TUHS January 2024

tuhs@tuhs.org

43 participants
31 discussions

Re: Original print of V7 manual? / My own version of troff

by Mychaela Falconia

Hi again John, > I only meant "professional" insofar as aptitude with graphics is concerned. > I won't accept money; I'm offering my labour out of love for typography, > computer history and its preservation, and of course, the technology that > got Unix the funding it needed to revolutionise computing. In any case, > there's no actual "design" work involved: it's literally just tracing > existing shapes to recreate an existing design. I do stuff like this > <https://github.com/file-icons/icons#why-request-an-icon-cant-i-submit-a-pr> > for *fun*, for crying out loud. Sounds great! If you are indeed serious about trying to recreate the ancient C/A/T character set in PostScript fonts (or some other font format that can be converted into a form that can be downloaded into a genuine PostScript printer), I'll try to find some time to produce the following: 1) A set of C/A/T binary files corresponding to that CSTR #54 manual, as well as BWK's troff tutorial which usually follows right after in book compilations. This step is simply a matter of running the original troff executable (with -t option) on the original source files for these docs - but since I actually run an OS that still includes that original version of troff and you said you don't, it would probably be easier for me to produce and publish these files. 2) A converter tool from C/A/T binary codes to PostScript, using whatever fonts you give it. I'll test it initially using the set of fonts which I developed for my 4.3BSD-Quasijarus pstroff - all characters will be there, all positioning will come from original troff, but it won't look pretty because most PS native font characters don't match those of C/A/T. Then as you progress with your font drawing project, you should be able to substitute your fonts instead of mine, and see how the output improves. > Nice! The more material I have, the merrier. As for the scan that Branden > and I were referring to, I've uploaded a copy to Dropbox Using pdfimages utility with -list option, I compared the image format and resolution in various scans I described in my previous mail, plus this new one you just shared, and concluded that the best quality is contained in these two: http://bitsavers.org/pdf/att/unix/7th_Edition/UNIX_Programmers_Manual_Seven… http://bitsavers.org/pdf/att/unix/7th_Edition/VA-004A_UNIX_Programmers_Manu… Here are extracted PNG images of just the relevant page from both PDFs: https://www.freecalypso.org/members/falcon/troff/cstr54-fontpage-sri.png https://www.freecalypso.org/members/falcon/troff/cstr54-fontpage-ucb.png Each PNG is a lossless extract from the corresponding PDF, made with pdfimages utility. Each image is described as being 600x600 DPI in PDF metadata, and the print is said to be in 12 point - numbers for converting from pixels to .001m units in font reconstruction... M~

1 year, 2 months

Re: Original print of V7 manual? / My own version of troff

by G. Branden Robinson

Hi John, At 2024-01-18T00:43:41+1100, John Gardner wrote: > I'm a professional graphic designer with access to commercial typeface > authoring software. Send me the highest-quality and most comprehensive > scans of a C/A/T-printed document, and I'll get to work. If you don't have my scan of CSTR #54 (1976), which helpfully dumps all of the glyphs in the faces used by the Bell Labs CSRC C/A/T-4, let me know and I'll send it along. I won't vouch for its high quality but it should be comprehensive with respect to coverage. > Thanks for reminding me, Branden. :) I've yet to get V7 Unix working > with the latest release of SimH, Let me know in private mail where you got stuck. Maybe I can help. > I'm still up for this, assuming you've not already started. No, I haven't--perhaps because I am an Ada fanboy, the prospect of coding in pre-standard C and its mission to turn anything that can be lexically analyzed into _some_ sequence of machine instructions has not stoked my excitement. (Which isn't to say that one _can't_ write safe code using K&R C; my fear is that having to remember all of the things the compiler won't do for you would overwhelm the task at hand. Too bad Unix V7 didn't have Perl, since this is basically a text transformation problem.) Regards, Branden

