Random Programming Language

What the random programming language generator does

You want a language. Maybe you are picking something to learn this month, maybe you are naming servers, maybe you are running a workshop and need to hand somebody a topic they did not choose.

Press the button and you get one.

Most of what comes back you will not recognise. That is not an accident and it is not padding. It is the single most honest thing about this subject, and it is worth understanding before you decide it is a flaw.

How to use it

1. Choose how many you want. One if you are deciding, a batch if you are browsing.
2. Press Generate. A fresh draw every time.
3. Press Copy to take the set with you.

The language written in a village with nothing to run it on

In 1945 a German engineer named Konrad Zuse was living in an Alpine village called Hinterstein, having got himself and his half-built computer out of Berlin while Berlin was being destroyed.

He had already done something remarkable and nobody had noticed. Four years earlier, in his parents' living room, he had finished the Z3, which was the first working programmable digital computer anyone had built. In Berlin. In 1941. While the government funding it had no idea what it was for.

Now, in a village, with a war ending around him, he sat down and did something stranger. He invented a way of writing programs.

He called it Plankalkül, which means plan calculus. And it was not a notation for machine instructions. It had assignments, conditionals, loops, subroutines, arrays, records, and a type system. Anybody who has learned to program in the last fifty years would look at the concepts and recognise almost all of them. He worked on it from around 1942 to 1945, alone, with no colleagues, no journals, no conferences and no correspondence, because there were none of those things to be had.

He wrote chess programs in it.

There was nothing to run them on. Not a shortage of machine time. No machine anywhere on Earth could execute this. He had designed a high-level programming language before the world had a computer capable of accepting one, and he knew it, and he wrote the chess programs anyway.

Then the interesting part, which is the waiting.

He tried to publish. A fragment appeared in a mathematics journal in the late forties and nothing happened, because for a long time after the war the idea that you would program a computer in anything other than machine code was not a controversial idea, it was simply not an idea. There was nothing to argue with. He wrote a whole book about it that went unpublished. The full description did not appear until 1972, twenty-seven years after he wrote it.

The first Plankalkül compiler was built in 1998.

Konrad Zuse died in 1995.

He designed the first high-level programming language, wrote programs in it, and spent fifty years not seeing a single one of them run. Somebody else got to press the button.

The compiler nobody would touch

Now cross the Atlantic and move forward seven years.

Grace Hopper was working on the UNIVAC I, and she had a theory that programmers were doing something stupid. Everybody was writing machine code by hand. Everybody was writing the same subroutines they had written last month, again, from memory, with fresh mistakes in them.

Her stated reason for wanting to fix this was that she was lazy. Her actual argument was that a programmer ought to be able to go back to being a mathematician and let the machine handle the clerical part, which is a claim about what computers are for, and in 1952 it was close to heresy.

So she built the thing that would do the clerical part. She called it the A-0 System. You wrote down which subroutines you wanted and what to give them, and the A-0 went and assembled the actual program for you. By the standards of what a compiler means now it was closer to a linker, and it does not matter, because it proved the thing that needed proving: a computer can be used to write a computer program.

It worked. In 1952 she had it running.

And nobody would use it.

Her own account of what happened is one of the best sentences in the history of the field. She had a working compiler, and nobody would touch it, and the reason they gave her was that computers could only do arithmetic.

Read that again, because it is easy to skim past as a quaint old opinion. It is not quaint. It is a perfectly reasonable position from inside 1952. A computer is a machine for doing sums very quickly. That is what it is. Asking it to read a description of a program and build the program is not a hard version of arithmetic, it is a category error, like asking a calculator to be disappointed in you. The people telling her this were not fools. They were working from a definition, and her compiler was outside the definition.

It took her two years of demonstrations to get management to accept it. She said afterwards that people are allergic to change and that a new idea has to be sold rather than proved.

Then it kept going. A-0 led to A-1, A-2 and A-3. A-3 shipped as ARITH-MATIC. Then came FLOW-MATIC, which was the first one that let you write instructions in words rather than notation, with commands like MULTIPLY and WRITE, and FLOW-MATIC went on to become the basis of COBOL, which is in this list, and which is still running payroll for a frightening proportion of the planet.

