The late Luiz Barroso used to say that the x86 ISA is the most successful ISA ever designed, by sheer numbers and longevity. I guess time will tell, but so far, I think he has been right. The instruction set is still out there, likely with more binaries compiled for one version or another than for any other ISA. Intel's recent attempts to remove backward compatibility for 16-bit Real Mode (X86-S, which failed) are a case in point. On Sun, Jun 14, 2026 at 9:30 AM Johan Helsingius via cctalk < cctalk(a)classiccmp.org&gt; wrote:

That's fair, though I think in the context Luiz used to make this point, it remains remarkable. He used to argue that the success of an architecture (be it software or hardware) is only loosely related to how well thought out and "elegant" it is, and is much more influenced by successfully building a long-lived ecosystem (which, arguably, the S/360 example demonstrates as well). Despite x86 being a stop-gap hodgepodge, it succeeded against much more elegant and modern designs because of extraneous factors unrelated to the design itself. He made that point in writing (to some extent, given the limited context of a one-pager) in his "Innovate Within" manifesto: https://datacenter-book.org/essays/innovate-within On Sun, Jun 14, 2026 at 11:17 AM Paul Koning &lt;paulkoning(a)comcast.net&gt; wrote:

On Sun, 14 Jun 2026, Johan Helsingius via cctalk wrote: The sole advantage of Intel's approach, is by making each one a hodgepodge of kludges from its predecessor, rather that a thorough redesign from scratch, is that less work was required to modify programs for the next one. Obviously starting over with a complete redesign makes for a much better final product, but longer delays to port software. 8086/8088 binary required VERY few changes to run up through Pentium. Micropro (later Wordstar International) ported Wordstar from 8080 to 8088 in about two weeks! It then took them much longer than that to revise the manuals. (Would using a different word processor have sped that up? :-) -- Grumpy Ol' Fred cisin(a)xenosoft.com

The 80XXX series was far from the best.  It was successful solely because the largest computer company in the world chose it. The saying back then was "You can never go wrong buying IBM" The original PC was really not a very good machine, especially not for more then $5,000.  It was three programs that made is useful. Wordstar, Visicalc (later Lotus 123) and DBASE II. IBM chose the 8088 because the 68000 was nearly equivalent to the 360 at the time and the NS32000 wasn't in silicon yet. On 6/14/2026 8:00 PM, Murray McCullough via cctalk wrote:

The uVAX-II CPU (KA630) ran at 5MHz (200ns machine cycle). On Sun, Jun 14, 2026 at 7:03 PM Jon Elson via cctalk &lt;cctalk(a)classiccmp.org&gt; wrote:

X86 is a zombie. It's dead but it won't lie down. On Tue, Jun 16, 2026 at 2:53 PM Ken Seefried via cctalk < cctalk(a)classiccmp.org&gt; wrote:

On Tue, 16 Jun 2026, ben via cctalk wrote: Is that the problem with the modern machines also?

On Tue, 16 Jun 2026, Ken Seefried wrote: Libraries are dead books are dead USA manufacturing command line mainframes serial/parallel composite video Floppy disks CRTs wired headphones manual transmission soldering (especially lead solder) DIY hand sanding hand drill non-power tools non cordless power tools shoe repair ANY repair of anything ("just buy a new one") using anything for as long as it works sewing clothesline non-cellular, landline phones music/movies and distribution other than streaming Over the air TV (except: "amazing new device lets you get TV without cable!" ("digital" rabbit ears)) non-microwave cooking cooking anything from scratch: spaghetti sauce, guacamole, chili, soup, cookies, . . . my entire world?

On Tue, 16 Jun 2026, Mike Katz via cctalk wrote: chemical photography, both B&W and color. 4x5 and 8x10 sheet film are still available, but where? Kodak has discontinued making photographic print paper. Kodak has discontinued roll film, other than 35mm and 120. Kodak has discontinued Kodachrome!! ("Don't take my kodachrome away.") Polaroid

Someone said assembler was dead, I have to disagree.  I still program in 6809, 68000, ARM, HP-41 and HP71 assembler from time to time.  I've used ARM assembler in my work recently. On 6/16/2026 8:39 PM, Ken Seefried via cctalk wrote:

I would gladly pay somebody to maintain code in each of these languages and more besides if it would get the “$x is dead!” forum trolls to sit on their hands for twenty minutes. We could start a foundation or something. ok bear.

