On 2026-09-27 5:40 p.m., Jim Davis via cctalk wrote: Kuth covers tertiary math, so half the logic is done there.How to handle boolean logic is the tricky stuff.Building the cool front panel would use 3 way switches and bidirectional red/green leds.

It's better than it's 2+2=5 (for large values of 2) and reporting the moon is made of green cheese days. last week, In a mildly pissed session, We/i/it did a sketch of an analog tensor cell, just op amps and floating gates Digital on the in and out, but a 1M^2 cells in the core. AI's agree, It looks good. Give it a try, Promise you won't get your hand blown off playing with improvised explosives. On Mon, Sep 28, 2026 at 1:06 PM Richard via cctalk <cctalk(a)classiccmp.org> wrote:

That depends. One answer is "all of them" -- because any general purpose computer can do both. If you meant "had machine instructions both for decimal and for binary arithmetic", the answer includes the PDP-11 and VAX, and I think the Intel 4004 had instructions used as building blocks in decimal arithmetic. By the way, on the subject of good books, Bell & Newell, "Computer structures: readings and examples" is very good. It categorizes computers a number of different ways, and covers a range of examples touching on most of the categories. The Google search mentions an authorized online copy at Gordon Bell's website: https://gordonbell.azurewebsites.net/cgb%20files/computer%20structures%20re… And here is the table of contents: https://ed-thelen.org/comp-hist/bell-71-t-o-c.html Apparently there is a second edition. I'm pretty sure I have the first edition. Lots of neat machines in there. Mainstream ones like the PDP-8 and IBM 360, stack machines (Burroughs 5500), vector processing (SIMD) machines (Illiac IV), and oddballs like the very early Dutch computer ZEBRA (a Dutch acronym for "extremely simple binary computer"). That one could be considered a microprogrammed machine where the horizontal microcode is what you use to write your applications. Since I learned about printing and typesetting at one point, I found it fascinating that this book comes in an unusual format: it is only one of two books I have ever seen that is square (the other being "Powers of ten" -- wonderful also). paul

Um, you forgot COBOL! Which is still around; GCC now includes a COBOL compiler. RPG maybe also, though I don't know that language. And nowadays there is Python, though I don't know if the decimal arithmetic library on Python/VAX would use those instructions. Nope. It's the "commercial instruction set" which was an optional feature on several PDP-11 models sold by DEC, clearly documented in the PDP-11 architecture manual, and supported by at least COBOL and probably RPG; the kernel would also use the CIS MOVC instruction for memcpy if present in the machine. There is of course the famous PDP-11/74 (one of two unshipped models of that name) which was an 11/70 with CIS added, canceled around the time the VAX came out. Internal rumors were that it was due to it being faster than the VAX for COBOL workloads. paul

Me neither. The closest I have come to RPG is my youngest sister, who was a commercial contract programmer around Y2K and made quite a good living updating COBOL and RPG programs. To her I owe this lovely quote: "COBOL -- the programming language that gives you writer's cramp". GCC has accumulated quite a bunch of languages at this point. C/C++ of course, and Objective C/C++, and Fortran. Also Ada, COBOL, Modula-2, D, Algol-68, and (on a fork) VHDL. I think that's it though I may have missed one or two. paul

And all our daily driver languages look like Algol or FORTRAN, excepting Forth and APL, both written by mad scientists, On Tue, Sep 29, 2026 at 4:51 PM Paul Koning via cctalk <cctalk(a)classiccmp.org> wrote:

aka curly-brace languages On Wed, Sep 30, 2026 at 2:31 PM ben via cctalk <cctalk(a)classiccmp.org> wrote:

Classic FORTRAN also (mostly) ignores white space, except for the first columns of a line. That is, the following are equivalent: PROGRAM P R O G R A M PRO GRAM PR OG RA M From: "Paul Koning via cctalk" <cctalk(a)classiccmp.org> To: "cctalk" <cctalk(a)classiccmp.org> Cc: "Paul Koning" <paulkoning(a)comcast.net> Sent: Wednesday, September 30, 2026 8:10:21 AM Subject: [cctalk] Re: Good computer books for 1960's computers Huh? The only language I can think of that ignores whitespace is TECO. Just about everything else treats whitespace as a token separator. paul