1 year, 2 months

PCS kernel sources

by Warren Toomey

All, I got this e-mail from Holger a while back. Somehow it went into a folder and has lurked unseen for way too long. Does anybody know any more about PCS Unix and, if so, where should I place the code that Holger has donated into the Unix Archive? Many thanks, Warren ----- Forwarded message from hveit01(a)web.de ----- From: hveit01(a)web.de To: wkt(a)tuhs.org Subject: PCS kernel sources Hi Warren, Some time ago I subscribed to the tuhs mailing list because of my interests in Unix. I have been working on regenerating the ancient PCS unix (see more details in the README file in the attached archive). Now it is in a state to publish the results. You may decide to put this up on the TUHS website for reference. PCS UNIX (dubbed MINUX) is special in the way that it is derived from an SVR3.2 UNIX with the Newcastle connection integrated. The Newcastle connection is an early attempt for multicomputer networking; it provides a shared file system over the network, similar to the later NFS. To my knowledge, it is the first time that source for it are described (beyond some publicly availablereasearch paper); I haven't yet managed to find the original sources of this. Regards Holger Veit ----- End forwarded message -----

1 year, 2 months

Original print of V7 manual? / My own version of troff

by tuhs＠cuzuco.com

No idea what COFF is, but in the early 1980s, two non-troff options on the software side were - 1) TeX. From Donald Knuth, which means tau epsilon chi, pronounced tech not tex. The urban legend was upon seeing an inital copy of one of his books sometime in the 1970s, he yelled "blech!" and decided that if you wanted your documents to look right, you need to do be able to it yourself, and TeX rhymes with blech. 2) Scribe. From Brian Reid, of Carnegie-Mellon See http://www.columbia.edu/cu/computinghistory/scribe.pdf -Brian Clem Cole clemc at ccc.com wrtoe: > Not really UNIX -- so I'm BCC TUHS and moving to COFF > > On Tue, Jan 9, 2024 at 12:19b/PM segaloco via TUHS <tuhs at tuhs.org> wrote: > > > On the subject of troff origins, in a world where troff didn't exist, and > > one purchases a C/A/T, what was the general approach to actually using the > > thing? Was there some sort of datasheet the vendor supplied that the end > > user would have to program a driver around, or was there any sort of > > example code or other materials provided to give folks a leg up on using > > their new, expensive instrument? Did they have any "packaged bundles" for > > users of prominent systems such as 360/370 OSs or say one of the DEC OSs? > > > Basically, the phototypesetter part was turnkey with a built-in > minicomputer with a paper tape unit, later a micro and a floppy disk as a > cost reduction. The preparation for the typesetter was often done > independently, but often the vendor offered some system to prepare the PPT > or Floppy. Different typesetter vendors targeted different parts of the > market, from small local independent newspapers (such as the one my sister > and her husband owned and ran in North Andover MA for many years), to > systems that Globe or the Times might. Similarly, books and magazines > might have different systems (IIRC the APS-5 was originally targeted for > large book publishers). This was all referred to as the 'pre-press' > industry and there were lots of players in different parts. > > Large firms that produced documentation, such as DEC, AT&T *et al*., and > even some universities, might own their own gear, or they might send it out > to be set. > > The software varied greatly, depending on the target customer. For > instance, by the early 80s, the Boston Globe's input system was still > terrible - even though the computers had gotten better. I had a couple of > friends working there, and they used to b*tch about it. But big newspapers > (and I expect many other large publishers) were often heavy union shops on > the back end (layout and presses), so the editors just wanted to set strips > of "column wide" text as the layout was manual. I've forgotten the name of > the vendor of the typesetter they used, but it was one of the larger firms > -- IIRC, it had a DG Nova in it. My sister used CompuGraphic Gear, which > was based on 8085's. She had two custom editing stations and the > typesetter itself (it sucked). The whole system was under $35K in > late-1970s money - but targeted to small newspapers like hers. In the > mid-1908s, I got her a Masscomp at a reduced price and put 6 Wyse-75 > terminals on it, so she could have her folks edit their stories with vi, > run spell, and some of the other UNIX tools. I then reverse-engineered the > floppy enough to split out the format she wanted for her stories -- she > used a manual layout scheme. She still has to use the custom stuff for > headlines and some other parts, but it was a load faster and more parallel > (for instance, we wrote an awk script to generate the School Lunch menus, > which they published each week). >