On Tuesday, June 16, 2026 at 21:09, Van Snyder via cctalk wrote: There is. I've been using SNOBOL periodically for the last 25 years, having written some two dozen string-processing programs (e.g., an e-mail reformatter and archiver, and a control-file generator for Eric Smith's "tumble" program to process TIFF scans into PDF manuals). Last one updated about a year ago. Largest program about 2600 lines including comments. I'm currently using a SPITBOL compiler, originally from Catspaw: http://www.snobol4.com/ ...now available from GitHub: https://github.com/spitbol Works like a champ! -- Dave

I cannot help but find this discussion of "dead languages" fascinating, particularly in the context of a Classic Computer mailing list. A language is not dead simply because somebody has not heard of it. Nor is it dead because the latest Python, JavaScript, or Ruby programmer has never encountered it. By that measure, half of computing history would vanish every few years. The whole reason groups like this exist is because there are still people actively using, studying, teaching, repairing, and preserving these technologies. Around the world there are enthusiasts rebuilding Apollo guidance computer rope memory, reproducing DSKYs, restoring S100 and STD bus systems, maintaining FORTH environments, repairing PDPs, and keeping countless other pieces of computing history operational. Many of these systems are not museum exhibits. They are running, being developed, and teaching new generations how computing worked before abstraction layers hid all the interesting parts. Perhaps a better definition of a dead language is one for which there is nobody left who can read it, write it, teach it, or care about it. By that standard, many of the languages being discussed here are very much alive. They may not be fashionable, they may not appear in job advertisements, and they may not be taught at university, but they are still being used by communities that value them. If anything, the existence of this mailing list is evidence against the idea that these languages are dead. We are quite literally the people keeping the knowledge alive and passing it on to those who come after us. Dead? Maybe. Mostly dead? Certainly not. Kindest regards, Doug Jackson em: doug(a)doughq.com ph: 0414 986878 On Wed, 17 Jun 2026 at 15:13, Fred Cisin via cctalk &lt;cctalk(a)classiccmp.org&gt; wrote:

Jun 17, 2026, 01:29 by cctalk(a)classiccmp.org: Neither.  They're under-educated at that point in time.  Just as all of us have been at one time or another. -- The Doctor [412/724/301/703/415/510] WWW: https://drwho.virtadpt.net/ My code, of course, begins with 4GH.

On 6/17/2026 9:48 AM, Paul Koning via cctalk wrote: I have stood in on conversations in Latin. I know a lot of people who can and do still converse in Latin. While my ability to converse has declined I can still read Latin easily and I have at least one friend whose Christmas greeting inside the card I still write in Latin. bill

Thanks for the clarification. I'm an amateur, though an interested one. My comment about Sanskrit was because I saw it in a list of languages spoken (as first language) in India, by a small population admittedly, 10k or so, but still, I had not expected that. That makes sense, and it reminds me of the revival of Hebrew. And the attempted revival of Cornish, or Manx, I forget. But doesn't that definition exclude pidgin languages (distinguished from creole by not being first languages)? paul

Yes, that's what I thought too. And then I found Sanskrit on a list of languages spoken in India today. https://en.wikipedia.org/wiki/List_of_languages_by_number_of_native_speaker… -- about 25k speakers of Sanskrit. Blew my mind. paul

Cos4? On Thu, Jun 18, 2026, 11:53 a.m. Bill Degnan via cctalk < cctalk(a)classiccmp.org&gt; wrote:

On Wed, 17 Jun 2026, Doug Jackson wrote: There is an enormous distinction between actually being dead, VS being CALLED "dead" by those who are simply ignorant of it. THAT is what I was saying the message that you quoted.

On 2026-06-16 11:13 p.m., Fred Cisin via cctalk wrote: I say it is dead when you can't find a copy to run. "A book on C". 1st ed. had a Rat C (small C), as a appendix in the Back. I found a copy a few years back,(but lost it with a disk crash). Now I can't find it at all online.

On 2026-06-17 3:04 p.m., The Doctor via cctalk wrote: Only the 1st one. Since I can't print it (pdf) , I can't type it in. Ben.

On 6/17/2026 1:13 AM, Fred Cisin via cctalk wrote: COBOL is "dead" because academia said it was. The price you pay for not letting the ivory tower masters drive the bus. bill

On Wed, 2026-06-17 at 10:59 -0700, Fred Cisin via cctalk wrote: While not directly affiliated with UC Berkeley, Damian Rouson, who works at the Lawrence Berkeley Lab, which is operated by UC Berkeley, is one of the lead developers of the Coarray system of SPMD parallel programming in Fortran (this was originally started by Bob Numrich at Cray, proposed for the Fortran standard in 1998 by John Reid, added as a project in 2008, and is now part of the standard). He is head of the Computer Languages and Systems Software Group at LBL, teaches classes on campus and UC Berkeley, adjunct faculty at UC San Diego, with former positions at Stanford, CUNY, and the University of Maryland.

In the early 70's, Mary Ann (Mark) Horton wrote an emulator in HP 2000C Basic named "Hortran" at Benson high and OMSI. I would suspect, no copy exists, unless Mary Ann has a copy. Dead? On Wed, Jun 17, 2026 at 11:17 AM Fred Cisin via cctalk &lt;cctalk(a)classiccmp.org&gt; wrote:

On Wed, 2026-06-17 at 10:50 -0800, Mychaela Falconia via cctalk wrote: The Summer Higgins character on the Yellowstone series asked Rip Wheeler (or maybe John Dutton) whether he killed rattlesnakes. He replied "Heavens no! Do you know how many rats and mice they eat?" I never molest snakes on my property. Snakes are terrified of people and will always try to escape if you don't block what they perceive as their way out.

