18

While reading about the ATOLL Checkout Language in the April 1965 issue of DATAMATION (pp 33-35), I came across one letter to the editor (pg 12), discussing the volume required to house the ~1QB memory belonging to an (I can only assume theoretical) new computer, 'Maclogal.' Intrigued, I searched google for any more information about this 'stranger-than-fiction' device, but found nothing mentioning the word 'maclogal' except for within the pdf I was already reading from. (Neither could I find the February issue of the magazine, which the letter referenced.)

Does anyone know any more about Maclogal, or, at the least, have a link to the Feb '65 issue of DATAMATION where I could (presumably) read more?

EDIT:

Thanks to the excellent answers, I now have plenty of information about this computer. However, one question remains: where did the claim of "storage capacity of 1260 × 10^27 alphanumerics" come from?

9
  • 2
    After reading the letter mentioned I just wonder where the Hungarians hid that computer...
    – UncleBod
    Mar 9, 2023 at 9:10
  • 4
    @UncleBod Behind an iron curtain?
    – Raffzahn
    Mar 9, 2023 at 13:48
  • 6
    It is very likely that the name of that computer just includes a typo or transcription error - It is likely that the Magyars named their computer not after an upcoming Apple product ;) , but after their country in Hungarian language. There are a number of mentions of Maglogal, a ferrite core-based computation device on the internet. (Apparently, that was a real system used in the print industry)
    – tofro
    Mar 9, 2023 at 13:48
  • 1
    That's the first time I've ever seen the prefix quetta- in actual usage! (For posterity, 1 QB, one quettabyte, is 10^30 bytes.) Mar 11, 2023 at 21:06
  • 1
    Is it plausible that 10²⁷ is just a typo for 1027? Mar 13, 2023 at 12:59

2 Answers 2

24

As tofro notes in a comment above, "maclogal" is probably a misprint of "maglogal". Indeed, searching Google for "maglogal" turns up a number of documents describing an actual Hungarian computing device by that name — all of them, alas, in Hungarian, which I don't really read. However, with the help of Google's translator and a small amount of hand-editing, I managed to (more or less accurately) translate some relevant paragraphs from these sources:

For example, in http://epa.uz.ua/00200/00220/00088/pdf/firka_EPA00220_1994_1995_02.pdf we find the following paragraph:

Különféle műszaki fejlesztések is folytak az SzK-ban, pl. Hatvany József NC-vezérlést fejlesztett szerszámgépekhez, Münnich Antal titokzatos elektronikus áramköröket tervezett és nyelvészeti problémák számítógépes megoldásával foglalkozott, mint Kiefer Ferenc és Varga Dénes is. Bóka András Ladányi Józseffel és Czili Gyulánéval ferritgyűrűs áramkörökkel (Maglogal, ferritválogató automata), Németh Pál pedig transzfluxorokkal kísérletezett. Különféle áramköri fejlesztéseket végzett Szűcs Károly és Bányai Ferenc.

Various technical developments were also carried out in the SzK [Computing Center of the Hungarian Academy of Sciences], e.g. József Hatvany developed NC control for machine tools, Antal Münnich designed mysterious electronic circuits and dealt with the computer solution of linguistic problems, as did Ferenc Kiefer and Dénes Varga. András Bóka experimented with József Ladányi and Gyuláné Czili with ferrite ring circuits (Maglogal, automatic ferrite sorting machine), and Pál Németh experimented with transfluxors. Various circuit developments were carried out by Károly Szűcs and Ferenc Bányai.

The same paragraph is also found in https://itf.njszt.hu/324rtr4/uploads/2019/08/KGY2_fuggelek.pdf, but with footnotes, including the following references:

20 Bóka András: Négyszöghiszterézisú ferritek mágneses mérései. MTA SzK Tájékoztató 6. szám, 1961. jún., 65–72. old. Bóka András és Ladányi József: Mágneses-félvezetős logikai alapegységrendszer (MAGLOGAL) MTA SzK Tájékoztató 7. szám, 1961. december, 111–126. old.

20 András Bóka: Magnetic measurements of square hysteresis ferrites. MTA SzK Bulletin No. 6, June 1961, pp. 65–72. András Bóka and József Ladányi: Magnetic-Semiconductor Logical Basic Unit System (MAGLOGAL) MTA SzK Bulletin No. 7, December 1961, pp. 111–126.