1 year, 2 months

V10 monk program

by arnold＠skeeve.com

Hi All. V10 had a program called "monk" which was a "document compiler". It produced troff and know how to run eqn, tbl, and pic and I'm guessing also refer. It seems to have been inspired by Scribe. The V10 files from Dan Cross have the device independent troff output for the paper that describes monk. G. Branden Robinson was kind enough to turn it into PostScript for me; his story is below, forwarded by permission. I'm also attaching the PostScript file. I'm curious, did this see a lot of use inside Research or outside of it? At first glance, it looks like the kind of thing that might have caught on, especially for people who weren't already used to troff. Thanks, Arnold > Date: Wed, 10 Jan 2024 12:25:53 -0600 > From: "G. Branden Robinson" <g.branden.robinson(a)gmail.com> > To: Aharon Robbins <arnold(a)skeeve.com> > Subject: Re: v10 ditroff output file > > Hi Arnold, > > At 2024-01-09T08:50:28+0200, Aharon Robbins wrote: > > Hi. > > > > The file of interest is attached. It's from vol2/monk in Dan Cross's > > V10 sources in the Distributions/Research directory from TUHS. > > > > If you can get postscript out of it somehow, > > Bad news and good news. > > ...unfortunately there was too much impedance mismatch with groff/grops. > > Some font names differ but that's not a big deal. (Also, today I > learned: the troff that generated this reloaded all the fonts at each > new page.) The troff(1) that generated this also attempted vertical > motion before starting the first page. That also wasn't a big deal. I > thought I was going to be able to text-edit my way to a solution...but > then... > > grops really wants the device resolution to be 72,000 dpi, not 720, and > we'd have to write a rescaling feature to support that. Just editing > the output file won't do because the file uses Kernighan's optimized, > anonymous output command pervasively. > > groff_out(5): > ddc Move right dd (exactly two decimal digits) basic units u, > then print glyph with single‐letter name c. > > In groff, arbitrary syntactical space around and within this > command is allowed to be added. Only when a preceding > command on the same line ends with an argument of variable > length a separating space is obligatory. In classical > troff, large clusters of these and other commands were used, > mostly without spaces; this made such output almost > unreadable. > > So all these two-digit motions would need to become five-digit motions > or all the glyphs would pile up on each other. (And that's exactly what > I happened after doing a couple of fixups and throwing gxditview(1) at > it.) > > Out of curiosity, I tried DWB 3.3 troff. > > It did well, until the fourth page, when it fell over and produced > PostScript that made Ghostscript very angry. > > So I tried Heirloom Doctools troff. > > 20 pages of goodness. > > But be advised: some sort of extension was used to embed other > PostScript files: > > ./bin/dpost: can't open picture file samples/tailor.ps (line 1493) (page 2) > ./bin/dpost: can't open picture file samples/memo.ps (line 1749) (page 3) > ./bin/dpost: can't open picture file samples/tmbody.ps (line 2151) (page 4) > ./bin/dpost: can't open picture file samples/tmcs.ps (line 2694) (page 5) > > So I went to minnie.tuhs.org to see if they were there... > > ...and they were. > > So here you go. Renders without errors (though Heirloom is nowhere near > as validation-happy as groff), and the output seems plausible. > > > I'll really appreciate it. :-) > > Enjoy! > > Regards, > Branden

1 year, 2 months

Re: Original print of V7 manual?

by Douglas McIlroy

> The <C>omputerphile Youtube channel did a video about 10 years ago about > "The Great 202 jailbreak:" https://www.youtube.com/watch?v=CVxeuwlvf8w It may be superfluous in this forum, but one should note that the video's characterization of Brian Kernighan as the father of typesetting at Bell Labs does great disservice to Joe Ossanna, who single-handedly brought the first phototypesetter to the labs, subjected it to computer control, and wrote troff (which lives on 50 years later) to drive it. In passing, the video denigrates the C/A/T because it had a fixed font repertoire and no general graphic capability. But without the antecedent of C/A/T and troff, the famous Linotron summer-vacation project would never have been undertaken. Doug

1 year, 2 months

XID register

by ron minnich

SunRPC, among other protocols, needs transaction IDs (XIDs) to distinguish RPCs.For SunRPC, it's important that XIDs not be reused (not for all protocols; 9p has no such requirement). Stateless protocols like NFS and reused XIDs can get messy. There is a vague, 30 year old memory, I have, that at some point SPARC got a time register, or some other register, that always provided a different answer each time it was read, even if read back to back, in part to enable creation of non-reused XIDs. Note that things like the TSC or RISC-V MTIME register make no such guarantee. I am pretty sure someone here can fill me in, or tell me I'm wrong, about my SPARC memory. thanks