dead (no longer used in production) vs lost (media dors not exist even if you wanted to revive it) Bill On Wed, Jun 17, 2026 at 1:13 AM Fred Cisin via cctalk &lt;cctalk(a)classiccmp.org&gt; wrote:

On Wed, Jun 17, 2026 at 11:57 AM Mark Green via cctalk < cctalk(a)classiccmp.org&gt; wrote: Columbus State University (Georgia, USA) has a CS degree specialization in mainframes, which includes COBOL, JCL and IBM assembly. I understand IBM and TSYS (now part of FIS) set it up. KJ

I'm sorry but I beg to differ with your, it is entirely possible to write the code generator for a compiler in a high level language. Most C-compilers are written that way. The initial after reset code may need to be written in assembler to properly setup some registers prior to the run time system for the high level language (stack, interrupt masks, cache initialization, etc.). On 6/17/2026 12:55 PM, Paul Koning via cctalk wrote:

You missed the point. The compiler backend is emitting assembly code, so you better have a very good understanding of the underlying hardware and the assembly language for the machine to successfully write a compiler backend for the architecture in question. Tom On Thu, 18 June 2026, 2:15 am Mike Katz via cctalk, &lt;cctalk(a)classiccmp.org&gt; wrote:

On 6/18/2026 1:13 PM, Van Snyder via cctalk wrote: Interesting. I did Autocoder but never thought of it as assembler. Programming the 1401 in assembler was when you took the output from an Autocder deck and hand optimized it to a single card. :-) As for translating from one language to another, that was pretty much standard practice for a long time. GNAT Ada produced C. P2C converted Pascal to C. F2C the same for Fortran. Early precursors of GNU COBOL did the same. Assembler knowledge is not necessarily needed to write a compiler if you have another language to function as the intermediate stage of the compiler. bill

True, but that approach has significant drawbacks and from where I sit it appears more like a technique that was used for a while as a shortcut to implementing a language. All the examples you gave are now obsolete, at least in the GCC system, replaced by full front ends that tie into the code generation machinery. A reason why this matters is debug info; having C debug info for your Fortran program, let alone COBOL or Pascal, is painful. Another reason is semantic mismatches. C doesn't like nested function definitions (though GCC sort of allows them as an extension) while other languages (Algol for sure, and Pascal also I think) consider them perfectly good things to do. paul

On Thu, Jun 18, 2026 at 9:19 PM Van Snyder via cctalk &lt;cctalk(a)classiccmp.org&gt; wrote: I'm currently working, off and on, on exactly this. I'm translating the Imp language (which is a very Algol60-like derivative of Atlas Autocode, which itself was a rival response to the '58 proposal that became Algol 60) to C. Yes, I am limited to gcc due to the nested procedures, but I'm doing it because I think gcc has a better chance of some longevity than a native bootstrapped Imp port. (which we currently have to x86 architecture as well as this translation project.) Imp had extensive runtime diagnostics and post mortem facilities as well as symbolic debugging on some platforms, and the linux/gcc environment has allowed me to reproduce many of them. Extensive use of #line allows most of the diagnostics to appear to be in the source language and the mapping to C is close enough that debugging with gdb is not an issue. I'm actually working on two translators - one is high-level and attempts to generate C code that looks and feels more or less like native C; the other is lower level and translates from the intermediate code generated by one of the last native Imp compilers into C that is not meant to be seen by the programmer - it's just another step in the translation. To give you a comparison, here's an imp program that assembles 6809 code: https://gtoal.com/history.dcs.ed.ac.uk/archive/languages/imp77-tmp/gtoal/re… (that's the compiler's listing file, I could point you at the .imp source instead but unfortunately my stupid web server causes it to download so you can't trivially view it in a browser) Here's the high-level generated C. This was created by writing a new parser and compiler from scratch: https://gtoal.com/history.dcs.ed.ac.uk/archive/languages/imp77-tmp/gtoal/re… Here's the more low-level generated C (which for now is a more accurate translation): https://gtoal.com/history.dcs.ed.ac.uk/archive/languages/imp77-tmp/gtoal/re… (You can see the differences in style at places like line 429 where an if/then/else statement is broken down into gotos in the low-level translation) The low-level translation is more or less a direct transliteration of the assembly code that the original imp77 front end (now translated into C) generated: https://gtoal.com/history.dcs.ed.ac.uk/archive/languages/imp77-tmp/gtoal/re… The GCC route lets me target multiple architectures (eg code comparison at https://gtoal.com/history.dcs.ed.ac.uk/archive/languages/imp77-tmp/gtoal/co… ) So I do believe there is still a place for language translation of this sort. Graham

On Sat, 20 Jun 2026, ben via cctalk wrote: I heard that all of the Lisa OS, and much of Macintosh, were written in Pascal. (obviously other than some low-level drivers and performance critical routines?) Is that correct? http://pascal.hansotten.com/apple-lisa-pascal/ When did Macintosh development switch to C? -- Grumpy Ol' Fred cisin(a)xenosoft.com

