12 ms·
Alonzo Church: Architect of computer intelligence
- cubefox 2y agoEspecially his philosophy of logic and sense/reference (a continuation of the work of Frege and Russell) is mostly forgotten. He published many papers on this topic, yet they aren't discussed e.g. in Wikipedia. At least the SEP entry is better: https://plato.stanford.edu/entries/church/ https://plato.stanford.edu/entries/church/ Though I heard that it also neglects some major parts of his work. I assume it was too philosophical for mathematicians and too technical for philosophers.
- magical_spell 2y agoRelated: E.J. Lemmon, in his book Beginning Logic, lists some important logic books and says that Chapter 0 of Church's Introduction to Mathematical logic ``deserves to be read several times by /all/ philosophers''.
- gpderetta 2y ago"Church’s lambda calculus and the Turing machine are equally powerful but differ in the fact that Turing machines use mutable state. To this day, there is a rift between functional and imperative programming languages, because of the separation of Church and state." [I have known the above quote forever, but I can't find an original source] edit: might be from Guy Steele: "And some people prefer not to commingle the functional, lambda-calculus part of a language with the parts that do side effects. It seems they believe in the separation of Church and state"
- drcwpl 2y agoThat is such a great quote - classic programming humor, but no idea of the original
- rbonvall 2y agoJust like Niklaus Wirth's quote about how people used to call him, or the joke about there being 10 kinds of people. Those are the ones that make me wish people knew just enough Computer Science to get them :)
- tialaramex 2y ago"There's two hard problems in computer science: we only have one joke and it's not funny" which I've seen credited to Phillip Scott Bowden Which is a reference to the "two hard problems" jokes, the most used is "There are two hard problems in Computer Science: Cache invalidation, naming things, and off-by-one errors" But there is also "Two hard problems in distributed systems: Exactly-once delivery, guaranteed order of messages, and exactly-once delivery".
- rramadass 2y agoI need more of these ...
- fuzztester 2y agohttps://news.ycombinator.com/item?id=42054476 https://news.ycombinator.com/item?id=42054476
- Swannie 2y ago> Which is a reference to the "two hard problems" jokes, the most used is "There are two hard problems in Computer Science: Cache invalidation, naming things, and off-by-one errors" It's just as funny today as when I first heard it in 2002/2003... to which my professor at the time would add: "But really, naming things is colouring. And cache invalidation is colouring. So really there's only one problem in computer science, colouring and counting... which doesn't sound so hard after all".
- fuzztester 2y agoI made a UDP joke once. You may get it.
- f1shy 2y agoI think it is from Peter Norvig, look the sister comment.
- tjr 2y agoGuy's quote there came from a mailing list at MIT that was formed out of the 2001 Lightweight Languages Workshop. Archive of the original post here: https://people.csail.mit.edu/gregs/ll1-discuss-archive-html/msg01134.html https://people.csail.mit.edu/gregs/ll1-discuss-archive-html/...
- gpderetta 2y agoPerfect! Thanks for the source.
- fuzztester 2y ago>seems they believe in the separation of Church and state" ha, good one. remind me of the joke about niklaus wirth's name. [ Quotes about Niklaus Wirth Whereas Europeans generally pronounce his name the right way ('Nick-louse Veert'), Americans invariably mangle it into 'Nickel's Worth.' This is to say that Europeans call him by name, but Americans call him by value. Introduction by Adriaan van Wijngaarden at the IFIP Congress (1965). ] from: https://en.m.wikiquote.org/wiki/Niklaus_Wirth https://en.m.wikiquote.org/wiki/Niklaus_Wirth
- inopinatus 2y agoSo sayeth the wise Alonzo.
- sourcepluck 2y agoA historical tidbit which I loved in Paradigms of Artificial Intelligence Programming (available in PDF and EPUB here - https://github.com/norvig/paip-lisp https://github.com/norvig/paip-lisp): > The name lambda comes from the mathematician Alonzo Church's notation for functions (Church 1941). Lisp usually prefers expressive names over terse Greek letters, but lambda is an exception. A better name would be make-function. Lambda derives from the notation in Russell and Whitehead's Principia Mathematica, which used a caret over bound variables: x̂(x + x). Church wanted a one-dimensional string, so he moved the caret in front: ^x(x + x). The caret looked funny with nothing below it, so Church switched to the closest thing, an uppercase lambda, Λx(x + x) . The Λ was easily confused with other symbols, so eventually the lowercase lambda was substituted: λx(x + x). John McCarthy was a student of Church's at Princeton, so when McCarthy invented Lisp in 1958, he adopted the lambda notation. There were no Greek letters on the keypunches of that era, so McCarthy used (lambda (x) (+ x x)), and it has survived to this day. So, yes, on the topic of this post - Church pops up in loads of Lisp retrospectives. Maybe he's "forgotten" by people with very little engagement in the history of computing.
