I found the regular expression package as a first thing to make in SML waaay too complex. You already need to have a good understanding of the elements of SML syntax to understand how it works. This is definitely not for SML beginners. I recommend "Elements of ML Programming" by Jeffrey D. Ullman for people starting to learn SML. It too has some corners, that are less great for not so mathematically inclined people, but you will definitely learn the language and have exercises.
This is valid criticism, and I agree. The regular expression example is not made easier by his use of non-standard regex syntax.
I have not read Dr. Harper's book cover to cover; I treat it as a reference. For that, it is quite useful.
The same criticism of his SML book could also be made of his Practical Foundations for Programming Languages book. It's a massive, dense tome that is probably great for other professors to use as the basis for a course, but (in my opinion) serves as a mediocre introduction if you have never seen the material before.
That's how I felt when I discovered FP after more than two decades writing procedural and OO code. It felt like I'd found a secret room where all the reasoning was kept.
Have a look at this similar course: https://www.coursera.org/learn/programming-languages
I enjoyed it quite a bit and the lecturer is fantastic. It also uses SML and teaches similar material, though maybe not as in depth since this is just the first section of a three-part semester long course.
Some assembly required; use the instructions from https://github.com/jacobneu/alleycat as inspiration, use the scripts (e.g. startLecture) and Makefile in /crucible
Unfortunately, it does not. These lectures are "mine", in the sense that I developed all of them myself, but the homeworks and lab exercises are the combined efforts of generations of TAs and instructors from the past. It wouldn't be right for me to give them away. (they are also reused from time to time, so there are academic integrity concerns with that also)
(but I have thought of developing my own exercises independently to go with the lectures, to post on my website. This is generally a lot of work, though, so this might take some time, depending on how much people would benefit from it.)
TIL. Given that 15-213 has been widely available for years I naively assumed this would also hold true for other undergrad CS courses, but apparently not.
Being a cranky shithead and scrolling through HN. Name a more iconic duo. Seriously though, if your first reaction to a neat source of information is to shit on it with snark and thinly veiled assholerry, please leave the internet for a week. There are better ways to give constructive feedback
In some ways I can only agree, having criticised others for apparently doing the same thing. I guess the difference is that it was completely bloody unnecessary to waste enormous amounts of bandwidth for the cost of just a little bit of basic checking. For me individually, bandwidth is virtually free, for others it is metered (not everywhere has Western-levels of Internetz). But bandwidth is not free – multiply up 49.2 MB by a number of expected page visits.
So, as a matter of principle we need to stop acting as if resources are free. Also if the author couldn't even get that simple HTML detail right then it potentially says something about the quality of his site's information. Perhaps.
> Also if the author couldn't even get that simple HTML detail right then it potentially says something about the quality of his site's information. Perhaps.
Perhaps is doing some heavy lifting here. I take it from the fact that the lectures are on Functional Programming that OP might not be a front-end developer and might not have the level of care needed for front end to compress the images in a way that renders nicely. So,
Perhaps you ought not judge someone releasing their free teaching materials for a backend class on their front end development, when the site is a centered <div> with purely static elements.
Perhaps you ought to slow down and do some critical thinking before being critical of others.
I'm a back-end dev (inc. FP) and RDBMS guy and even I'd know not to do this; instead to pre-shrink images presented as decorative thumbnails. But let's be constructive, what form should my critical thinking have taken, how should I have done better (edit: or criticised more carefully)? Serious question. Thanks.
As a matter of principle, those in bandwidth-constrained situations should be browsing via proxies or browsers that automatically compress, such as Opera Mobile back in the day, or Google’s more modern compression baked in to Android/Chrome. (If it still exists, it’s been awhile…) That said… yeah, it couldn’t hurt to put a free CloudFlare CDN in front with automatic image optimization, for example.
> "it was completely bloody unnecessary to waste enormous amounts of bandwidth"
Thats not enough. It was completely bloody unnecessary to waste enormous amounts of co2.
And critical thinking should lead to the realization that not many people have to solve the - albeit small - mistake of an individual via Cloudflare or similar, but rather someone should do it correctly right away. And if he has no knowledge of frontend, he should not do things he has no knowledge of.