Meanwhile https://itf.njszt.hu/324rtr4/uploads/50_eves_az_M3.pdf includes the following:

A műszakiak között is voltak, akiknek nem az M-3 körül volt dolguk, így például Hatvany József, szerszámgépek automatizálásához készített jelfogós eszközöket, Németh Pál az M-3-hoz épített volna ferrit memóriát, de a deszkamodellből sohasem lett működő memória, amit a számítógéphez illeszthettünk volna. A memóriához szükséges ferritgyűrűket Bóka András, vezetésével a Vaskutatóban fejlesztették ki és az MTA KKCs-ben készült válogató készülékkel szelektálták. Bókával dolgozott, majd önállósult Ladányi József, aki egy ferritmagos logikai rendszert fejlesztett, később - Maglogal néven - szabadalmaztatott is. A rendszert egy fényújság építésénél alkalmazták. Utána elvesztettem a szemem elől. Ugyancsak velünk együtt dolgozott egy oktatógép kifejlesztésén Balázs Bélának, a Marx Károly Közgazdaságtudományi Egyetem professzorának a vezetésével egy kis csoport – Dauerbach Béla és Jámbor Antal – akik egy filmvetítőből, egy magnetofonbólés egy jelfogós vezérlőből állítottak össze egy oktatógépet. A készülék működött, de – azt hiszem - sohasem került felhasználásra.

There were also technical people who had nothing to do with the M-3, such as József Hatvany, who made signal capture devices for the automation of machine tools. Pál Németh worked on building a ferrite memory for the M-3, but the board model never became a working memory which we could have adapted to the computer. The ferrite rings required for the memory were developed under the leadership of András Bóka at the Vaskutata and selected with a sorting device made at the MTA KKCs [Cybernetics Research Group of the Hungarian Academy of Sciences]. József Ladányi, who worked with Bóka and then became independent, developed a ferrite core logic system and later patented it under the name Maglogal. The system was used in the construction of a light newspaper. I lost sight of it after that. A small group - Béla Dauerbach and Antal Jámbor - led by Béla Balázs, professor at the Károly Marx University of Economics, also worked together with us on the development of an educational machine. The device worked, but - I think - it was never used.

Unfortunately, that's as far as I managed to get with Googling — if the papers from the MTA SzK Tájékoztató mentioned above are available online somewhere, I have not been able to find them.

Thus, all I can say with confidence is that Maglogal was apparently the name of some kind of a ferrite core based computing device developed by József Ladányi and András Bóka at the Hungarian Academy of Sciences in the late 1950s or early 1960s. While it was apparently a real, working system, what it did or how big it was (in terms of physical size or computing capacity) still remains a mystery to me.


Ps. Based on the paper title cited above, tofro's otherwise insightful comment does seem to have been wrong in one respect: the "mag" in "maglogal" does not come from magyar, but from mágneses ("magnetic").*

Specifically, the name appears to be an abbreviation of mágneses-félvezetős logikai alapegységrendszer, which Google translates (accurately, AFAICT) as "magnetic-semiconductor logical base unit system" (alap = "base" + egység = "unit" → alapegység[hu] = "base unit"; rendszer = "system"). Don't ask me what that really means, though — the words all make sense individually and even more or less fit together, but at least to me they reveal very little about what the device actually did.

*) Of course I cannot rule out the possibility that it actually comes from both. It wouldn't be the first or the last abbreviation with a deliberate double meaning, after all.


Update: I managed to find (what I believe to be) the Maglogal patent mentioned in the last translated quote above via the Hungarian Intellectual Property Office patent database. Alas, the database doesn't seem to support direct linking, but the patent in question is #149613 (application LA-488), titled Mágneses logikai alegységek ("Magnetic logic subassemblies") and filed on November 11, 1960, with the rightsholder listed as MTA Számítástechnikai Központ; Budapest (HU) and the inventor as Jószef Ladányi.

Also, it's of course all in Hungarian, but at least the PDF file (Google Drive link) has been OCRed decently well. It's also not very long, so let me quote a translation (by Google translate, with some hand editing) below:

Magnetic logic subassemblies

MTA Computing Center, Budapest
Inventor: technician József Ladányi, Budapest
Date of filing: November 11, 1960.