On Sat, Jun 20, 2026 at 3:02 PM Fred Cisin via cctalk &lt;cctalk(a)classiccmp.org&gt; wrote: I remember seeing articles about Pascal and the Lisa in 1983, and I definitely read in "Inside Macintosh" (the "hernia manuals") in 1984 that there was a Pascal calling convention for various toolbox functions. As I read it, you were expected to write your apps in Pascal or Assembler. C wasn't an option for me (I didn't start learning it until the following year, on a VAX running UNIX), and I wanted no part in Pascal. Totally put me off developing for the Mac. System 7 (early 90s) AFAIK. -ethan

On Mon, 2026-06-22 at 13:52 -0400, Paul Koning via cctalk wrote: Pascal was Wirth's reaction to Algol 68, which he thought was overly complicated. His first step along the way was Algol W ("W" for "Wirth"). That's where he introduced the dot notation for structure members. My first compiler class, at the UC Santa Cruz "Summer Institute in Computer Science" in 1983, used MetaWare Pascal, which the co-author Tom Pennello described as "overextended Pascal."

On 22/06/2026 7:52 pm, Paul Koning via cctalk wrote: Ah yes, Pascal was my first properly structured language, after Fortran, M6800 assembler/machine code and Basic. Pascal on a Decsystem-20 (sometimes referred to as PDP-20, but it was a LP10 running TOPS-20). Still really enjoyed C on UNIX (PDP-11 and M68000). Julf

On Thu, 18 Jun 2026, Bill Gunshannon via cctalk wrote: Hand otimizing was more likely machine language programming rather than "programming in assembler". Sixty years ago, I took a course in 1401 machine language. The 1401 was inconvenient to access regularly, but there was a [homemade?] "compatability"? "conversion" program to be able to run small 1401 machine language programs on the 1620. I had just had a few years of FORTRAN programming, and a course using 1620 PDQ FORTRAN, but 1401 machine language changed my view of programming. Most fun that I EVER had in a programming class. I was going to then be taking a course in SPS ("Symbolic Programming system"?)(a low level assembler) the next semester, but I got hired away for a gig that got out of hand, almost into a career, and it was decade(s) before I got back. The 1401 was long gone :-( I never experienced Autocoder; I understand that it is a much more "advanced" assembler. -- Grumpy Ol' Fred cisin(a)xenosoft.com

On Thu, 2026-06-18 at 16:28 -0400, Paul Koning via cctalk wrote: I just now had a conversation with Google (not Google Gemini, just the search engine AI). I mentioned that I had written an IBM 1130 emulator in Varian V-73 microcode. It said "You must be Paul Koning." Did you wrote one too?

On Fri, 2026-06-19 at 09:54 -0400, Paul Koning via cctalk wrote: V73 fit in a standard 19" rack, about six inches high. Flat panel. Not membrane switches, IIRC. Made by Varian Data Machines, not DEC. Varian sold the Data Machines division to Sperry Univac, who produced the V77 but couldn't actually run with it. Then they pounded Data Machines into the ground. In the late 1960s one of my colleagues needed to process some telemetry data. The client couldn't afford time on the IBM 7094 so he wrote the code for the 1620. To do octal arithmetic he changed the tables. He exploited the CADET — "Can't Add Doesn't Even Try" — nature of the 1620.

On Thursday, June 18, 2026 at 14:10, Bill Gunshannon via cctalk wrote: The comp.lang.ada FAQ disagrees: Last-posted: 22 April 1996 comp.lang.ada Frequently Asked Questions (FAQ) [...] 4.2.1: GNAT, The GNU NYU Ada Translator -- An Ada 95 Compiler GNAT is a compiler for Ada 95 that accepts Ada 95 source code and generates executable (machine) code (GNAT is a compiler and does not, repeat: DOES NOT, generate C code). It is based on the Free Software Foundation (FSF)'s gcc, a portable compilation system for a variety of languages. From "The GNAT Project: A GNU-Ada 9X Compiler" by Edmond Schonberg and Bernard Banner of the New York University Courant Institute: The front-end of the GNAT compiler is thus written in Ada 9X. The back-end of the compiler is the back-end of GCC proper, extended to meet the needs of Ada semantics. I have read that the "Translator" part of the name may be the source of the confusion. -- Dave

On Friday, June 19, 2026 at 8:53, Bill Gunshannon via cctalk wrote: My first exposure was with GNAT 3.04 from June 1996. I've been using later versions off-and-on ever since. No version that I've used has generated C code. Indeed, I'm not sure how the "GNU" part of the "GNU NYU Ada Translator" would fit in. What part of gcc generates C code for a front end? Dewar and Schonberg were, of course, colleagues on the GNAT project, which was targeted specifically at Ada 95 (then Ada 9X) and intended as a free compiler. The CLA FAQ, which was written at the time of the Ada 95 revision, lists some two dozen companies offering Ada compilers. Ada 83 pre-dates that and undoubtedly had a number of commercial compiler vendors, given the US DOD Ada mandate. I don't think you're wrong regarding an early Ada compiler that generated C code. I seem to recall such a beast too, although I can't lay my hands on a reference. -- Dave

On Fri, 19 Jun 2026, Paul Koning via cctalk wrote: Original GCC was a plain RTL compiler, other internal representations were only retrofitted later on, in early 2000s. You are right in that their use enabled optimisations which were infeasible with RTL. Maciej