- drcwpl 2y agoGreat tidbit, (thanks for the Paradigms share) - in the footnote I mention about CS folks awareness.
- agumonkey 2y agoBut wait, who ever first coined the term 'lambda calculus' ? was it before or after McCarthy started lisp ?
- seanhunter 2y agoWell before. The original paper[1] introducing the lambda calculus was in the 1930s, but it wasn't called "lambda calculus" until a bit later. [1] https://raw.githubusercontent.com/emintham/Papers/master/Church-%20A%20Set%20of%20Postulates%20for%20the%20Foundation%20of%20Logic.pdf https://raw.githubusercontent.com/emintham/Papers/master/Chu...
- tromp 2y agoThe Alonzo programming language named after him [1] has mostly been forgotten. [1] https://dl.acm.org/doi/pdf/10.1145/68127.68139 https://dl.acm.org/doi/pdf/10.1145/68127.68139
- breck 2y ago[dead]
- dang 2y ago[stub for offtopicness] Side remark: "Please don't take the bait" is a good analogue to "please don't feed the trolls". We've taken the bait out of the title now, but when a thread has already filled up with comments reacting to it, rather than anything interesting, that's bad for HN threads.
- seanhunter 2y ago"Forgotten" in the sense that everyone who knows anything about CS knows who he is because of the Church-Turing hypothesis, Church numerals, lambda calculus etc, and anyone who reads "On Computable Numbers" (only the most famous paper in all of computer science) knows that Turing actually quotes and credits Church and his work in that paper. Here's a link to "On Computable Numbers" for easy reference for anyone who wants to read it/read it again. It's a cracker https://www.cs.virginia.edu/~robins/Turing_Paper_1936.pdf https://www.cs.virginia.edu/~robins/Turing_Paper_1936.pdf
- abstractbeliefs 2y agoForgotten not in the sense of lost knowledge, but more that the individual is not known proportionally to the importance of his work, or perhaps consistently when compared to his peers. While specialists in his field know his work and his name (but not even everyone in software does), the public do not. While your parents and friends see the dramatised exploits of Turing in films like The Imitation Game, or his face on the currency, the same is not said for Church. Every field has it's public heroes, usually related to the stature of their work - Fleming and Jenner, Marconi, Ford, Bell. Turing. Anyone will at least recognise these names, but not so for Church.
- gilleain 2y agoOk, and if I asked a random member of the public for the name of a mathematician (excepting Turing, for clarity) what name do you think they would come up with? Pythagoras? Euler? Erdős? I think the reality is that only a very small number of scientists, mathematicians, and similar are household names.
- dang 2y agoIf you want to read something incredible about Church, read Rota's reminiscence. It's the first section of https://www34.homepage.villanova.edu/robert.jantzen/princeton_math/pmcxrota.htm https://www34.homepage.villanova.edu/robert.jantzen/princeto.... Related: Alonzo Church, 92, Theorist of the Limits of Mathematics(1995) - https://news.ycombinator.com/item?id=12240815 https://news.ycombinator.com/item?id=12240815 - Aug 2016 (1 comment) Gian-Carlo Rota on Alonzo Church (2008) - https://news.ycombinator.com/item?id=9073466 https://news.ycombinator.com/item?id=9073466 - Feb 2015 (2 comments)
- drcwpl 2y agoThat is brilliant, thank you - I added the Rota one as an addendum under the further reading at the end.
- devindotcom 2y agoThis was entertaining, thanks for the link.
- svat 2y agoRota's reminiscence (not just the part about Church but the entire webpage, namely "Fine Hall in its golden age: Remembrances of Princeton in the early fifties") is, as the asterisked footnote alludes to, a chapter of his book Indiscrete Thoughts. The whole book is great reading.
- sourcepluck 2y agoVery enjoyable reading! I've noted the book too now, and hope to get to it.
- tomgp 2y agoNot really the point BUT... I really wish people would refrain from using AI illustrations on blog posts. There are actual pictures of Church in the public domain, this illustration doesn't particularly look like him and is already appearing in image search results for the man thanks to the blog post's popularity. If the illustration has so little value that you can't be bothered to spend more than 5 minutes generating it then perhaps just leave it out? [edit] ... and if you really must use "AI" generated imagery then at least caption it as such.