The development of technical science is moving towards high operational reliability, long service life and standardization of equipment.

All branches of technical cybernetics, including computer technology, follow this path. Logical subunits are the building blocks and mosaics of computer equipment. These subassemblies were initially built from tubes, diodes, and then from tubes, transistors, and diodes. In the last decade, the use of magnetic elements has become more and more popular, precisely because they possess all of the above-mentioned properties. Logic sub-units are created from magnetic elements and diodes, or only from magnetic elements. These subunits, like computers, mostly work in the binary number system and can have one, two, or more inputs and one, two, or more outputs. Generally, they have such a system that if there is a signal it means "1", if there is no signal it means "0". This is not always reliable, because in many cases the interference signal is comparable to the useful signal, so it is difficult to detect it. More reliable and therefore more and more popular is the solution where "0" and "1" signals are of opposite polarity.

The aim of the invention is to develop a reliable, high-speed, inexpensive system.

A possible embodiment of the invention can be seen in the figures marked below:
Figure 1: 1 input,
Figures 2–3: 2 inputs.

Ferrite cores with a square hysteresis curve —a—, —b— form a pair of current pulse deflecting cores. (Fig. 1.) Coil —2—, which adjusts the core pair, forms the input of the sub-unit. The current pulse driving core pair works on mixing transformer —c— (Figure 1). The secondary winding of the mixing transformer —3— is directly connected to the input of the second stage —4—. In the single-input sub-unit, the second stage is completely identical to the first. The logical sub-unit works in two phases. (I—II; Fig. 1.) In the first stage, the first stage transfers the information to the second stage, in the second stage this information appears at the output of the sub-unit, which is none other than the secondary coil —5— of the mixing transformer of the second stage. The secondary can consist of several coils, i.e. the sub-unit can have any output.

The first stage of the two-input sub-unit is different from the one-input one and can have two different solutions. The second stage is the same regardless of the number that can be entered.

A two-input sub-unit can have a solution such that in the first stage of the single-input unit, the mixing transformer is not operated by one pair of diverting cores, but by two (Figure 2). In this way, a subunit with any input can be obtained. Second, it can also be solved by adjusting the two rings of the core pair separately with coils —7— —8—. In this case, the inputs of the two coils form the inputs of the sub-unit (Figure 3). On the mixing transformer of the first stage of each even-numbered input — thus two-input — sub-unit, there is also a blocking coil —6— connected in series with the pair of current pulse diverting magnetic cores —1—. (Fig. 2-3.)

In addition to performing a logic function, the sub-units of the system also perform pulse amplification. The essence of amplification is that a relatively small current pulse is sufficient to set the diverting core pairs, but the diverted current can be many times that amount.

The first stage of the single-input logic sub-unit can therefore be used as a read-write amplifier for a magnetic data storage (Figure 4). The signal read from the database is sent to the input of the sub-unit directly or, if necessary, through a magnetic amplifier —E—. The signal obtained on the secondary of the mixing transformer is suitable for writing back to the board and transmitting it to the machine. New information is entered into the database in such a way that the content of the sub-unit is changed by the new information with a special coil —10— prepared for this purpose. This time takes place between the reading and writing periods.

Patent claim:

Magnetic logic sub-unit and amplifier, characterized by the fact that the output wires of the magnetic core pair of the current pulse driver are connected to the primary windings of the mixing transformer, the secondary winding(s) of the mixing transformer form the output of the sub-unit.

1 drawing:

Figures 1–4 from Hungarian patent #149613 (application LA-488) titled Mágneses logikai alegységek ("Magnetic logic subassemblies") by Jószef Ladányi, filed on November 11, 1960. The figures are schematic drawings of what appear to be wires (represented by thin lines) and ferrite cores (represented by thick lines), with several points and sections of the drawings labelled with letters, numbers and Roman numerals referenced in the patent text. Arrow heads at the ends of lines appear to distinguish output from inputs.

So, if I'm reading it correctly, the patent describes a method of constructing logic gates based on ferrite cores(!). What it doesn't do, however, is describe any kind of computing devices these gates might have been used in.

Also, curiously, while the title of this patent also abbreviates to "maglogal", it is not the same as the title of the 1961 paper cited earlier above (and the abbreviation is never mentioned in the patent).