There were some issues with UNCOL which Peter Robertson attempted to remediate in his more successful portable intermediate code, ICODE, which eventually was used in implementations of Imp, Pascal, Fortran, Occam and C. https://www.gtoal.com/imp77/psrthesis/psrthesis.html (that's the same intermediate code that I based my recent Imp77 to C translation on as mentioned in an earlier post) It would still be usable but no-one has produced a native code generating back end using ICode since the late 80's. (well, except our x86 recreation). Back in the day there were the languages I mentioned above as front ends and about 20 different architecture back ends. It was quite a successful example of an UNCOL-like system. G On Sun, Jun 21, 2026 at 8:43 PM Van Snyder via cctalk &lt;cctalk(a)classiccmp.org&gt; wrote:

Intercal lives.

On Fri, 19 Jun 2026, J. David Bryan via cctalk wrote: Perhaps it was merely intended to be run under GNU, or become part of the GNU binutils? Or GNU tool than the "GNU Compiler Collection"?

Wonder if the Ada -> C translator was the commercial Ada development system they had to bootstrap from ? At least one animal like that existed at one time. Another such animal was Telelogic's SDT ie SDL92 development system there the tool generated C.

On Fri, 2026-06-19 at 08:53 -0400, Bill Gunshannon via cctalk wrote: In addition to GNAT there are a number of excellent commercial compilers. I don't know whether any are free, or have evaluation or student licenses. The Ada standard committees (ARG and WG9) are active.

On Wed, 2026-06-17 at 13:55 -0400, Paul Koning via cctalk wrote: A bit tangential, but… speaking of Assembler, I remember hearing decades ago that Multics was 85% PL/1 and 15% assembler. Has anybody gotten Multics running on modern hardware, such as Intel? — probably a better platform for it than the GE 645 or Honeywell 6180.

On Jun 17, 2026, at 2:29 PM, David Wade via cctalk &lt;cctalk(a)classiccmp.org&gt; wrote: Multics has been running on modern hardware for years via emulation. It works just fine on a Raspberry Pi, and it amuses me to have it running on a system the size of a deck of cards, when I think about the DPS-8’s I used to work on (running GCOS-8). Sadly, as far as I know, no one has gotten GCOS running, as I don’t know of anyone with the software. https://gitlab.com/dps8m/dps8m https://dps8m.gitlab.io/dps8m/ https://multics-wiki.swenson.org/ Zane

Jun 17, 2026, 10:56 by cctalk(a)classiccmp.org: It also calls into question the Operating Systems courses said academic's employers offer.  As embarrassing (well, maybe un-sexy) as it would be to admit, a non-trivial amount of platform specific assembly language is written for OSes (usually for drivers, but there are other bits of the OS involved as well).  Even for simulation environments like OSP. -- The Doctor [412/724/301/703/415/510] WWW: https://drwho.virtadpt.net/ My code, of course, begins with 4GH.

Definitely alive. See Zeptoforth, for current ARM microcontrollers. Obviouslly alive, given that millions use emacs which uses a LISP dialect. 60 or 68? Are Univac/Burroughs machines still around in production? If so, 60 is alive. And Algol 68 runs on current machines, via GCC. It's WATFIV. And those are FORTRAN implementations, not languages in their own right. I don't use it a lot, but I stil run it on my Mac from time to time when I need to do some magic that it does more easily (for me) than emacs. Some other older languages: SNOBOL SYMPL CYBIL Tutor (probably not alive by your definition but alive surprisingly recently) PL/I POP-2 paul

Here's one. IITRAN. It was written at IIT at least for the IBM 7044, IBM 360, and Univac 1108. It was an ALGOL like language used for teaching new students. Does the executable interepreter exist on any platform? Does the source code exist on any platform? <https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail> Virus-free.www.avast.com <https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail> <#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>

Also used to write games on the Atari System 2 arcade platform (using a 10MHz T11) played by millions of unsuspecting prepubescent miscreants, including yours truly (Paperboy and 720). Probably the most prominent use of the language that no one ever noticed. -- ------------------------------------ personal: http://www.cameronkaiser.com/ -- Cameron Kaiser * Floodgap Systems * www.floodgap.com * ckaiser(a)floodgap.com -- Po-Ching Lives! ------------------------------------------------------------

With regards to LISP, do i need to mention AutoCAD ? Extensions was at least done using a version of LISP.