1 year, 2 months

Re: Original print of V7 manual?

by Brian Walden

Mychaela Falconia falcon at freecalypso.org wrote: . . . > > It was made under Solaris 2.6, on an Ultra 2 ("Pulsar"), using the troff, tbl, > > eqn, pic, refer and macros as supplied by Sun at that time, and NOT any GNU > > ones. Why? These were the versions written by AT&T that Sun got directly from > > them during their SVR4 collaboration. I used the PostScript output option to > > troff (which obviously did not exist in 1979). > > You did the right thing: the version you used certainly feels much more > "right" than anything from GNU. I was just tryting to use the tool that would give the path of least resistance for that troff source. Even between flavors of UNIX in the 1980s, there were issues getting correctly formatted output bewteen Documenter's Workbench (DWB) and UCB. > > That code to produce PostScript > > outout, had a high probability of being written by the graphics group run by > > Nils-Peter Nelson in Russ Archer's Murray Hill Computer Center (department > > 45268). > > So it is a different ditroff-to-PS chain than psdit from Adobe > Transcript? I am not too familiar with the latter, as I ended up > writing my own troff (derived from V7 version, just published) that > emits PS directly, but it is my understanding that Back In The Day > most people used psdit for this type of workflow. The DWB way of troff to PostScript is -- $ pic file | tbl | eqn | troff -mm -Tpost | dpost >file.ps $ # if you want to print it near the "bird cage" printer, near a famous stair case in MH $ i10send -dbirdie -lpost file.ps $ # which would eventually call postio for you $ postio -l /dev/tty?? file.ps As this is pre-ethernet time, QMS printers are connected via RS-232 serial lines and postio does the communication to the printer. You can find dpost at https://www.tuhs.org/cgi-bin/utree.pl?file=OpenSolaris_b135/cmd/lp/filter/p… (or at https://github.com/n-t-roff/DWB3.3/blob/master/postscript/dpost/dpost.c ) Looking at the last few lines of https://www.tuhs.org/cgi-bin/utree.pl?file=OpenSolaris_b135/cmd/lp/filter/p… it is signed, Richard Drechsler MH 2F-241 x7442 mhuxa!drexler Which is that group I mentioned. Rich wrote dpost for sure, also if you look at the last person thanked in the Preface of The C programming Language, Second Edition (1988) -- Rich Drechsler helped greatly with typesetting. On a sad side note, Carmela L'Hommedieu, I was going to say "recently," but it's been almost four years now, who also worked in that group, has passed https://www.tributearchive.com/obituaries/10822663/Carmela-Scrocca-LHommedi… > > > I did have a volume 2A that also had the correct 7th Edition C Reference > > Manual > > in it. The one you get in my 1988 PDF is from the 6th Edition, notice it is > > the old =+ syntax and not the += one. Dennis said that not even Lucent could > > provide that as a free PDF, as it was a published book by Prentice-Hall. I > > was asked to destroy all PDFs that had that version in it. > > Ouch, until you pointed it out in this ML post, I hadn't even noticed > that the C Reference Manual doc is "wrong" in your PDF version! But > here comes the really important question: if you once had a PDF reprint > with the "right" version of this doc, where did you get the troff > source for it? This is the source that was actually censored from the > V7 tape: > > https://www.tuhs.org/cgi-bin/utree.pl?file=V7/usr/doc/cman Yes that is the missing C Reference Manual. I was gifted the troff source for it, and unfortunately I do not have that gifted copy anymore. > > I don't have this problem for my 4.3BSD reprint: the source for 4.3BSD > version of this doc is included on the tape; the corresponding SCCS > log begins with "document received from AT&T", checked in on 86/05/14, > and then revised by BSD people into what they wanted printed in their > version of the manual. But if someone wishes to do a *proper* reprint > of the V7 manual (or 4.2BSD, where this doc and many others were > literally unchanged duplications from V7 master at the plate level), > we need the troff source for the V7 version of this doc. > > If this source is totally lost, we as in community can probably do an > OCR from a surviving scan (for example, the one in 4.2BSD PSD book) > and then painstakingly produce a new troff source that would format > into an exact replica - but if there is a leaked copy of the original > source somewhere, it would certainly make our job way easier. > > > Larry McVoy asked me for my modified files to make the PDFs too, in 1999 or > > 2000, for bitkeeper or bitsavers. But since I was not allowed to share them > > and I had moved companies, I had lost them. I thought I had saved a copy but > > I could no longer find it. I asked Dennis if he still had them, he did not. > > This work is truly lost. > > Aside from the unresolved issue of "cman" document, we as in community > can produce an even better work if we so wish. I am deferring a more > detailed discussion until I put out my 4.3BSD PS reprint, so I can > point to it as a reference for how I like to do things, and maybe by > then we'll have some clarity on what happened to V7 "cman" troff source. You will need to check on the legality of that. It is missing because it was published as Appendix A of the first edition of The C Programming Language in 1978 by Prentice-Hall, which means they (not Bell Labs, nor successor compaies, AT&T, Lucent, Alcatel, Nokia) contractually own the rights to it for some period of time. I you read Dennis' old home page at https://www.bell-labs.com/usr/dmr/www/ you'll see this verbage -- "The version of the C Reference Manual Postscript (250KB) or PDF, (79K) that came with 6th Edition Unix (May 1975), in the second volume entitled ``Documents for Use With the Unix Time-sharing System''. For completeness, there are also versions of Kernighan's tutorial on C, in Postscript or PDF format. There is also a slightly earlier (January 1974) version of the C manual, in the form of an uninterpreted PDF scan of a Bell Labs Technical Memorandum, visible here, if you can accommodate 1.9MB. No updated version of this manual was distributed with most machine readable versions of the 7th Edition, since the first edition of the `white book' K&R was published about the same time. The tutorial was greatly expanded into the bulk of the book, and the manual became the book's Appendix A. However, it turns out that the paper copies of the 7th Edition manual that we printed locally include not only what became Appendix A of K&R 1, but also a page entitled "Recent Changes to C", and I retyped this. I haven't been able to track down the contemporary machine-readable version (it's possible that some tapes were produced that included it). This is available in PostScript or PDF format." As we know from the recent public domaining of Mickey Mouse, copyright is retained 70 years past the date of death of the (last surviving) author. So if Brian Kernighan lives to the ripe old age of 101, this work cannot be used without permisson until 2113, unless the rights holders place it into the public domian before hand. Since the 1st edition is out of print, it's rights *may* have reverted back, but to which companies? Probabaly Nokia and AT&T jointy. But there is no way to know if you can use it, without an official notice of such. -Brian