Perhaps Ladányi first came up with "maglogal" as an abbreviation for mágneses logikai alegység ("magnetic logic subassembly"), and then liked the name enough to reuse it for the computer built out of these things? That would at least explain the weird "base unit system" (alapegységrendszer) part of the name, since he would've had to come up with something that starts with "al" in Hungarian and sort of makes sense the new context.


Update 2: I still haven't managed to track down the 1961 paper, but I did find a few more references to Maglogal, some of which may shed some light on how (a garbled reference to) it ended up in Datamation.

First of all I tried to see if the MTA SzK Tájékoztató might be archived in the REAL-J journal repository of the Hungarian Academy of Sciences. Unfortunately that doesn't seem to be the case — at least all my search attempts came up empty, and the 1961 index doesn't seem to include any journals with similar names at all.

However, the REAL-J full text search did turn up a reference to "Maglogal" in Akadémiai Értesítő / Magyar Tudomány ("Academic Bulletin / Hungarian Science"), 69 (1-12), 1962 (PDF). Alas, it's just a passing mention in a list of "inventions developed in the research institutes of the Hungarian Academy of Sciences":

Ami az egyéb tudományterületeket illeti, az automatizálás fejlődése szempontjából figyelmet érdemel a Számítástechnikai Központban kidolgozott, ún. "Maglogal" találmány, amely az eddigi megállapítások szerint az automatikus irányítás és vezérléstechnika korszerű és nagy üzembiztonságot jelentő eszköze.

As far as other fields of science are concerned, the development of automation deserves attention in the so-called "Maglogal" invention, which, according to the findings so far, is a modern tool for automatic management and control technology with high operational reliability.

(I can't tell how much of the awkward phrasing is due to the machine translation and how much is just the stilted way the original text is written. I suspect a bit of both.)


My next idea was to transliterate the name "Maglogal" into cyrillic and Google for "Маглогал" instead. This indeed turned up several hits, probably the most interesting of these being a PDF copy of a January 1965 issue of the "Aeroflot Engineer" magazine, which includes this short item:

В МИРЕ ИНТЕРЕСНОГО

Патент „Маглогал"

Универсальный логический аппарат «Маглогал» создан в центре вычислительной техники Академии наук Венгрии. Конструктор его – тридцатилетний ученый Йожеф Лакданьи. «Маглогал» составляется из крошечных логических элементов площадью з один квадратный сантиметр зысотой два миллиметра.

Из элегентов «Маглогала» мож чо создать любые кибернетические конструкции. Область их применения практически бесконечна. С гомощью «Маглогала» можно правлять крупными заводами, сотнями машин. Запоминающий центр, построенный из логических элементов, заставляет станки в нужный момент начать работу, произвести необходимые операции, а затем тщательно проверить качество деталей. Производство аппаратов будет налажено в начале будущего года.

(«Известия» № 295 за 1964 г.).

which translates as:

IN THE WORLD OF INTERESTING

Patent "Maglogal"

The universal logical apparatus "Maglogal" was created in the center of computer technology of the Hungarian Academy of Sciences. Its designer is a thirty-year-old scientist Jozsef Lakdanyi [sic!]. "Maglogal" is made up of tiny logical elements with an area of one square centimeter and a height of two millimeters.

Any cybernetic constructions can be created from the elements of "Maglogal". The scope of their application is almost endless. With the power of "Maglogal" one can manage large factories, hundreds of machines. The memory center, built from logical elements, makes the machines start working at the right time, perform the necessary operations, and then carefully check the quality of the parts. Production of the devices will start at the beginning of next year.

(“Izvestia” No. 295, 1964).

Another Google result, http://www.hrono.ru/1964_.php, also features an abbreviated version of the text above, but also gives a date: December 12, 1964. What that's the date of is anyone's guess, but my own guess would be that it's when the Izvestia article was published.

Unfortunately, while there are several online archives of Izvestia, most of them seem to require a university account to view that and none of the freely available ones that I could find have that particular issue. Thus, whether the Izvestia version had any more detail than quoted above remains unknown to me.

In any case, as the official newspaper of the Soviet government, Izvestia was one of the biggest and most influential newspapers of the USSR and the entire Eastern Block, and there were certainly also a lot of people keeping an eye on it in the West. I find it very likely that the Izvestia article, perhaps garbled in translation and transmission, was the source of the Datamation piece.


