prompt on the first try (legend doesn't record whether it also did so on the second try). [End of quote] 

Now to the point of this message. 


The home page of the website states: "We are one of the most extensive web archives of vintage computers."

The following text is part of the commentary about an Intellec 8 with photos indicating the author(s) have one (Serial No 0188, Type 884, 115v and 60Hz):

"Intel Intellec
Long before the Altair was a gleam in Ed Roberts eye, Intel was quietly producing one of first microcomputers, and in fact, the first complete useable microcomputer, the Intel Intellec. The Intellec was developed by Intel for industry to help in the prototyping of computer systems using their microprocessors. The Intellec contains front panel programming, paper-tape or teletypewriter I/O, 16 slots, PL/M compiler, 16K RAM for a price of $2395.  Intel did not sell a lot of these computers as they didn't know what they had and did very little marketing, as such, few were made.

The Intellec-8 was used by Ed Roberts to help develop the Altair, and if you look closely at the front panel (and at the whole computer) you will noticed many similarities. The Altair front panel has just about all the data/address/control functions and LED's as the Intellec. Arrange in almost the same way. (The Altair used toggle switches instead of the push button switches on the Intellec). The Altair has the same type of motherboard, oriented in the same position as the Intellec. The Altair is slightly larger then the Intellec, due mostly by the slightly larger boards.
I believe the same 100-pin connector but the buses were completely different.


The Intellec is a much more robust system, however, more solidly made and reliable. Of course, the Intellec did cost about 5 times the price of the Altair. A nifty hinged hood makes for easy access of the internal boards.

As a part time programming consultant for Intel in 1972, Gary Kildall received an Intel Intellec-8 as partial payment for his services. Fascinated by computers, Gary created the first microcomputer language, PL/M on the Intellec in 1972, when Basic was just a gleam in Bill Gates eyes. PL/M was a simplified version of the mainframe language, PL/I, but significantly more sophisticated then the first BASIC introduced with the Altair.

Later, Gary realized that a cheap storage medium would be essential to making the personal computer practical. So he wrote another program, using the Intellec 8, that interfaced directly with the new floppy disk drives being produced at that time, he called this program CP/M (1974). Shortly afterwards he started his new company, Digital Research. CP/M became the standard programming language for personal computers until IBM announced their PC in 1983.

[End of quote - I think some details in the text above could do with re-wording for clarity/accuracy]
Last line of last paragraph - CP/M became the standard OS for PCs....


Given your (Herb's) extensive information about S-100 and the history of development of this specification, do you have any information to corroborate the assertion about the extent to which the design of the Altair 8800 was 'influenced' by that of the Intellec/8?

Ed Roberts was talking to Intel about purchasing considerable numbers of the 8080 - so they presumably had a good relationship. I have a memory of seeing a comment that Intel had loaned an Intellec 8/80 to Ed Roberts, but it must have been in a different article.

Note: the search {"ed roberts" altair intellec} produced links to retrotechnology.com - of course.

Postscript:
The practical implication - if this link turns out to be corroborated - is that the Centre for Computing History should place one or both of its Intellec 8/80s alongside the very prominent display of Altair 8800, 8800b and Imsai 8080 machines.

Best regards

Jon
Bill

On Fri, 30 Apr 2021 at 09:03, mark.pm.ogden via intel-devsys <intel-...@googlegroups.com> wrote:
All
The control structure parts of ISIS 1.1 that haven't  been hand optimised are consistent with being compiled with the FORTRAN cross-compiler, although the compiler mvi l,xx and inr l/h usage has been eliminated, probably because this would have been difficult to manage by hand. Other apps and ISIS.T0 were definitely cross compiled.

What is interesting is that the copyright notice is for 1975, 1976, which hints at an earlier 1975 version.
For reference AP-7 was printed in June 1975 which aligns with these dates

Note ISIS II 2.2 was also cross compiled and has the same copyright dates, suggesting possibly parallel development with the low memory footprint variant. It also suggests a potential earlier version or possibly that ISIS II was based on the same source code as ISIS I, without the copyright being updated.

For reference ISIS II 3.4 is copyright 1978, this has DD support.



Mark




--
You received this message because you are subscribed to the Google Groups "intel-devsys" group.
To unsubscribe from this group and stop receiving emails from it, send an email to intel-devsys...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/intel-devsys/001501d73d97%245efcc9a0%241cf65ce0%24%40btinternet.com.
--
You received this message because you are subscribed to the Google Groups "intel-devsys" group.
To unsubscribe from this group and stop receiving emails from it, send an email to intel-devsys...@googlegroups.com.

Herb Johnson

unread,
Apr 30, 2021, 9:02:35 PM4/30/21
to intel-...@googlegroups.com, Bill Beech (NJ7P)


On 4/30/2021 8:38 PM, Bill Beech (NJ7P) wrote:
> Herb,
>
> Good data!  We are closing in on it all.  So CPM was first a SD system
> on an MDS800 then moved onto the S-100 bus?
>
> Bill

Well, ISIS was developed on the MDS-800. Still sorting out the controller.