On Wed, Sep 30, 2026 at 11:10 AM Paul Koning via cctalk <cctalk(a)classiccmp.org> wrote: Microsoft 8-bit BASIC tokenizes when you enter a new line of BASIC but whitespace is not required between tokens. It can get confused in very specific edge cases but mostly, it works. When executing tokenized code, outside of quotes, the actual next-char fetch routine in low mem (CHRGET) explicitly skips whitespace, so the token evaluator never sees it. The stream of tokens and untokenized args presented to the BASIC statement evaluator sees exactly the same stream from the following two lines... 10FORX=1TO100:PRINTX:NEXTX and 10 FOR X = 1 TO 100 : PRINT X : NEXT X Those spaces are stored in RAM but CHRGET skips them on each fetch. -ethan

And Lisp. And TECO. Also SNOBOL, I think. And POP-2. But yes, Algol-60 is the ancestor of most programming languages, in particular C and C++ (though C++ also borrowed a bunch from Algol-68). paul

On 9/29/26 12:36, Paul Koning via cctalk wrote: Anything with the bit slice 2901 or 2903 chips from AMD had the capability to do hardware binary and packed decimal operations. Of course, the microcode would need to have implemented those. Jon

Paul Koning wrote: The Motorola 68000 has the instructions ABCD, NBCD, and SBCD for decimal arithmetic.

