Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Transpiler, a Meaningless Word (2023)

people.csail.mit.edu

Post reply on HN
Print view View original

41–50 of 117 posts

Re: Transpiler, a Meaningless Word (2023)

#41

Post by kragen »

lesam wrote: post #38
mjburgess wrote: post #28

Earlier quoted context omitted.

'computer' is an ambiguous word. In a mathematical sense a computational process is just any which can be described as a function from the naturals to naturals. Ie., any discrete function. This includes a vast array of processes. A programmable computer is a physical device which has input states which can be deterministicaly set, and reliably produce output states. A digital computer is one whose state transition is…

In fact ‘computer’ used to be a job description: a person who computes.

Yes, definitely. And "nice" used to mean "insignificant". But they don't have those meanings now.

Re: Transpiler, a Meaningless Word (2023)

#43

Post by NuclearPM »

kragen wrote: post #25
oersted wrote: post #21

I don't understand what the issue is: a transpiler is a compiler that outputs in a language that human programmers use. It's good to be aware of that from an engineering standpoint, because the host language will have significantly different limitations, interoperability and ecosystem, compared to regular binary or some VM byte-code. Also, I believe that they are meaningfully different in terms of compiler architectu…

GCC outputs textual GNU assembly language, in which I have written, for example, a web server, a Tetris game, a Forth interpreter, and an interpreter for an object-oriented language with pattern-matching. Perhaps you are under the illusion that I am not a human programmer because this is some kind of superhuman feat, but to me it seems easier than programming in high-level languages. It just takes longer. I think tha…

You’re being \__

Re: Transpiler, a Meaningless Word (2023)

#45

Post by lo_zamoyski »

kragen wrote: post #35
mjburgess wrote: post #34

Earlier quoted context omitted.

Ah well, that's true -- so we can be more specific: discrete, discrete computable, and so on. But to the overall point, this kind of reply is exactly why I don't think this is a case of L vs. S -- your reply just forces a concession to my definition, because I am just wrong about the property I was purporting to capture. With all the right joint-carving properties to hand, there is a very clear matrix and hierarchy o…

Word definitions are arbitrary social constructs, so they can't really be correct or incorrect, just popular or unpopular. Your suggested definitions do not reflect current popular usage of the word "computer" anywhere I'm familiar with, which is roughly "Turing-complete digital device that isn't a cellphone, tablet, video game console, or pocket calculator". This is a definition with major ontological problems, incl…

> Word definitions are arbitrary social constructs, so they can't really be correct or incorrect, just popular or unpopular.

If you mean that classifications are a matter of convention and utility, then that can be the case, but it isn’t always and can’t be entirely. Classifications of utility presuppose objective features and thus the possibility of classification. How else could something be said to be useful?

Where paradigmatic artifacts are concerned, we are dealing with classifications that join human use with objective features. A computer understood as a physical device used for the purpose of computing presupposes a human use of that physical thing “computer-wise”, that is to say objectively, no physical device per se is a computer, because nothing inherent in the thing is computing (what Searle called “observer relative“). But the physical machine is objectively something which is to say ultimately a collection of physical elements of certain kinds operating on one another in a manner that affords a computational use.

We may compare paradigmatic artifacts with natural kinds, which do have an objective identity. For instance, human beings may be classified according to an ontological genus and an ontological specific difference such as “rational animal“.

Now, we may dispute certain definitions, but the point is that if reality is intelligible–something presupposed by science and by our discussion here at the risk of otherwise falling into incoherence–that means concepts reflect reality, and since concepts are general, we already have the basis for classification.

Re: Transpiler, a Meaningless Word (2023)

#46

Post by cyco130 »

jasode wrote: post #16

Whenever someone argues the uselessness or redundancy of a particular word, a helpful framework to understand their perspective is "Lumpers vs Splitters" : https://en.wikipedia.org/wiki/Lumpers_and_splitters An extreme caricature example of a "lumper" would just use the word "computer" to label all Turing Complete devices with logic gates. In that mindset, having a bunch of different words like "mainframe" , "pc" , "…

It's all about context, isn't it? "Humans vs. animals" is an important distinction to make in some contexts and useless in others. Insisting on the fact that humans are also animals if we're talking about, say, "language in humans vs. animals" is unproductive. It just makes discussions harder by forcing everyone to add "_non-human_ animals" to every mention. But if we're talking about, say, cellular biology, it's unproductive to force everyone to write "human and animal cells" instead of just "animal cells".