1 year, 2 months

Re: Original print of V7 manual?

by Brian Walden

Arnold, Thank you, it's nice to have one's work appreciated. And I know, you were doing exactly what I was doing, trying to make it more accessible to more people. And Dennis, being who he was, always gave credit where credit was do. There's nothing else he could or would have done. And like I said, it was long ago and has not bothered me in a very long time. Thanks for your continued dedication to gawk. Awk still just flows out of my fingers without even needing to think much or at all. Professionally, I have programmed in python for years, and have never gotten to the same level intimacy I have with awk. -Brian arnold at skeeve.com wrote: > Thanks for this history Brian. > > It was a long time ago, but I think all I did was figure out how > to turn the PDF back into postscript, since I had a postscript printer > at the time and it was easier for me to print postscript. > > I sent the files to Dennis _only_ with the thought that they might be > useful to other people, and certainly with no intent to steal any credit. > > Your files were great; I printed out hardcopy at the time and > still have them on a shelf in my basement. > > Thanks! > > Arnold

1 year, 2 months

SunOS 4.1.3 and 4.1.4 Release Manuals

by Kevin Bowling

Would anyone happen to have a link to these? I am curious to read the changes in the final SunOS 4.1 versions and see how it yielded to Solaris during that period. Here is the 4.1.2 version for reference https://vtda.org/docs/computing/Sun/software/800-6645-10_SunOS4.1.2ReleaseM…. Regards, Kevin

1 year, 2 months

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

1996

1995

1994

1993

1992

1991

1990

TUHS January 2024