It looks like their current workflow keeps exams and homeworks off the internet effectively, but there's a 6-year-old codebase at https://github.com/zhengguan/15150-1 with 10-year-old homeworks and such.
Rewriting standard list functions (map, fold, sum, etc.) is a good entry-level exercise.
A λ-calculus interpreter can be used as an intermediate level exercise. It is in particularly valuable in the context of solidifying one's understanding of functional programming.
You can also use "standard" textbooks, such as the SICP [0], and perform the exercises using the language of your choice, instead of Scheme/LISP.
Great resource! Forgive my ignorance but why do so many modern functional programming courses use Standard ML instead of a Lisp dialect? Is it because of its built-in type-checking, or is it just how it's always been taught?
The value of purely functional programming languages, as opposed to functional programming languages like lisps, is that you get referential transparency, which means that when you define `a = b`, you know that you can always replace any instance of `a` with `b` and get the same answer. This is a very natural property in mathematics (algebraic rewritings are basically just this property writ large) and so it helps to draw nice parallels between the familiar notation of functions from mathematics and the "new" and "confusing" notion of functions in functional programming and other declarative languages.
As other posters have said, strong typing is also a nice property for lots of reasons, most notably it gives a platform to talk about ad-hoc and parametric polymorphism.
(I lecture on Functional Programming at the University of Warwick, where we use Haskell.)
The first problem with this argument is that referential transparency is a property of syntactic positions, not of languages.
The second is that languages like Lisp, SML, C, Pascal and BASIC all have referentially transparent and referentially opaque positions in exactly the same way that languages like Haskell do.
This means that all these languages enjoy referential transparency in the same way, because when you unpack the notion of equivalence, referential transparency itself is within a whisker of being a tautology: if a is equivalent to b, then you can substitute a for b or b for a. The relevant sense for "is equivalent to" can really only be contextual equivalence, which is all about meaning-preserving substitutability.
That said, not having to reason about effects within one's program equivalence sure makes things simpler in a pedagogical setting. But that's not to do with referential transparency per se.
For those of us who are unfamiliar with Lisps, can you expand on how they break referential transparency (and how Standard ML contrasts in that regard)?
more importantly there are functions (using scheme as an example) like set! and set-cdr! that mutate existing values and totally break referential transparency.
this isn't just user facing - for example let* kind of depends on creating bindings up front so they work across clauses, and then mutating them afterwards
let* permits expressions on the right refer to arbitrary other symbols bound by the let*. in particular it allows for construction of recursive lambdas that may not be linearlizable.
> let* permits expressions on the right refer to arbitrary other symbols bound by the let*
In what language? I just checked Elisp, SBCL, and Guile, and they all error out if you refer to a variable not previously defined by a left-to-right traversal of the varlist:
(let* ((a (+ b 1)) (b 1)) a)
Edit: This doesn't work either:
(let* ((a (lambda () (+ b 1))) (b 1)) (funcall a)) ; (funcall a) -> (a) for Schemes
Ah, that does look to be the case. I didn't know about that one.
Signature
(letrec BINDERS &rest BODY)
Documentation
Bind variables according to BINDERS then eval BODY.
The value of the last form in BODY is returned.
Each element of BINDERS is a list (SYMBOL VALUEFORM) that binds
SYMBOL to the value of VALUEFORM.
The main difference between this macro and let/let* is that
all symbols are bound before any of the VALUEFORMs are evalled.
Lisp allows you to mutate, you can certainly write non-mutating code in lisp. Why do you think you need to use mutation with let*? let* is just a sequential let
"Lisp" is pretty broad. Whilst it was inspired by Lambda Calculus (the core of most FP languages), a lot of Lisp code is quite imperative (loops, mutable variables, control flow separate from data flow (e.g. exceptions/errors), etc.).
Scheme (and its dialects/descendants) tend to stick to a more functional style (although they also like to do stack-gymnastics with continuations, etc.). Many courses are based around Lisp.