Similarly, distinguishing between transpilers and compilers might be important in some contexts and useless in others. Transpilers are source-to-source compilers, a subset of compilers. Whether it matters depends on the context.

Re: Transpiler, a Meaningless Word (2023)

#47

Post by Joker_vD »

JonChesterfield wrote: post #23
fulafel wrote: post #7

It would be good if we had a term that didn't confuse linking with translation. In English compiling means joining together many parts, after all.

That one is historically interesting. I suspect the first compilers were named that because they were making compilations of assembly routines, probably slightly modified/specialised to the rest of the routines. Compilers still do that. Some of the input is your source, but there's also "the compiler runtime" which is essentially a lot of extra routines that get spliced in, and probably "the language runtime" which g…

The first compilers were called "translators". The first linker/loader (kinda, A-0 was a... strange tool, by modern standards) was actually called "compiler", precisely because of the generic meaning of the word "compile".

Re: Transpiler, a Meaningless Word (2023)

#48

Post by crazygringo »

kragen wrote: post #25
oersted wrote: post #21

I don't understand what the issue is: a transpiler is a compiler that outputs in a language that human programmers use. It's good to be aware of that from an engineering standpoint, because the host language will have significantly different limitations, interoperability and ecosystem, compared to regular binary or some VM byte-code. Also, I believe that they are meaningfully different in terms of compiler architectu…

GCC outputs textual GNU assembly language, in which I have written, for example, a web server, a Tetris game, a Forth interpreter, and an interpreter for an object-oriented language with pattern-matching. Perhaps you are under the illusion that I am not a human programmer because this is some kind of superhuman feat, but to me it seems easier than programming in high-level languages. It just takes longer. I think tha…

https://news.ycombinator.com/item?id=45912557

Re: Transpiler, a Meaningless Word (2023)

#49

Post by JoBrad »

kragen wrote: post #33

The link to Lindsey Kuper's post https://decomposition.al/blog/2017/07/30/what-do-people-mean... is great! I think the note about generators may be a good definition for when one language is "more powerful" than another; at least it's a good heuristic: > The input and output languages have the syntax of JavaScript but the fact that compiling one feature [generators] requires a whole program transformation gives away…

Thanks for the last link! At first read, the regeneration code is nuts: using a switch to assign a value, then comparing hard coded values. I only used generator functions in TS after they were supported in JS, so I’m going to step through that, just to understand it more.

Re: Transpiler, a Meaningless Word (2023)

#50

Post by kragen »

lo_zamoyski wrote: post #45
kragen wrote: post #35

Earlier quoted context omitted.

Word definitions are arbitrary social constructs, so they can't really be correct or incorrect, just popular or unpopular. Your suggested definitions do not reflect current popular usage of the word "computer" anywhere I'm familiar with, which is roughly "Turing-complete digital device that isn't a cellphone, tablet, video game console, or pocket calculator". This is a definition with major ontological problems, incl…

> Word definitions are arbitrary social constructs, so they can't really be correct or incorrect, just popular or unpopular. If you mean that classifications are a matter of convention and utility, then that can be the case, but it isn’t always and can’t be entirely. Classifications of utility presuppose objective features and thus the possibility of classification. How else could something be said to be useful? Wher…

No, I don't mean that classifications are a matter of convention and utility, just word definitions. I think that some classifications can be better or worse, precisely because concepts can reflect reality well or poorly. That's why I said that the currently popular definition of "computer" has ontological problems.

I'm not sure that your definition helps capture what people mean by "computer" or helps us approach a more ontologically coherent definition either. If, by words like "computing" and "computation", you mean things like "what computers do", it's almost entirely circular, except for your introduction of observer-relativity. (Which is an interesting question of its own—perhaps the turbulence at the base of Niagara Falls this morning could be correctly interpreted as finding a proof of the Riemann Hypothesis, if we knew what features to pay attention to.)

But, if you mean things like "numerical calculation", most of the time that people are using computers, they are not using them for numerical calculation or anything similar; they are using them to store, retrieve, transmit, and search data, and if anything the programmers think of as numerical is happening at all, it's entirely subordinate to that higher purpose, things like array indexing. (Which is again observer-relative—you can think of array indexing as integer arithmetic mod 2⁶⁴, but you can also model it purely in terms of propositional logic.)

And I think that's one of the biggest pitfalls in the "computer" terminology: it puts the focus on relatively minor applications like accounting, 3-D rendering, and LLM inference, rather than on either the machine's Protean or universal nature or the purposes to which it is normally put. (This is a separate pitfall from random and arbitrary exclusions like cellphones and game consoles.)

Post reply on HN