When I arrived at Rice University in 1973, they still had their Burroughs 5500 and were doing research with BCPL. I was simply too late to participate in it (the 5500 was retired after my first year). But if you find some 70-something-year-old Rice graduates you'll still find fondness for the language. mcl

Microcode is also alive and well, and not just the Amd 29xx bit slice paradigm. How else would you control (SIMD) mills implemented in FPGA fabric : VLIW. etc etc Martin -----Original Message----- From: Mike Katz via cctalk [mailto:cctalk@classiccmp.org] Sent: 17 June 2026 01:54 To: Ken Seefried via cctalk &lt;cctalk(a)classiccmp.org&gt; Cc: Mike Katz &lt;bitwiz(a)12bitsbest.com&gt; Subject: [cctalk] Re: "It's dead, Jim" Was: Re: Intel 3086 Someone said assembler was dead, I have to disagree.  I still program in 6809, 68000, ARM, HP-41 and HP71 assembler from time to time.  I've used ARM assembler in my work recently. On 6/16/2026 8:39 PM, Ken Seefried via cctalk wrote:

On 6/16/26 20:54, Mike Katz via cctalk wrote: Last time I used assembler was to create an unpacking routine on the PRU of a BeagleBone computer (200 MIPS 32-bit microcontroller aux processor.) It unpacks run-length encoded pixel data on a photoplotter in real time.  5 us between pixels, and the PRU is just loafing. Jon

On Wed, 2026-06-17 at 10:03 -0500, Jon Elson via cctalk wrote: Modula-2 is a descendant of Pascal, actually developed by Wirth and his students. Then Wirth had second thoughts about separating definition and implementation, a cornerstone of Modula-2 and Ada, and developed Oberon. I had some correspondence with him about my proposal for a similar separation of definition and implementation in Fortran, now standardized as "submodules" — analogous to the Implementation Module in Modula-2 or the Package Body in Ada. He agreed that in large or complicated projects, it's important. It also allows commercial providers such as NAG to publish definitive machine-readable interfaces without publishing the source code for intellectual property or trade secrets. Fortran is alive and well. The national and international committees are meeting in Coventry (and Zoom) this week to put the finishing touches on Fortran 2028.

On Wed, 2026-06-17 at 13:48 -0600, ben via cctalk wrote: Very few FORTRAN IV statements have been deleted from modern Fortran (the official spelling was changed in 1990). ASSIGN and assigned GO TO, PAUSE, and the "H: edit descriptor  were eliminated. Real and double precision DO loops (not in FORTRAN IV and mistakenly added in Fortran 90) were deleted. The interpretation of the first character of "printed" output as vertical format control was deleted in 2008, without changing any syntax, because there was no way to control it, and no way to detect whether output to a particular unit did it. A post-processor is trivial so no effort was done to add control and inquiry. Arithmetic IF and nonblock DO were deleted in Fortran 2018. Fixed form has NOT been deleted, even though nearly-trivial preprocessors exist to convert to free form. Read Annex B of the standard. The Intel delegate, who had been a DEC VAX Fortran developer (the late Stan Whitlock), remarked to the committee "Go ahead and delete it; I still have to compile it forever." The NAG compiler has a -dusty switch that interprets some statements that have the same syntax as in FORTRAN IV with FORTRAN IV semantics if there are subtle differences. So you can almost certainly still compile your dusty FORTRAN IV decks (if you have a card reader). And if you can't compile them, there's a good chance the necessary changes are simple. Polyhedron Software's SPAG code might handle all of them automagically — and produce a nice structured code instead of spaghetti.

On Tue, 2026-06-16 at 21:39 -0400, Ken Seefried via cctalk wrote: Honeywell provided a COBOL to Ada translator more than forty years ago. I don't know whether anybody actually used it.

On 17/06/2026 4:49 pm, Bill Gunshannon via cctalk wrote: Sure, but would anyone starting a new software project choose COBOL as the implementation language? Julf

On 17/06/2026 15:56, Bill Gunshannon via cctalk wrote: Nothing, but we will disagree on defining its domain which I believe must include hardware and software. So if I am extending something written in COBOL on an IBM Mainframe or Mid-range then yes, that Is the domain of COBOL. For something thats going to run in the cloud with a web front end unlikely. Not really. The issue with COBOL as I see it is that employers think COBOL staff have no other skills, are desperate for work, so they can pay 1950s rates. Dave G4UGM Dave

On 17/06/2026 19:01, Paul Koning wrote: That is the trouble with this list. Pure technical. You CAN run COBOL on anything, but could you convince a 30 year old project manager to let you.. .. the domain must be the hardware, software and liveware... Dave

On Wed, Jun 17, 2026 at 10:52 AM Johan Helsingius via cctalk < cctalk(a)classiccmp.org&gt; wrote: If you're a COBOL shop, have COBOL platforms, COBOL devops in place with COBOL programmers, you just might pick COBOL for your project. I know we do (large financial). There is more to IT than what the programming social media sites tell you to think. Maybe check it out some time. KJ

On Thu, Jun 18, 2026 at 3:09 AM Johan Helsingius via cctalk < cctalk(a)classiccmp.org&gt; wrote: By that definition, most programming is legacy programming, particularly in an enterprise setting. Most of us don't green-field everything every time a new project starts. By what I'd consider a more reasonable definition our COBOL folks are starting new projects like "create Zelle service" or "enable custodianship of digital assets" even though those things are intimately bound to existing infrastructure. But I will concede that if you define 'new COBOL project' as "First you buy a new mainframe that runs nothing we're previously written...", then you get to be right. One mans pointed correction is another mans slur. You should assume (and respect the fact) that many of us have been on this list since the 1990s and don't need to be condescended to with regards to list etiquette.