On 29/09/2026 18:22, ben via cctalk wrote: Oddly enough I've been compiling a list of  non-power-of-two machines and Ai does a good (or at least credible) job of turning web pages I don't understand into ones I think mostly match DeepL and Google Translate. Here's a starting list. I've not checked it but a few of the machine son there (like some of the Odra ones) are real (in the sense that I've independently found info about them, not that I've  ever seen them ... but then I've never seen a PDP-1 either). Now I have to decide whether the Dnepr-1 counts as 26 bits for my list ... should parity/sign count?? Oh, and I was missing a 29-bit example ... so more checking to do (you never know, it might have imagined it!) Antonio ChatGPT | Country | Computer       |        Date |                 Width | English description / notes                                                      | | ------- | -------------- | ----------: | --------------------: | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | USSR    | **M-4**        |        1961 | **23-bit** arithmetic | Real-time radar-control computer; used 23-bit fixed-point numbers. RAM stored 24-bit numbers and program ROM 30-bit words. ([Computer Museum][1])                                         | | USSR    | **M4M**        |        1962 | **23-bit** arithmetic | Second machine of the M-4 family, produced in 1962; same 23-bit arithmetic architecture. ([Computer Museum][1])                                                        | | USSR    | **M4-2M**      | 1964 onward |            **29-bit** | High-performance real-time control computer; 29-bit instruction/constant words and 29-bit main-memory words. Produced at Zagorsk for about 15 years. ([Computer Museum][1])  | | USSR    | **Dnepr**      |   1961–1971 |            **26-bit** | General-purpose process-control computer, developed in Ukraine. Its designers deliberately chose a 26-bit machine word as sufficient for process-control calculations. ([glushkov.su][2]) | | USSR    | **Ural-11**    |   1965–1975 |            **24-bit** | First of the Ural-11/14/16 compatible family; fixed- and floating-point arithmetic and alphanumeric processing. 123 produced. ([Computer Museum][3])     | | USSR    | **Ural-14**    |   1965–1974 |            **24-bit** | Intermediate Ural model; addressed word 24 bits, with optional 48-bit arithmetic. 201 produced. ([Computer Museum][4])                                                         | | USSR    | **Ural-16**    |       1967– |            **24-bit** | Senior member of the Ural-11/14/16 family. The family used the same 24-bit basic architecture. ([Computer Museum][5])                                                          | | Poland  | **ZAM-21**     |        1965 |            **24-bit** | Experimental transistorised computer for numerical/scientific computing. Three machines were produced. ([Pismo Hebrajskie][6])                                                           | | Poland  | **ZAM-41**     |   1966–1970 |            **24-bit** | Large second-generation transistor computer for data processing; 24-bit words, 24- or 48-bit numbers. 16 produced. ([Pismo Hebrajskie][7])                                                | | Poland  | **ODRA 1204**  |   1967–1972 |            **24-bit** | Second-generation transistor computer developed and manufactured by ELWRO. 179 produced. ([Pismo Hebrajskie][8])                                                       | | Poland  | **ODRA 1304**  |   1969/1970 |            **24-bit** | Polish computer compatible with the ICL 1900 series; pilot series in 1969 and serial production from 1970. ([ResearchGate][9])                                                       | | Poland  | **ODRA 1305**  |        1970 |            **24-bit** | Third-generation ELWRO computer based on the ICL 1900 architecture; 24-bit words. ([Software Preservation Group][10])                                                                  | | Poland  | **ODRA 1325**  | early 1970s |            **24-bit** | Third-generation ODRA 1300-series machine; specification gives 16/32K 24-bit words. ([https://historiainformatyki.pl][11])                                                         | | Romania | **DACICC-200** |        1968 |            **24-bit** | Romanian-designed and manufactured computer; 24-bit fixed-point words, with 37-bit mantissa + 10-bit exponent for floating point. ([ictp.acad.ro][12])                                    | [1]: https://computer-museum.ru/english/m4.htm?utm_source=chatgpt.com "M-4, M4M, M4-2M and M4-3M Control Computers. Russian Virtual Computer Museum" [2]: https://glushkov.su/eng/nauchnaia-deiatelnost-vm-glushkova?utm_source=chatg… "Scientific activity of V.M. Glushkov" [3]: https://www.computer-museum.ru/english/ural_11.php?utm_source=chatgpt.com "Electronic Digital Computer Ural-11 (Урал-11). Russian Virtual Computer Museum" [4]: https://computer-museum.ru/english/ural_14.php?utm_source=chatgpt.com "Electronic Digital Computer Ural-14 (Урал-14). Russian Virtual Computer Museum" [5]: https://computer-museum.ru/english/ural_11.php?utm_source=chatgpt.com "Electronic Digital Computer Ural-11 (Урал-11). Russian Virtual Computer Museum" [6]: https://xn--meb.pisz.pl/ZAM-21?utm_source=chatgpt.com "ZAM-21 - Encyklopedia internetowa" [7]: https://www.xn--meb.pisz.pl/ZAM-41?utm_source=chatgpt.com "ZAM-41 - Encyklopedia internetowa" [8]: https://www.xn--meb.pisz.pl/Odra_1204?utm_source=chatgpt.com "Odra 1204 - Encyklopedia internetowa" [9]: https://www.researchgate.net/publication/290997919_The_History_Present_Stat… "(PDF) The History, Present State and Future of Information Technology" [10]: https://softwarepreservation.computerhistory.org/prolog/warsaw/doc/Kluz301n… "The 'Marseille Interpreter' -" [11]: https://historiainformatyki.pl/skan.php?doc_id=395&for_download=1&t… "ODRA 