Kildall used an intellec 80 and a wired controller from Torode, to get
CP/M running. Some early CP/M licensees used their own hardware. John
Torode took his Kildall controller, and made it and CP/M a product for
his Digital Systems. Torode also made an Altair bus interface card for
his controller - I have a few of both. So that got CP/M 1.2 onto the
Altair bus and then bought by owners of IMSAI's and other S-100 systems.

CP/M 1.3 was IMSAI's IMDOS.
CP/M 1.4 was generally sold by Digital Research.
CP/M 2.0 isolated all I/O and drive stuff to BIOS. CP/M 2.2 was fixes.

regards, Herb

mark.p...@btinternet.com

unread,
May 1, 2021, 5:10:54 AM5/1/21
to intel-...@googlegroups.com
All
I have just been scanning the ISIS internals document and it references other versions of ISIS, specifically
1) In the ISIS: Anatomy of a Real-World Operating System ISIS 1.2 is mentioned (August 1976). There is no reference to a particular variant of ISIS i.e. I or II, suggesting that at this time there was only ISIS. It does mention that ISIS runs on the Intellect MDS and that the MDS disk controller is documented in the MDS-DOS Hardware Reference Manual.
Also mentioned is V1.6 of 16K ISIS

2) There is a description of ISIS II v4, including a table of supported configurations
3) The source fragments, references ISIS II v3, with a hand annotation of 3.0

Mark


-----Original Message-----
From: mark.pm.ogden via intel-devsys <intel-...@googlegroups.com>
--
You received this message because you are subscribed to the Google Groups "intel-devsys" group.
To unsubscribe from this group and stop receiving emails from it, send an email to intel-devsys...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/intel-devsys/001501d73d97%245efcc9a0%241cf65ce0%24%40btinternet.com.

Herb Johnson

unread,
May 1, 2021, 11:15:32 AM5/1/21
to intel-...@googlegroups.com
Please provide a link to that document when it's made available. Thanks
for your diligence. - regards Herb

On 5/1/2021 5:10 AM, mark.pm.ogden via intel-devsys wrote:
> All
> I have just been scanning the ISIS internals document and it references other versions of ISIS, specifically
> 1) In the ISIS: Anatomy of a Real-World Operating System ISIS 1.2 is mentioned (August 1976). There is no reference to a particular variant of ISIS i.e. I or II, suggesting that at this time there was only ISIS. It does mention that ISIS runs on the Intellect MDS and that the MDS disk controller is documented in the MDS-DOS Hardware Reference Manual.
> Also mentioned is V1.6 of 16K ISIS
>
> 2) There is a description of ISIS II v4, including a table of supported configurations
> 3) The source fragments, references ISIS II v3, with a hand annotation of 3.0
>
> Mark
>
>
> -----Original Message-----
> From: mark.pm.ogden via intel-devsys <intel-...@googlegroups.com>
> Sent: 30 April 2021 09:04
> To: intel-...@googlegroups.com
> Subject: RE: intel-devsys ISIS 1.1 and 2.2 - versus ISIS, ISIS-II, versus CP/M
>
> All
> The control structure parts of ISIS 1.1 that haven't been hand optimised are consistent with being compiled with the FORTRAN cross-compiler, although the compiler mvi l,xx and inr l/h usage has been eliminated, probably because this would have been difficult to manage by hand. Other apps and ISIS.T0 were definitely cross compiled.
>
> What is interesting is that the copyright notice is for 1975, 1976, which hints at an earlier 1975 version.
> For reference AP-7 was printed in June 1975 which aligns with these dates
>
> Note ISIS II 2.2 was also cross compiled and has the same copyright dates, suggesting possibly parallel development with the low memory footprint variant. It also suggests a potential earlier version or possibly that ISIS II was based on the same source code as ISIS I, without the copyright being updated.
>
> For reference ISIS II 3.4 is copyright 1978, this has DD support.
>
>
>
> Mark
>
>
>
>

--

mark.p...@btinternet.com

unread,
May 1, 2021, 5:51:33 PM5/1/21
to intel-...@googlegroups.com
Herb
Poor choice of words. I was scanning through the document rather than scanning the document. Isis internals is already widely available
Mark

Regards
Mark

From: intel-...@googlegroups.com <intel-...@googlegroups.com> on behalf of Herb Johnson <hjoh...@retrotechnology.com>
Sent: Saturday, May 1, 2021 4:14:25 PM
To: intel-...@googlegroups.com <intel-...@googlegroups.com>
Subject: Re: intel-devsys ISIS 1.1 and 2.2 - versus ISIS, ISIS-II, versus CP/M
 
--
You received this message because you are subscribed to the Google Groups "intel-devsys" group.
To unsubscribe from this group and stop receiving emails from it, send an email to intel-devsys...@googlegroups.com.
Reply all
Reply to author
Forward
0 new messages