Somebody wrote: What's HN? If I haven't heard of it, does that mean its dead? Or does it mean I'm dead? ... It appears my telephone line may also be dead. I just phoned the telephone company to enquire about my account balance. The automaton on the other end said: "If you are calling about the number you have dialled from, please press 1." so I pressed 1. Then it asked for my account number which I provided. Then it said: "Your service is currently impacted by a network outage. Our technicians are working to repair the fault and restore service as quickly as possible." "Estimated restoration time is 11:00 AM on 17th June. Thank you for your patience." (This was at 11:45). a) Why is this stupid thing trying to tell me about an non-existant fault on my line? b) Why does this stupid thing think it can tell me about a fault on my line using the line it thinks is faulty in order to tell me? (I'm not sure this Artificial Insemination thing is any use at all...) Regards, Peter Coghlan.

On 6/16/2026 8:11 PM, Fred Cisin via cctalk wrote: And COBOL... Don't forget COBOL.... :-) bill

On Tue, 16 Jun 2026, Bill Gunshannon via cctalk wrote: That's the best example. There is also FORTRAN Fortran APL ADA Pascal RPG BASIC Assembly language (35? years ago, UC Berkeley's revised curricula for CS, Clancy and Harvey said, "Nobody prograns in assembly language any more, nor ever will again")

COBOL and Fortran won’t die, they’ll just be certified with StarDates. //m On Wed, 17 Jun 2026 at 11:30 am, Paul Koning via cctalk < cctalk(a)classiccmp.org&gt; wrote:

I agree with that. I have spent the last 45 years doing embedded programming mostly in C or Assembler. Even if some of the architectures were a bit unusual (Intel 80XXX, TI 9900, RCA 1802, HP-41 and HP-71) or a bit sparse (PDP-8 and Epson 4 bit watch CPU).  The only assembly I really didn't enjoy was 80XXX and the 6502. For obscure languages APL, LISP, Prolog, ADA, Occam & Pilot come to mind.  Focal, though widely used on DEC computers seems to have died with the demise of DEC itself. On 6/16/2026 8:48 PM, Ken Seefried via cctalk wrote:

On Tue, 2026-06-16 at 21:11 -0500, Mike Katz via cctalk wrote: I was told, by whom I do not remember, that the IBM 360 was designed in APL, starting in 1962. There's a paper about it, "A Formal Description of System/360," which was entirely written using APL notation. The first APL compiler that IBM used was written by Larry Breed and Philip Abrams in Fortran IV and ran on the IBM 7090 under IBSYS.

VMT heads up that A68 is in GCC 16 I have a 50 yo programs to run Martin -----Original Message----- From: Paul Koning via cctalk [mailto:cctalk@classiccmp.org] Sent: 17 June 2026 01:30 To: cctalk(a)classiccmp.org Cc: Paul Koning &lt;paulkoning(a)comcast.net&gt; Subject: [cctalk] Re: "It's dead, Jim" Was: Re: Intel 3086 COBOL was recently added to the GCC compiler suite (as was Modula-2 and Algol 68). And gfortran sees lots and lots of development; it seems that various fields use it extensively still. Weather forecasting seems to be one example. And ADA has been and remains an active part of GCC, too. Not to mention that it's the foundation of VHDL (and the open source GHDL simulator is written in Ada). As for assembly language, that certainly is still alive. A FORTH package I use extensively on the Raspberry Pico is written mostly in Forth, but the rest in assembler -- no C to be found anywhere at all. For that matter, you can't program the Pico's programmable state machine engines in anything other than assembler, for the simple reason that nothing else will serve when you only have 32 words of program memory. paul

On Wed, 17 Jun 2026, Bill Gunshannon via cctalk wrote: On THIS list, "Recently" can mean in the last 40 years,

On Wed, 2026-06-17 at 10:23 -0400, Bill Gunshannon via cctalk wrote: The newest revision of the ISO/IEC 1989 COBOL standard was published three years ago. If COBOL is dead, the revision must have been written by psychics reading Tarot cards. (1989 is the document number, not the year it was first published).

On Sun, 14 Jun 2026, Mike Katz via cctalk wrote: Those good software items are what made it successful. Similarly, some say that Apple's success was reliant on VsiCalc. It was only $5,000 if you bought a ridiculous bundle from IBM with seriously inflated prices for commodity hardware, and all of the software that you did NOT want. I hope that nobody other than college administrators paid THAT. Without unwanted software bundle, it was more like $2500 My first was $1320? plus about $600 for CGA and disk controller boards, and then I put in my own 4116 RAM and floppy drives (same as TRS80), and used my own CCTV composite monitor. Google Gemini says $1565, including CGA (~$300) True, IBM wanted about $3000 more for FDC and one drive, 48K (three rows of 9 4116s), CGA board, and more than $600 for a color monitor! Floppy controller board, and CGA were only available through IBM initially, but the rest was same as TRS-80, etc. at about 1/10 the price. IBM sold VisiCalc; Wordstar was independently from MicroPro; dBase II was from Ashton Tate. For a word processor, IBM tried selling "Easy Writer" by John Draper! (included in the $5,000 bundle?) Supercalc came out soon later, and was much cheaper than IBM's price for VisiCalc, Immediately after I started teaching in the college, my department chair had me come in on the weekend, and he and I put RAM, FDC, CGA, and drives into a couple dozen bare PCs. -- Grumpy Ol' Fred cisin(a)xenosoft.com

