[TUHS] phototypesetter C (was nm on Third Edition .o files)

Clem Cole clemc at ccc.com
Sat Dec 24 01:28:05 AEST 2016

On Fri, Dec 23, 2016 at 7:50 AM, Noel Chiappa <jnc at mercury.lcs.mit.edu>

>     > I don't think I ever heard the appellation "phototypesetter C"
>     > before.
> Interesting data point; thanks for passing that along.


>     > Certainly C and the phottypesetter developed independently, though
> in
>     > the same room. But the explanation that they got linked by appearing
> in
>     > the same tape release makes perfect sense.
> I have this vague memory of being told, or reading somewhere, that many of
> the
> changes from 'vanilla V6' C to 'phototypsetter' C were added because they
> were
> needed for that project, hence the name. Alas, I have no idea where I might
> have gotten that from (I had a quick look at a few likely documentary
> sources,
> but no joy).

​My memory also.   I may have gotten that from Dennis or srb in
conversation at some point​ at an early USENIX conference.

> It's quite possible that this was a supposition on someone outside Bell's
> part
> (or perhaps inside Bell, but outside the Unix group), because the two came
> out
> in the same tape.

​Yes - in fact why many of us called it "Typesetter C" because the way we
got that compiler was with the independent Phototypesetter release from
AT&T.   It was all a licensing thing - and the folks in the labs may
otohave realized it -- it was a moniker we used outside of the labs to
differentiate between the different versions of the compiler in the wild.​

I'm trying to remember all of the details/sequences and frankly, I may not
have everything.   But IIRC at CMU I >>believe<< the path we ran 5th, then
6th, got Phototypesetter, then TS, than V7.  i.e. Ken made RK05 images for
us for the 5th edition and that was brought up circia 74/75.   They were
updated to v6 shortly there after via 9track, IIRC, as was the
Phototypsetter code.   tjk brought TS to us on RK05 images on boot-able
9-tracks and V7 was a 1600 bpi tape.

IIRC the CS csv/cret changes were done to a V6 based compiler and then
updated to the Phototypesetter compiler.   When Ted brought us TS ( which
had a different compiler ) we had to hack the compiler per my previous

When V7 was about to be be released we were pumping Ted for info, which he
had some but not all what what we wanted to know.   I remember having a
conversation with some of the BTL folks a @ USENIX (I think it was with
Dennis, but srb may have been part of it also ) asking about the soon to be
delivered 7th edition.  Again, IIRC Klein and I were trying to figure out
what had be changed and how we CS was going to keep going (dvk's domain)
and EE/Mellon Institute (mine own).   The own issue was that CS UNIX
systems were lot of mostly 11/40e's with a couple maybe 45's, but EE,
Mellon, Bio et all were running 11/34's.

> Reading the notes about the upgrades (in particular, "newstuff.nr") makes
> it
> seem like a more likely driver of _some_ of the changes was the Interdata
> port
> (which was also happening at around the same time, if I have the timeline
> correct).

​Right -- I'm trying to line up what we were doing around the same period
to see if I can help date it a little.​

> And of course some might have been driven by general utility (e.g.
> the ability to initialize structures).

 I want to say the Mergenthaler and the APS5 must have been on the
horizon at BTL?  Doug or aps - can either of you offer any enlightenment?
I remember a different conversation with Brian and he saying that when he
was were doing the ditroff work and Ken was reverse engineering the
mergenthaler, Dennis was adding things to make ditroff easier to write.

> It would be interesting to see if anyone remembers why these changes were
> made.

​We should ask Brian,  I would think, he would have been mixed up in all.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://minnie.tuhs.org/pipermail/tuhs/attachments/20161223/5e552afa/attachment.html>

More information about the TUHS mailing list