1325 Jednostka Centralna Central Processor" [12]: https://ictp.acad.ro/ro/dacicc-200/?utm_source=chatgpt.com "DACICC-200 (1968) - primul calculator cu sistem de operare şi compilator, cel mai puternic calculator construit în România" DeepSeek | Country | Computer Model | Year(s) | Word Size (bits) | Notes | | :--- | :--- | :--- | :--- | :--- | | **Soviet Union** | Argon-1 | 1969 | 24 | Vehicle-based computer, fixed-point numbers | | **Soviet Union** | Radon | 1964 | 24 (command) / 20 (operand) | Dual-machine system, single-address instruction set | | **Soviet Union** | Ural-11 | 1966 | 24 | General-purpose computer, fixed-point binary-decimal numbers | | **Poland** | ZAM-21 | 1961–1964 | 24 | Experimental transistor computer, used for calculations and data processing | | **Poland** | ZAM-41 | 1961 | 24 | Transistor computer for data processing, 16 units produced | | **Poland** | Odra 1001 | 1960–1961 | 18 | Vacuum-tube computer, 2048 words of magnetic drum memory | | **Poland** | Odra 1204 | 1966 | 24 | Transistor computer, 4K–64K words of core memory | | **Poland** | Odra 1304 | 1970–1973 | 24 | Second-generation transistor computer, developed by Elwro | | **Hungary** | EMG-830 | 1968–1969 | 24 | Second-generation transistor computer, 25,000 operations/second | | **Yugoslavia** | CER-10 | 1963 | 30 | First computer developed in SFRY, 4096 words of core memory | Gemini Between **1958 and 1975**, computer developments in Eastern Europe and the Soviet Union utilized word-addressed architectures with word sizes falling into the **18-bit to 30-bit** range. During this era, 24-bit architectures were particularly common for scientific computing, industrial automation, and mainframes. Below is a categorized list of computer systems developed in Eastern Bloc nations matching your criteria (18–30 bit word size, 1958–1975), organized by country and translated into English. --- ### 1. Poland (ELWRO & Academic Institutions) * **Odra 1204 (1967)** * **Word Size:** **24 bits** * **Description:** Designed and manufactured by ELWRO in Wrocław, the Odra 1204 was an original Polish-designed second-generation transistorized computer intended for scientific and industrial calculations. * **Odra 1300 Series — 1304, 1305, 1325 (1970–1973)** * **Word Size:** **24 bits** * **Description:** Built by ELWRO under an official license agreement with Britain's ICT/ICL. These systems were hardware-compatible with the ICT 1900 series and could run ICL operating systems (Executive, GEORGE 2/3). * **K-202 (1971)** * **Word Size:** **16/24 bits** (16-bit word size, but featured extended 24-bit logical addressing space allowing up to 8 MB of memory). * **Description:** Designed by Jacek Karpiński, it was an advanced 16-bit minicomputer with 24-bit addressing capabilities that surpassed many Western equivalents at the time. --- ### 2. Soviet Union (USSR) * **Ural Family — Ural-11, Ural-14, Ural-16 (1964–1969)** * **Word Size:** **24 bits** (addressable in 12-bit or 24-bit half-words; expandable up to 48 bits for floating-point calculations). * **Description:** Developed at the Penza Computing Machines Plant under Penza's design team led by Bashir Rameev. These were semiconductor-based mainframe computers designed for engineering, economic planning, and scientific data processing. * **Dniepr / Dnepr-1 (1961) & Dnepr-2 (1968)** * **Word Size:** **26 bits** (24 data bits + 2 parity/sign bits). * **Description:** Developed by the Glushkov Institute of Cybernetics in Kyiv (Ukrainian SSR), the Dnepr series was the premier general-purpose process-control computer used for industrial automation across the Eastern Bloc. * **M-220 & M-222 (1968)** * **Word Size:** **24 bits** (extended / half-word variants of the 45-bit M-20 architecture). * **Description:** Medium-class scientific minicomputers/mainframes derived from the tube-based M-20 line, rebuilt using discrete semiconductor logic. * **MIR Series — MIR-1 (1965) & MIR-2 (1969)** * **Word Size:** **Microprogrammed / Variable** (processed data in 12-bit to 24-bit chunks with dynamically configured bit structures). * **Description:** Developed under Viktor Glushkov in Kyiv, MIR ("Machine for Engineering Calculations") was designed specifically to execute high-level symbolic mathematical operations directly in hardware. --- ### 3. East Germany (GDR / VEB Robotron) * **Robotron R 4200 / PR 2100 (1968–1970)** * **Word Size:** **18 bits** * **Description:** Manufactured by VEB Kombinat Robotron, these process-control systems used an 18-bit word structure (inspired in part by DEC's PDP-9/PDP-15 architectures) for industrial automation and real-time monitoring. --- ### 4. Czechoslovakia (TESLA / ZPA) * **MSP 2 (1966)** * **Word Size:** **24 bits** * **Description:** Developed at the Research Institute of Mathematical Machines (VÚMS) in Prague, the MSP 2 was a small-scale transistorized computer built for scientific calculations and data gathering. * **ZPA ZTA 900 / ZPA 600 (Late 1960s)** * **Word Size:** **24 bits** * **Description:** Modular industrial process-control computers built primarily for automation in manufacturing plants across Czechoslovakia. --- ### Key Historical Architecture Context During the 1950s–1970s transition period, Eastern European design teams routinely chose **24-bit** structures because 24 bits provided a favorable balance: 1. **Character Storage:** It cleanly fit four 6-bit characters. 2. **Fixed/Floating Point:** It offered sufficient precision for general engineering without requiring expensive 36-bit or 48-bit hardware components. By 1974–1975, most Eastern Bloc development pivoted toward the **ES EVM** (Unified System, based on 32-bit IBM System/360) and **SM EVM** (based on 16-bit DEC PDP-11), making 18-to-30-bit non-standard word sizes obsolete. -- Antonio Carlini antonio(a)acarlini.com