[usual long rambling tangents] Microsoft was not a big player in any of that until 1981. They had no major input in OEMs decisions of processor until 1980? Although their earliest product was a BASIC for 8080, they tried to branch out into more platforms of BASIC. They wrote the BASIC for the TRS80 (Z80, which was based on 8080), they also wrote Applesoft BASIC (6502). They did eventually make some hardware, most notably the Z80 Softcard for the ApplelII, and later, mice. Tim Paterson was developing computers based on 8086 at Seattle Computer Products. He wrote a "placeholder" OS called "Quick and Dirty OS" to use while waiting for the overdue CP/M-86. When IBM decided to "take over" home/personal computers, they looked at the 8086. They chose 8088, which I liken to an "8086 cut back, to make an 8 bit version of 8086", because the computer could be designed and built cheaply, as if it were 8 bit. [insert whole long CP/M - MS-DOS/PC-DOS history] Microsoft providing the OS created a long-term working association. The popularity of the 5150 PC created an unbelievable market for clones, imitations, similar, etc. running MS-DOS. And caused Microsoft to become what it is. In contrast, Jobs locked in the Mac, so that hardly anybody could build an imitation. IBM used the "8 bit version of 8086". But not an 8086. They used an 8086 in the Displaywriter (1980?) Later, they went to 80286 (5170 PC/AT). Around 1985-1986?, IBM came out with a "whole new line" of computers ("PS/2") In 1987? They made the PS/2 model 25 and the PS/2 model 30 which used 8086. I've heard that the PC/JX (Japanese market) had an 8086. I don't know of any other IBM devices with 8086, although 8088 and 8086 overlapped in design so much that that conflation is excusable. Intel was the biggest in processors; IBM dominated PC design; Microsoft had a virtual monopoly on Operating system. UCSD Pascal never really caught on, for several reasons. DRI's CP/M-86, "Concurrent CP/M-86", and later "Concurrent PC-DOS" and DR-DOS were welcome alternative choices, although never big enough to seriously challenge Microsoft. Microsoft trademarked "MS-DOS", IBM did NOT trademark "PC" (and it would have been absurd to try), and IBM DID NOT trademark "PC-DOS", declaring "PC/Personal Computer" and "PC-DOS to be a "desctiptive terms, not really names"! So, although it created confusion, particularly assumed affiliation with PC-DOS (IBM specific version of MS-DOS), DRI, IBM, and Microsoft did not end up in court. Over "Comcurrent PC-DOS". Another factor was that MS-DOS was based on CP/M (not copied nor plagiarized, but indisputably based on it), so IBM and Microsoft were loathe to fight with DRI over the trademark. so, Intel, IBM, and Microsoft did a lot of business together and relied on each other, but they never managed to create a cartel. So, IBM's dominance of the market made the 8088 extremely poplular, and many companies eventually branched out creating variants, some of which used 8086. A small number of 80186, then lots of 80286, 80386, 80486, and then other chip manufacturers got in on the action. Nec made the V20, V30; AMD, Cyrix, etc. After the 80486, Cyrix came out with a fancy 80486 that they called the 80586, . . . The consensus was that you "can't trademark a number" (?? How about the Oldsmobile 88, Oldsmobile 442, etc?) So Intel chose to go to a name, instead. They came up with "Pentium". I heard that that was the winner of an in-house contest; what were the losers? Even so, there were possible problems that they didn't think of: Asahi Pentax was acquired by Honeywell. Honeywell, who made some computers, could have come out with a "Pentaxium" processor (and contracted with a chip maker) with enough prior usage claim to be protected. But, it also gave the other chip makers impetus to come out with their own variation, rather trying to exactly copy Intel Some people think of Intel, IBM, and Microsoft as being a cartel, but they never got it together enough to organize anything like that. -- Grumpy Ol' Fred cisin(a)xenosoft.com On Sun, 14 Jun 2026, Murray McCullough via cctalk wrote: > What role does the duopoly, Intel/Microsoft, play in the success of the > earlier years of the 8086? When choice is restricted sometimes the best > isn't the most successful! > > Happy computing, > > Murray 🙂

Intel has shipped approximately 12 Billion CPUs since it's introduction in 1978. There have been more than 350 Billion ARM based CPUs sold since it's introduction in 1985. That means Intel is selling, on average, 250 Million per year in the 48 years since it's introduction. Whereas, ARM, on average, there have been 8.5 billion ARM based CPUs per year in the 41 years since it's introduction. ARM wins in sheer numbers and wins even more in terms of architecture refinement and structure. On 6/14/2026 12:56 PM, Fred Cisin via cctalk wrote:

Fred Cisin wrote: I have seen it claimed that Intel made the 8086 as a stop-gap until the iAPX 432 was finished.