So there are two people, on two continents, seven years apart, neither of whom knew what the other was doing, and both of whom had to argue with the same objection: that a machine cannot read words.

Everything after that is detail.

Why there are so many of them

Once the argument was settled, the floodgates went, and this is where the unrecognisable names come from.

A programming language is cheap to invent and enormously expensive to keep. Inventing one is a weekend if you are unambitious and a PhD if you are not. What costs is everything after: a compiler that works on the machines people own, a standard library, documentation, a package ecosystem, a community that answers questions at two in the morning, and thirty years of somebody caring enough to fix it. Almost nobody has that. So almost every language ever made is a corpse.

They are not failures, though, and this is the part people get wrong. Most of them were built to answer one question, they answered it, and they stopped. A language written in a lab in 1968 to see whether an idea would work is not a tragedy if the idea worked and got absorbed into something else. The ideas migrate. The names do not.

That is why the field looks the way it does. Garbage collection was a strange idea in a symbolic-processing language for artificial intelligence research. Now it is in almost everything and nobody credits it. Type inference, pattern matching, closures, generics and iterators all arrived in languages most working programmers have never opened, waited twenty or thirty years, and then turned up in whatever you are writing today wearing different clothes.

So when this hands you a name you have never heard of, you are not looking at a mistake. You are looking at the normal condition. The dozen languages everybody argues about are the exception, and the graveyard is the field.

Where these come from

The languages are drawn from a list our team researched and checked by hand, so the obscure ones are real rather than invented filler. Your browser picks from them at random. Nothing you do here leaves your device.

Questions people ask

So which one was actually first?

It depends on what you count. Plankalkül was designed first, in the mid-1940s, and was not run until 1998. The A-0 System was working in 1952 and got used. Fortran, from 1957, is usually given the title because it was the first high-level language that was implemented, released, and widely adopted, which is three different achievements bundled into one word. All three answers are defensible and they are answers to different questions.

Is COBOL really still running?

Yes, and not as a curiosity. It sits underneath a great deal of banking, insurance and government payroll, and it is there because those systems work and rewriting them is a genuinely terrifying proposition. It is a direct descendant of Grace Hopper's FLOW-MATIC, which means the language keeping the world's payroll running traces back to somebody in 1952 who was told computers could only do arithmetic.

Why are there so many I have never heard of?

Because that is what the field is. Languages are easy to make and hard to keep alive, so nearly all of them stop, usually after answering the one question they were built to answer. The handful you have heard of are survivors, and survivorship is a strange thing to mistake for the whole population.

Is a thing like XSLT or SQL a programming language?

People argue about this constantly and the argument is usually about the word rather than the thing. The old test was whether a language can compute anything computable, which lets in a lot of oddities and keeps out some genuinely useful tools. A more practical test is whether people write, debug and maintain programs in it, and by that test the answer is usually yes, whatever the purists say.

Which should I actually learn?

Whichever one the people around you are using, if you are starting. Nearly everything transfers. The concepts that matter, the ones described above, drift between languages on a thirty year cycle, and learning your second language is roughly a tenth of the work of learning your first.

Can the same language come up twice?

Not within a single draw. Ask for ten and you get ten different ones. Press again and everything is back in play.

References

1. Konrad Zuse's Plankalkül: The First High-Level, "non von Neumann" Programming Language. IEEE Annals of the History of Computing, 19(2), 1997. https://doi.org/10.1109/85.586068
2. Plankalkül: The First High-Level Programming Language and its Implementation. Freie Universität Berlin. https://www.researchgate.net/publication/250809396_Plankalkul_The_First_High-Level_Programming_Language_and_its_Implementation
3. O'Connor, J. J. and Robertson, E. F. Grace Brewster Murray Hopper. MacTutor History of Mathematics Archive, University of St Andrews. https://mathshistory.st-andrews.ac.uk/Biographies/Hopper/
4. A-0 Compiler and Initial Development of Automatic Programming, 1951-1952. Engineering and Technology History Wiki, IEEE. https://ethw.org/Milestones:A-0_Compiler_and_Initial_Development_of_Automatic_Programming,_1951-1952