Pps. Based on the dates, I decided to search the Hungarian patent database for patents by Jószef Ladányi filed near 1964 to see if I could locate "the" Maglogal patent referenced by Izvestia. I did not, at least not by that name, but I did find one (patent #153136, application TA-822, filed on May 28, 1964) titled Transzformátorcsatolású logikai rendszer (Tralogal) ("Transformer-coupled logic system (Tralogal)") which looks a lot closer to an actual computing device than the "magnetic logic subassembly" patent I quoted above:

A hagyományos két vagy toíbb változós logikai rendszeréknél, ha a változókfaal többfajta műveletet kívánunk egyidőben elvégezni, külön-külön ki kell építeni a kapcsolásokat és a változókat egyidőben párhuzamosan kell a kapcsolások megfelelő bemeneteire adni. Ez több egyidőbeli logikai művelet esetén akár diódás, akár tranzisztoros megoldás esetén igen körül ményes eljárás.

Jelen találmány biztosítja, hogy egy meglevő kapcsolás esetén egy további művelet beiktatása csupán két idarab ferritmag elhelyezésével és megfelelő befűzéslével jár. Ha nondestruktiv (tartalmat automatikusan visszaíró) ferritmagot (biax, oldalán átfúrt mag, transzfluor stb.) használunk transzformátorként, akkor valamennyi magon átmennek a vezetékek és kívülről tudjuk, az általunk szükséges tartalmat beírni. Így minden időben átf űzés nélkül tetszőleges tartalmat tudunk beállítani.

In the case of a traditional logic system with two or more variables, if we want to perform several operations with the variable tree at the same time, the connections must be built separately and the variables must be assigned simultaneously to the appropriate inputs of the connections. This is a very delicate procedure in the case of several simultaneous logical operations, either in the case of a diode or a transistor solution.

The present invention ensures that, in the case of an existing connection, the introduction of an additional operation only involves the placement of two pieces of ferrite core and proper threading. If a non-destructive (automatically writes back content) ferrite core (biax, side-drilled core, transfluor, etc.) is used as a transformer, the wires pass through all the cores and we can enter the content we need from the outside. In this way, we can set any content at any time without connecting. […]

Of course, the obvious difference here is the name: "Tralogal" instead of "Maglogal". Were these two different systems, or two different names for the same thing? I have no idea, but they do seem closely related.

(Also, a curious feature is that "Tralogal" doesn't really work as an abbreviation of the patent title — it seems like the part that the "al" supposedly stood for got dropped at some point! Make of that what you will. At least to me it seems to reinforce the idea that the desire for a catchy abbreviation may, to some extent, have driven the choice of the full name rather than vice versa.)

3
  • Lovely research. Cant't wait for more extensions!
    – Raffzahn
    Mar 10, 2023 at 22:43
  • Regarding the "Mag" and "log" in maglogal, it may also be influenced by the fact that logic implemented with ferrite cores was usually called magnetic logic.
    – RETRAC
    Mar 11, 2023 at 14:44
  • @RETRAC: Considering that (apparently, based on the 1961 paper title) the first two parts literally come from the Hungarian words mágneses ("magnetic") and logika ("logic"), that seems likely indeed. The same roots are shared by many languages, though, and it's quite possible that the inventor also had e.g. the Russian words магнитный and логика in mind. Mar 11, 2023 at 15:02
17

Bitsavers also provides the February issue that's referenced by the letter.

The mentioned section does not really give away any additional information - except that the numbers do not make any sense:

"A peep behind the curtain" article

So far I'd file that whole report under "time appropriate joke".


Then again it might as well be the usual mix-up of information presented in a foreign language at a time before jargon was settled. So lets follow the pointers and look at the state of computing in Hungary at the time (late 1964), which does yield some additional information.

The mentioned

  • Hungarian Academy of Sciences (Hungarian: Magyar Tudományos Akadémia, or MTA) does still feature an
  • Institute for Computer Science and Control (Hungarian: Számítástechnikai és Automatizálási Kutatóintézet or MTA SZTAKI). Except, today's incarnation was formed in
  • 1973 by joining MTA's
  • Computing Center (Számítástechnikai Központjával, or MTA SZK for short) and its
  • 1964-founded Automation Research Institute (Automatizálási Kutató Intézete, MTA AKI).