- Robotenomics 2y agoThank you. Noted and sorry - I was loathe to take an online photo … this one was actually 7th effort as I did not want the ‘fake’ likeness, I felt it captured a resemblance… I managed a good one of JvN but you are right to address this. I think in future I will use a symbol rather than a ‘character (un) likeness ’
- priprimer 2y agothe real challenge of understanding lamba calculus is realizing its simplicity it is a sacred idea by the fact that executing it does not really help understand the fact that it is equivalent to any and all computation
- mrkeen 2y agoI don't know. You don't have to do much bootstrapping to get from lambda calculus primitives all the way up to if-elses, pattern matching, higher-order functions, etc.
- beepbooptheory 2y agoI can't fully argue for this, but my gut says that Turing and all he represents is ultimately valorized by AI-stuff, but Church is the opposite. The former started from the point of view of purity, minimum possible conditions/possibilities, and abstraction/"pure" computation; the latter cared about ways we could actually think, and cared not for implementation, but articulation and the furthering of abstraction.
- drcwpl 2y agoThis is a strong point, "the latter cared about ways we could actually think, and cared not for implementation, but articulation and the furthering of abstraction." I will look into this more. Thank you
- WorldMaker 2y agoOne point of view for this is that Turing built practical computers during the war effort and then was forbidden by his own government to continue to build computers so had to retreat to theory. Church had no practical experience with computers and was looking as much to expand the theory of Mathematics itself. In their collaborations together and communications across an ocean they married a lot of the practical and the theory and solidified core parts of the theory (imperative/functional dualism [the Church-Turing Theorem often shortened to the Turing Theorem ignoring the contributions of both]; the Halting Problem [and its relationship to Church's Theorem]; more). I think it's wrong to see it as a competition, because their collaboration did so much together. Computer Science has "two dads" and that feels appropriate for several reasons, including how Turing died. It is somewhat important to Computer Science that Church and Turing met in the middle, one with the deep raw mathematical theory background and the other with the practical (but at the time unacknowledged) background. Also, it would be somewhat remiss to ignore that Turing did care about implementation and wanted to get back to practical implementation but wasn't allowed. There's a deep tragedy there and lot of questions about what would have been different if things hadn't been classified in the same way by the UK government. Though also in that case perhaps it would have cost us the Church collaboration that helped firm up the theory so well in our timeline.
- wodenokoto 2y agoFirst edition of Probabalistic Models, uses a lisp-like language called Church as a teaching tool. Obviously named after Alonzo Church. http://v1.probmods.org/ http://v1.probmods.org/
- sourcepluck 2y agoExcellent! Looks like a very cool book, and this comment very nearly passed me by. Thank you for sharing!
- sourcepluck 2y agoI'm going for a browse of the book now and realise they've a newer second edition, for anyone interested, it's here: https://probmods.org/ https://probmods.org/
- wodenokoto 2y agoWhich doesn’t use church. Second edition uses a more JavaScript like language.
- ogogmad 2y ago"Architect of computer intelligence" - I'm sure Church was an accomplished logician, but his contribution to "computer intelligence", which I imagine is supposed to mean AI/ML, is absolutely nil. On a separate note, I don't think lambda calculus is really maths. It's more like clever notation. The merits of any notation are subjective. It's interesting how Church never cared about how that particular idea of his inspired the design of certain programming languages - which are ultimately notations, and therefore subjective in their merits.
- cubefox 2y ago"Lambda calculus" is sometimes used to mean "simply typed lambda calculus", which is mainly used to refer to simple type theory (STT), which is also called "Church's type theory". STT is also often identified with higher-order logic, since just two primitive types (basic "entities" like chairs and the "truth values" T/F) and the function type (a --> b) suffice to express arbitrary logical objects. STT is indeed an invention of Church, and it has had a major influence on other modern type theories, which are often described as extensions of simple type theory. These also influenced some programming languages with complex type system, like Haskell.
- ogogmad 2y agoOh yes, that's fair. I was thinking about that, but somehow dismissed it.
- Upvoter33 2y agoOne of Church's big contributions was his students; an amazing collection of thinkers coming out of one place...
- neonscribe 2y agoI was lucky enough to meet Alonzo Church and Haskell Curry at the ACM Symposium on LISP and Functional Programming at CMU in August 1982. Curry was clearly not well and only lived about two weeks after the conference, but Church was in good form and lived about 13 more years. Gerry Sussman was very excited while he went around the room at the reception introducing them, and of course it was a great thrill for us to meet them.