That's certainly a start but just the bare beginning. I suspect that, unti lthe IBM/360, most machines had word lengths not a power of two. For example, none of the DEC machines before the PDP-11. Older IBM machines had assorted word sizes; CDC and Burroughs ditto. I worked on a 1970s era Philips machine with 24 bits, and there were other minicomputer with that word length (Modcomp?). Electrologica had 27 bit words. The MC research machines had word lengths of 30 and 34 bits (the 34 bit one, ARMAC, was a successor of the 30 bit ARRA 2 with a very similar instruction set, two instructions per word, but the address space grew from 1k to 4k so the word was made 4 bits larger). ENIAC was 40 bits I think. I would count the sign since it's part of the data representation, but not the parity bit. And on Burroughs mainframes the 3 tag bits are usually not mentioned; if you do include them (sensible, since they carry semantic information) you end up with a word size of 51 bits. Not often you run into prime numbers for the word size. paul

You can find the original papers about his invention of semaphores in the EWD archive at U Texas. The oldest ones are in Dutch but I think they were translated. One mentions the original name -- (railroad) signposts, which explains the P and V function names, based on the Dutch words "pass" and "leave" -- passing the signpost at the start of a railroad block, and leaving the block. I put much of this into the Wikipedia article about semaphores. The X8 used semaphores in the API of the I/O coprocessor CHARON, to guard the request queues and to manage interrupt delivery. They are also used intensively in Dijkstra's THE operating system, with "hardware semaphores" in device drivers to block until an interrupt, and "software semaphores" to do mutual exclusion, task switching, etc. The underlying magic is an X8 instruction added to the base X1 set. There is the regular add and subtract operations (add/sub memory to register) but also memory-update variants, which add or subtract memory to the register but also atomically deliver the result to memory as well. paul

how many bits count? does the flag bit count in the decimal IBM 1620? It indicated the high order digit of a decimal number, and nothing else. There was a character "record mark" for indicting the right end of a character type field. nb, in the model 2 machine, the flag also indicated the special feature index register, if any. <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> On Tue, Sep 29, 2026 at 4:10 PM Fred Cisin via cctalk &lt;cctalk(a)classiccmp.org&gt; wrote: -- --Carey

On Tue, Sep 29, 2026 at 03:18:43PM -0400, Paul Koning via cctalk wrote: To a first approximation... :) Each accumulator operated on a number stored as 10 decimal digits (each in a ring counter) plus a sign. In pure binary, this would be equivalent to a little over 34 bits. Data read from cards was staged in relays before being transmitted to accumulators. There each digit was stored on 4 relays (but not in BCD). So for a 10-digit number taken from the cards, there were 41 relays (40 for the digits and 1 for the sign) storing it. (Card data could also be treated as 5-digit numbers.) Around '52 or '53, they added a core memory unit. I haven't dug into it as much as the rest of the machine, but I'm pretty sure numbers were stored in it in essentially the same format as the relays holding card data. So while calculation and transmission didn't take place in that format, there were places where numbers were represented with 41 bits. BLS

On 29/09/2026 22:11, Paul Koning wrote: In terms of machines in the wild, I'd expect that copies of IBM systems and PDPs and VAXen would outnumber most things. But most of  the oddball  word sizes look like original machines. Here's what would seem to be an authoritative source: https://simh.trailing-edge.com/semi/cvax.html (near the bottom). And an image (which I remember seeing long enough ago that if it was faked, it was done so by human hands): https://micro.magnet.fsu.edu/creatures/pages/russians.html Antonio -- Antonio Carlini antonio(a)acarlini.com