Talk:Supercombinator

This Page Is Wrong

The definition of supercombinator is totally wrong!!!!!!! What is very informally defined in the article refers to a combinator, without the super! The super- prefix in computer science in things like this subject, refers to actually inexistent languages, or computation systems. You may talk about Super-Turing machines, which refers to a machine capable to perform computations that Turnig machines can not perform, i.e. a super set of what a Turing machine can compute. Combinatory logic is equally powerfull than Turing machines (and Lambda calculus, recursive functions, etc.), a super combinator refers to something like a super-Turing machine.

The confusion of the use of super-combinator in this article, comes from the use that Simon Peyton Jones, (a definite expert in this subject!!!) gives to a combinator basis in the context of functional programming languages compilation, the book cited in this article.

Bringing that term out of context may confuse with what I told above which is what a philosopher of computation may think about in first instance. This article is tagged in the subject of philosophy. The use that the well known great computer scientist Simon Peyton Jones gives to this term should be included in an article about functional programming languages compilation.

I am not going to fix this article, because I had a very bad experience with OTHER pages, monopolized by ignorants which react to corrections thinking they know everything because it is in their high school notes. I do not have time to earn privileges to monopolize wikipedia and active alarms to be aware of every change, having no other thing to do than revert every change. Let me emphasize that this had happen to me in OTHER pages, but seems an extended practice that should stop!

I do not know if this could be the case, but in my previous experience, suppose that I fix this article, I donate my time to write something better and well documented. After that, the change is notified to the original writer and my changes are reverted, after all I am not the guy who wrote the book that was misinterpreted by the one who posted the article. I could insist, but I am blamed of vandalism. I DO NOT SAY THAT THIS HAPPENS IN THIS PAGE, but I had happen to me in similar cases, for minor changes. I say this here because I think that this should be stopped.

If someone wants to take the risk of fixing this page. It should contain a story of the intents to create super-combinators which is related to super-Turing machines super-recursion, etc. There are some intents published but all are wrong, got it? Good luck! — Preceding unsigned comment added by 189.178.40.254 (talk) 04:17, 20 November 2015‎

How does this match the definition at https://wiki.haskell.org/Super_combinator ? Are they also wrong to talk about supercombinators? — Carl (CBM · talk) 11:34, 20 November 2015 (UTC)

I can't find any reference to anyone using the term supercombinator in the sense described by the OP (i.e., to refer to a class of expressions that are "more powerful" than the class of combinators). I only find it used in the sense given by the referenced textbook by Jones (who attributes its original use to Hughes, 1984). However, I agree that the definition given in this wikipedia page isn't very good. The informal definition as "fully bound and self-contained" could easily be mistaken as the informal definition of a (non-super) "combinator". The definition given in "mathematical terms" is mostly okay, but why not excerpt the precise definition from the referenced text, instead of trying to paraphrase (resulting in the rather bizarre syntax "A lambda expression S is a ..." as used on this page making it look as if we're trying to define lambda expressions, instead of "A supercombinator is a lambda expression of form ..." as is standard for mathematical definitions and which is the phrasing given in the referenced textbook)? In fact, it appears that the much clearer textbook definition was in the original page and was included until it was (in my opinion badly) paraphrased in October 2007. — KABuhr (talk) 19:16, 3 May 2020 (UTC)

I am coming from https://fr.wikipedia.org/w/index.php?title=Logique_combinatoire&oldid=223860207#cite_note-6 which got me curious about what could be a 'supercombinator'.

The philosophical usage of 'super' as prefix explained by OP seems to ring a bell but I wouldn't have thought about it first.

This discussion make me searched and found usage of this philosophical prefix here https://groups.cs.umass.edu/binds/super-turing-computation/ and https://www.alanturing.net/turing_archive/pages/pub/super/super.pdf as the intended meaning by OP. I also came across before coming here on the the usage of ITTM (Infinite Time Turing Machine) such as https://www.jstor.org/stable/pdf/2586556.pdf which are more "powerful" in the same sense as the "super" I believe.

It seems like https://wiki.haskell.org/Super_combinator appeared closer to the creation of this page, respectively 20:18, 5 October 2006‎ and 00:13, 8 February 2005. I can't exclude the possibility that this mistake spread from here.

According to the reference cited, I do find the definition of a Supercombinator as a construction based on a combinator. It refer to Hughes 1984 for the original term :

>> "The combinator  technique of section 3.5 is simple, and has the advantage that compilation  into those combinators is easy. Turner's method, on the other hand, requires a more complex compiler and breaks the execution down into very small steps: an application of S. .... [t]he final section presents the results of an experimental comparison of our  super-combinators with Turner's approach.".

This seems to contradict the fact that supercombinator are "moins bavarde que la logique combinatoire fondée sur S et K.", the footnote above, which led me here.

Hopefully the ambiguity could be resolved.

--

Hughe 1984 The Design and Implementation Programming language https://www.cs.ox.ac.uk/publications/publication3788-abstract.html Oversleep9 (talk) 12:30, 6 July 2025 (UTC)