Considering that any news mentioned in the February 1965 issue must at least date from late 1964 (*1) AKI's foundation might have played a role in spreading some PR.

Another notable data point may be the main computer in use at the SZK, the Hungarian M-3, active since 1960.

In 1959 the MTA's Cybernetics Research Group (Kibernetikai Kutató Csoport, MTA KKC) was founded to build the Hungarian M-3 computer. The machine first ran in 1959 and was handed over to the computing center in 1960. It was, despite often noted(*2) otherwise, an implementation of the M3 design developed my Nikolai Matjuchin at the Laboratory for Electrical Systems of the Energy Institute of the USSR Academy of Sciences (Лаборатории электросистем Энергетического института АН СССР) (*3).

The M3 was built as 3 prototypes:

Plus at least one unit built

While intended for serial production, only about a dozen machines with many modifications were built. The design was then used for the Minsk-1 which delivered close to 300 systems starting in 1960.

The Hungarian M3 was moved in 1965 from SZK in Budapest to the Cybernetics Laboratory of the József Attila University in Szeged, while being replaced by a BESM-2 (*5,*6)

It was used in Szeged until 1968 before being dismantled. Some parts, including one drum, made it into an exhibition in Szeged, seen here.

The reason why the this may be relevant here are additions made in Hungary. The machine had at first a dingle drum offering 2048 31-bit words of storage. Not a lot to solve real world tasks, but getting drums from the USSR wasn't easy, so soon KCC/SZT developed their own drum design with 1.6 KiWords storage (*7). While this reduced the storage pressure, it still restricted the machine to about 50 operations per second, so development of a 1 KiW ferrite core was done in 1963 (*8) which resulted in increasing the speed of the very same machine from 50 IPS to 1000 IPS (*9).

And that is what started production of core in Hungary - a possible point to search on from - always with an eye what numbers may have been mixed up on the way from Budapest to Los Angeles.


Also a great reminder that there is a whole world of interesting designs out there that need to be looked at.


*1 - Yes, there was an era before the Internet, where publishing a monthly newspaper did need considerable preparation :))

*2 - Research is a bit encumbered as most internet resources (Wiki maybe a bit less) put a strong emphasis on the M3 being a very domestic development.

*3 - In some aspect the M3 can be seen like the IAS family, a series of computers inspired by the freely distributed plans of the Princeton IAS machine. It includes machines like the Munich PERM, Russia's BESM or IBM's 701. Other than with the M3, resulting machines were usually not compatible with each other.

*4 - Where it was used by Koroljow to turn the R7 into a space launch system, still used in today's incarnation Soyuz-2.

*5 - Looks like Ferranti did not make the intended sale :))

*6 - Which BTW shows that translation does not really remove all hurdles. The Hungarian language uses usually transcription, while in English transliteration, often 1:1, is more common. So (Russian) Cyrillic 'БЭСМ' is not written 'BESM' as in English but 'BESZM' or ' BESzM'.

*7 - That drum type was later as well used for the first domestic development in Romania, the MECIPT-1,

*8 - There was already a production of core memory in Minsk, but as with drums, getting a stack was hard.

*9 - Other sources claim up to 3000 IPS, which I would attribute to peak performance.

15
  • 4
    Your ability in finding the "weird" impresses me. I suppose it comes down to experience and memory.
    – Fred
    Mar 9, 2023 at 8:27
  • 3
    If we assume that each core weighs 100mg, that's... on the order of 90% of earth mass used just for cores. I think mining 90% of earths mass into cores would have been noted, even if it was happening behind the Iron Curtain.
    – vidarlo
    Mar 9, 2023 at 13:37
  • 5
    Valid point. Maybe this is the reason the Iron Curtain ripped? It was mined for use in a computer?
    – vidarlo
    Mar 9, 2023 at 13:45
  • 7
    @vidarlo A computer built using most of the Earth -- wasn't that one of the plot points in the Hitchhiker's Guide to the Galaxy?
    – Barmar
    Mar 9, 2023 at 14:15
  • 3
    @davidbak or the original referred to decimal numbers each with 27 digits, but something was lost- or rather gained- in translation. Mar 10, 2023 at 15:22

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .