[TUHS] Apollo Domain/OS

[TUHS] Apollo Domain/OS

[Pete Wright]

Apologies that this isn't specifically a Unix specific question but I 
was wondering if anyone had insight in running domain/OS and it's 
relationship to Plan 9 (assuming there is any).

One of my early mentors was a former product person at Apollo in Mass. 
and was nice enough to tell me all sorts of war stories working there.  
I had known about Plan9 at the time, and from what he described to me 
about domain/OS it sounded like there was lots of overlap between the 
two from a high level design perspective at the least.  I've always been 
keen to understand if domain/OS grew out of former Bell Labs folks, or 
how it got started.

As an aside, he gifted me a whole bunch of marketing collateral from 
Apollo (from before the HQ acquisition) that i'd be happy to share if 
there is any historical value in that.  At the time I was a 
video/special effects engineer are was amazed at how beneficial having 
something like domain/OS or Plan9 would have been for us, it felt we 
were basically trying to accomplish a lot of the same goals by duct 
taping a bunch of Irix and Linux systems together.

Cheers,
-pete

-- 
Pete Wright
pete@nomadlogic.org
@nomadlogicLA

Re: [TUHS] Apollo Domain/OS

[Larry McVoy]

On Tue, Jan 21, 2020 at 11:02:38AM -0800, Pete Wright wrote:
> Apologies that this isn't specifically a Unix specific question but I was
> wondering if anyone had insight in running domain/OS and it's relationship
> to Plan 9 (assuming there is any).

It would be news to me if there was overlap.

I used apollos at my first job and hated them.  Diskless and slow as
molasses.  The first thing I did was port the cross compiler to the
Sun file server and did all my work there.

Re: [TUHS] Apollo Domain/OS

[Dan Cross]

[-- Attachment #1: Type: text/plain, Size: 1759 bytes --]

To my knowledge, there was never a link between Plan 9 and Domain/OS.

As I understood it, Domain/OS was more influenced by Multics and maybe the
Pr1me machines than anything else. Plan 9 came from the Unix heritage (and
thus had Multics as a grandparent, in a manner of speaking).

If there was any connection at all, I suspect it was tenuous and more in
the form of an indirect influence, such as an idea someone got from a paper
or something like that.

        - Dan C.


On Tue, Jan 21, 2020 at 2:11 PM Pete Wright <pete@nomadlogic.org> wrote:

> Apologies that this isn't specifically a Unix specific question but I
> was wondering if anyone had insight in running domain/OS and it's
> relationship to Plan 9 (assuming there is any).
>
> One of my early mentors was a former product person at Apollo in Mass.
> and was nice enough to tell me all sorts of war stories working there.
> I had known about Plan9 at the time, and from what he described to me
> about domain/OS it sounded like there was lots of overlap between the
> two from a high level design perspective at the least.  I've always been
> keen to understand if domain/OS grew out of former Bell Labs folks, or
> how it got started.
>
> As an aside, he gifted me a whole bunch of marketing collateral from
> Apollo (from before the HQ acquisition) that i'd be happy to share if
> there is any historical value in that.  At the time I was a
> video/special effects engineer are was amazed at how beneficial having
> something like domain/OS or Plan9 would have been for us, it felt we
> were basically trying to accomplish a lot of the same goals by duct
> taping a bunch of Irix and Linux systems together.
>
> Cheers,
> -pete
>
> --
> Pete Wright
> pete@nomadlogic.org
> @nomadlogicLA
>
>

[-- Attachment #2: Type: text/html, Size: 2308 bytes --]

Re: [TUHS] Apollo Domain/OS

[Chris Hanson]

My understanding is that there’s a direct line from MULTICS to Prime to Apollo, in that Apollo was founded by former Prime folks who took their philosophy with them.

Apollo’s operating system (Aegis, Domain, Domain/OS) had a lot of good and interesting ideas and was quite influential despite the relatively small number of people who used it. A lot of what we take for granted today in distributed computing came via Apollo more than anywhere else, as Apollo users and alumni took what they learned to other systems.

  — Chris

Re: [TUHS] Apollo Domain/OS

[Dennis Boone]

 > My understanding is that there’s a direct line from MULTICS to Prime
 > to Apollo, in that Apollo was founded by former Prime folks who took
 > their philosophy with them.

Among others, and perhaps key, Bill Poduska, Prime's VP Engineering.

De

Re: [TUHS] Apollo Domain/OS

[Clem Cole]

[-- Attachment #1: Type: text/plain, Size: 2566 bytes --]

On Tue, Jan 21, 2020 at 5:03 PM Chris Hanson <cmhanson@eschatologist.net>
wrote:

> My understanding is that there’s a direct line from MULTICS to Prime to
> Apollo,

Yes in some ways...



> in that Apollo was founded by former Prime folks who took their philosophy
> with them.
>
Actually not quite.  MIT/Multics to Honeywell to Pr1me to Apollo (to
Belmont/Stellar)  Bill P and his former TA from MIT (Mike Spourer) actually
wanted to break a little when they did Apollo.  The whole ideas was too
risky for Honeywell, so he created Pr1me.    Apollo was too risky for Pr1me
so he created Apollo.  Stellar was too risky for Apollo so he created
Belmont - a.k.a. Stellar.

[By the way, I spoke to Bill over the holidays.  He's a still the same].

>
> Apollo’s operating system (Aegis, Domain, Domain/OS) had a lot of good and
> interesting ideas and was quite influential

I absolutely agree.

But a number new idea were from an influx of MIT and ex-DEC folks
actually.  And that that terrible sin called the registry that lives on
winders came from Paul (some of us thought it was a bad idea then too BTW).

IMO: The best idea was the typed file system and the ability to run user
code specific with a file type.  That's how IP, TCP, UDP are all
implemented.  Very, very cool.  There is a USENIX paper that describes it
I'll have to dig up the reference.  It's worth reading.  But I have never
seen it implemented again as well.




> A lot of what we take for granted today in distributed computing came via
> Apollo more than anywhere else, as Apollo users and alumni took what they
> learned to other systems.
>
Yes and no.   I agree it was a wonderful intellectual playground for some
very cool ideas.   Some worked pretty well, but not all did.  For example,
as Larry said earlier today, the "twin'ed" or diskless nodes were awful
(replace the S with C for what many of us think about them).  But it got
Sun to make them too and ended up being a great add-in disk business
later.  I refused at Masscomo and ended up losing war, even though the cost
of a WS500 was less than a Sun2 with a disk, people bought Sun's diskless
and then after they discovered they sucked, spent another 6K to buy a desk
system for them (we lost for economics, but I was technically right -
a.k.a. Cole's law of economics vs. sophisticated technology).

Anyway, we (as a community) are better for having that system but other
than the registry, I can think of little actual technology that we continue
to use from Aegis.

[-- Attachment #2: Type: text/html, Size: 4495 bytes --]

Re: [TUHS] Apollo Domain/OS

[Kevin Bowling]

On Tue, Jan 21, 2020 at 3:39 PM Clem Cole <clemc@ccc.com> wrote:
>
>
>
> On Tue, Jan 21, 2020 at 5:03 PM Chris Hanson <cmhanson@eschatologist.net> wrote:
>>
>> My understanding is that there’s a direct line from MULTICS to Prime to Apollo,
>
> Yes in some ways...
>
>
>>
>> in that Apollo was founded by former Prime folks who took their philosophy with them.
>
> Actually not quite.  MIT/Multics to Honeywell to Pr1me to Apollo (to Belmont/Stellar)  Bill P and his former TA from MIT (Mike Spourer) actually wanted to break a little when they did Apollo.  The whole ideas was too risky for Honeywell, so he created Pr1me.    Apollo was too risky for Pr1me so he created Apollo.  Stellar was too risky for Apollo so he created Belmont - a.k.a. Stellar.
>
> [By the way, I spoke to Bill over the holidays.  He's a still the same].

I respect someone that is willing to put their time (and money?) where
their mouth is like that.  Would he be willing to write up some
post-mortems on some of these?  I believe at least late Apollo and
Stellar can qualify as unix?

>>
>>
>> Apollo’s operating system (Aegis, Domain, Domain/OS) had a lot of good and interesting ideas and was quite influential
>
> I absolutely agree.
>
> But a number new idea were from an influx of MIT and ex-DEC folks actually.  And that that terrible sin called the registry that lives on winders came from Paul (some of us thought it was a bad idea then too BTW).
>
> IMO: The best idea was the typed file system and the ability to run user code specific with a file type.  That's how IP, TCP, UDP are all implemented.  Very, very cool.  There is a USENIX paper that describes it I'll have to dig up the reference.  It's worth reading.  But I have never seen it implemented again as well.
>
>
>
>>
>> A lot of what we take for granted today in distributed computing came via Apollo more than anywhere else, as Apollo users and alumni took what they learned to other systems.
>
> Yes and no.   I agree it was a wonderful intellectual playground for some very cool ideas.   Some worked pretty well, but not all did.  For example, as Larry said earlier today, the "twin'ed" or diskless nodes were awful (replace the S with C for what many of us think about them).  But it got Sun to make them too and ended up being a great add-in disk business later.  I refused at Masscomo and ended up losing war, even though the cost of a WS500 was less than a Sun2 with a disk, people bought Sun's diskless and then after they discovered they sucked, spent another 6K to buy a desk system for them (we lost for economics, but I was technically right - a.k.a. Cole's law of economics vs. sophisticated technology).
>
> Anyway, we (as a community) are better for having that system but other than the registry, I can think of little actual technology that we continue to use from Aegis.

Re: [TUHS] Apollo Domain/OS

[Charles H Sauer]

On 1/21/2020 4:36 PM, Clem Cole wrote:

> Anyway, we (as a community) are better for having that system but other 
> than the registry, I can think of little actual technology that we 
> continue to use from Aegis.

don't forget UIDs/UUIDs/GUIDs

-- 
voice: +1.512.784.7526       e-mail: sauer@technologists.com
fax: +1.512.346.5240         Web: https://technologists.com/sauer/
Facebook/Google/Skype/Twitter: CharlesHSauer

Re: [TUHS] Apollo Domain/OS

[Theodore Y. Ts'o]

On Tue, Jan 21, 2020 at 05:36:57PM -0500, Clem Cole wrote:
> > A lot of what we take for granted today in distributed computing came via
> > Apollo more than anywhere else, as Apollo users and alumni took what they
> > learned to other systems.
> 
> Anyway, we (as a community) are better for having that system but other
> than the registry, I can think of little actual technology that we continue
> to use from Aegis.

The OSF/DCE's RPC system came from Apollo, as does the predecessor for
the UUID layout still in use today (RFC 4122).  Paul Leach brought
both to the OSF, and as the Kerberos V5 Tech Lead, I worked with Paul,
and used an early Internet-Draft spec of the OSF UUID and implemented
it in e2fsprogs for labelling ext2/3/4 superblocks, and from there the
infection vector spread to the GNOME project.  When Paul brought both
of those technologies to Microsoft when he went there after OSF went
belly up.

	     		       	       - Ted

Re: [TUHS] Apollo Domain/OS

[Ronald Natalie]

[-- Attachment #1: Type: text/plain, Size: 1196 bytes --]

Apollo gave us the long term use of a DN1000.    I ported our image processing system to it.     The amusing thing was when you fired up an X window manager on the thing, it wrapped all three segments of the little status bar at the bottom with window borders.    The DN10000 docs came with a beautiful 5 volume set that described all the internals.   It was quite detailed for what was there.   What I really needed was the documentation on the graphics system which was in Volume 3.   My set was missing this.   After much digging around Apollo, it turns out that they never wrote volume 3, just assigned the volume number to some yet to be written document.

I learned the hard way not to open the gull wing doors on the thing while someone was using it.

When HP/Apollo merged and they formally pulled the plug on Domain, I was over at the FAA and they showed me the “latest” computer they got in that was running a system to show where all the planes in the airspace system were (think of FlightAware.COM <http://flightaware.com/>, this was a first stab at pulling that data from the ATC computers).    It was a DN4000? running domain.   Always using the latest stuff at the FAA.


[-- Attachment #2: Type: text/html, Size: 1539 bytes --]

Re: [TUHS] Apollo Domain/OS

[Clem Cole]

[-- Attachment #1: Type: text/plain, Size: 1357 bytes --]

Ted point taken. As I was driving home tonight I thought about both DCE and
UUIDs.  The later is clearly important and we do owe the guys in Chelmsford
a great thank you for that one.  But the registry ...  really now.

On Tue, Jan 21, 2020 at 6:00 PM Theodore Y. Ts'o <tytso@mit.edu> wrote:

> On Tue, Jan 21, 2020 at 05:36:57PM -0500, Clem Cole wrote:
> > > A lot of what we take for granted today in distributed computing came
> via
> > > Apollo more than anywhere else, as Apollo users and alumni took what
> they
> > > learned to other systems.
> >
> > Anyway, we (as a community) are better for having that system but other
> > than the registry, I can think of little actual technology that we
> continue
> > to use from Aegis.
>
> The OSF/DCE's RPC system came from Apollo, as does the predecessor for
> the UUID layout still in use today (RFC 4122).  Paul Leach brought
> both to the OSF, and as the Kerberos V5 Tech Lead, I worked with Paul,
> and used an early Internet-Draft spec of the OSF UUID and implemented
> it in e2fsprogs for labelling ext2/3/4 superblocks, and from there the
> infection vector spread to the GNOME project.  When Paul brought both
> of those technologies to Microsoft when he went there after OSF went
> belly up.
>
>                                        - Ted
>
-- 
Sent from a handheld expect more typos than usual

[-- Attachment #2: Type: text/html, Size: 1834 bytes --]

Re: [TUHS] Apollo Domain/OS

[Jon Steinhart]

Chris Hanson writes:
> My understanding is that there’s a direct line from MULTICS to Prime to Apollo,
> in that Apollo was founded by former Prime folks who took their philosophy with them.
>
> Apollo’s operating system (Aegis, Domain, Domain/OS) had a lot of good and
> interesting ideas and was quite influential despite the relatively small
> number of people who used it. A lot of what we take for granted today in
> distributed computing came via Apollo more than anywhere else, as Apollo
> users and alumni took what they learned to other systems.
>
>   — Chris

I stil have some Domain/OS books here if anyone wants 'em.

Re: [TUHS] Apollo Domain/OS

[Richard Salz]

[-- Attachment #1: Type: text/plain, Size: 669 bytes --]

Paul was still at Apollo at the time that OSF/DCE picked Apollo's NCA/NCS
as the RPC system.  I don't know when he left Apollo (or HP/Chelmsford
after they bought it) for Microsoft, but it was before OSF died.  That
latter happened because Sun (Rob Gingell?) insisted OSF stop doing all
development as the price of ending the Unix wars and joining everyone in
opposing Redmond.  We could no longer say OSF stood for Oppose Sun Forever.

I liked that NCA could handle env vars in pathnames, which Locus also had
IIRC.  Domain also had "fat executables" similar to what Apple did when
migrating across CPUs.  And their build and source system was recreated as
ClearCase.

[-- Attachment #2: Type: text/html, Size: 763 bytes --]

Re: [TUHS] Apollo Domain/OS

[Chris Hanson]

On Jan 21, 2020, at 3:53 PM, Ronald Natalie <ron@ronnatalie.com> wrote:
> 
> Apollo gave us the long term use of a DN1000.    I ported our image processing system to it.     The amusing thing was when you fired up an X window manager on the thing, it wrapped all three segments of the little status bar at the bottom with window borders.    The DN10000 docs came with a beautiful 5 volume set that described all the internals.   It was quite detailed for what was there.   What I really needed was the documentation on the graphics system which was in Volume 3.   My set was missing this.   After much digging around Apollo, it turns out that they never wrote volume 3, just assigned the volume number to some yet to be written document.

Also the DN10000 is extremely heavy. Mine’s in a basement and claims are it’s never coming out…

If anyone has DN10000 documentation like Roland describes above, I would love to get my hands on it and ensure it’s preserved. My DN10000 is a long term restoration project, and we could use some docs.

  -- Chris

Re: [TUHS] Apollo Domain/OS

[Lars Brinkhoff]

Chris Hanson wrote:
> My understanding is that there’s a direct line from MULTICS to Prime
> to Apollo, in that Apollo was founded by former Prime folks who took
> their philosophy with them.

Other Multics inspired systems:

M.A.G.I.C. / MagicSix (Interdata minis at MIT Architecture Machine)
Amber (S-1 super computer)

Any recollections?  *COFF* *COFF*

Re: [TUHS] Apollo Domain/OS

[Paul Winalski]

On 1/21/20, Richard Salz <rich.salz@gmail.com> wrote:
>
>And their build and source system was recreated as
> ClearCase.

That was the Domain Software Engineering Environment (DSEE).  One of
my fellow engineers in DEC's Software Development Methods & Tools
group was Dave Leblang,  He had been working on CMS (Code Management
System), DEC's source code versioning system for VAX/VMS.  At Apollo
he developed DSEE, and he was a co-founder of Atria.  CMS took many
ideas from SCCS on the PWB.

Apollo also did some early work in the field of CGI ray-tracing.  They
released the very impressive 4-minute film "Quest: A Long Ray's
Journey into Night" in 1986.  The ray-tracing was done by a network of
100 Apollo workstations and was revolutionary for its time.

-Paul W.

Re: [TUHS] Apollo Domain/OS

[Pete Wright]

On 2020-01-22 08:31, Paul Winalski wrote:
>
> Apollo also did some early work in the field of CGI ray-tracing.  They
> released the very impressive 4-minute film "Quest: A Long Ray's
> Journey into Night" in 1986.  The ray-tracing was done by a network of
> 100 Apollo workstations and was revolutionary for its time.
Thanks Paul, I vaguely remember being told about this effort by my 
mentor which is what sparked my interest.  At the time we were trying to 
optimally use all available CPU resources for rendering and while we did 
have working solutions it required lots of perl and shell.  so the idea 
of CPU resources being available and shared as a foundational concept on 
the network was really interesting to me.

I also seem to remember him telling me about working on the patriot 
missile system, although i am not certain if i am remembering correctly 
that this was something he did at apollo or at another company in the 
boston area.

-pete

-- 
Pete Wright
pete@nomadlogic.org
@nomadlogicLA