One of the main features of the ML family is static typing, with algebraic datatypes, pattern-matching, etc. (i.e. the stuff that new languages like to call "modern", because they first saw it in Swift or something). That gives a useful mathematical perspective on code ("denotational semantics", i.e. giving meaning to what's written; as opposed to the common "operational semantics" of what it made my laptop do), and having type checking and inference makes it easier to do generic and higher-order programming (dynamic languages make that trivial in-the-small, but can make large systems painful to implement/debug/maintain/understand). This course seems to take such abstraction seriously, since it covers modules and functors too (which are another big feature of the ML family).
NOTE: In ML, the words "functor" and "applicative functor" tend to mean something very different (generic, interface-based programming) to their use in similar languages like Haskell (mapping functions over data, and sequencing actions together)
My opinion: marketing. In my experience, if I tell someone „look here is Lisp“ they turn off and roll the eyes with a „ohh that DEAD origramming language“.
If I instead say „look this new state of the art language called ML“ I get full attention and respect.
No, it's definitely not because of marketing. SML is more of a dead language than most lisps. SML is used because it has a strong type checker, which is much more conducive to learning, and just having resilient programs in general.
Why would they use a Lisp dialect instead of Standard ML? It looks ugly, it looks different from math, and different concepts have the same syntax. Standard ML looks nice, it looks like math, and different concepts have different syntax.
The instructor is clear, energetic, has a decent flow in the first lecture e.g. emphasizes the important points with vigor and repetition, switches media every 10 or 15 minutes, has a conversation with students. Slides are relatively noise-free and are informative; terms to know are highlighted.
I haven't watched a full functional programming lecture series yet, but I'd gladly audit this one.
As a "new" instructor (as in, not a TA anymore) he's got a bit of teaching talent. I hope he keeps a tight feedback loop and improves every year and continues publishing material.
on the choice of language to teach the course
why sml
i think there are a lot of nicer choice
OCaml , its basically sml only more popular and used more in real life
Haskell , again more popular , and used more in real life
Idris , newer and said to be more progressive
F# , a more practical choice and similar to sml
a lisp , well if you want to focus on the functional part and less on the types part
There's a few things which go into this (hi, I'm the instructor!).
One such reason is historical. Standard ML is a research language, and a significant amount of work on it was done by professors at Carnegie Mellon, who developed the curriculum for this course.
Even setting that aside though, I fully agree with the choice to teach it in SML. For transparency, I work professionally in OCaml, so I am not unfamiliar with it, and I enjoy it quite a bit. That being said, I think that the approach taken by CMU is best summarized as the fact that languages are ephemeral, and the concepts are what matters. We don't teach programming languages, we teach concepts -- so even if SML is not widely used, the tradeoff for having students have a simpler, less distracting, and better learning experience is well worth it.
OCaml has its own intricacies that make things difficult. For instance, you can go down a lot of rabbit holes with `dune` and `utop` and `ocamlc` and `ocamlopt` and all of these things, versus SML/NJ's simple interactive REPL. Another thing is that the language is just generally more "bloated" -- you can teach modules, but then what if a student starts running into first-class modules, recursive modules, or even beyond that, GADTs and classes and objects?
(as an aside, this is my primary reason for why I would not want to teach an introductory course in Haskell. To do anything, you suddenly need to understand the concept of type classes and lazy evaluation, and that's simply too much. I don't know much about the other languages.)
I think teaching is as much enabling students to succeed as it is to prevent them from shooting themselves in the foot. For an anecdote, there is an `Option.valOf` function (of type `'a option -> 'a`), which essentially is just a bad function that should be avoided where possible. Every semester, without fail, even though we never tell students that function exists, students are smart enough to use Google, and will use it anyways, ultimately harming themselves.
I think that same mentality applies to programming language choice, here. Keep it simple, keep it neat, and make sure that the students see what is necessary for their education, and not have to spend mental energy thinking about much more.
As a professional who uses F# every day, I appreciate this answer, even though it concerns me. Separating concepts from practical details is worthwhile, but can be taken too far. I think the FP community probably focuses a bit too much on theory. I would encourage you to consider teaching a more practical language, like F#, in the future.
I feel like F#'s commercial viability over other more computer sciencey FP languages is actually kind of a downside, from a "vibe" perspective. It seems like CS folks are not generally very interested in moving more towards the software engineering end of the discipline. Which is fair, because they are different things.
i think someone with a good grounding in sml could get proficient ocaml or f# in under a month. sml is absolutely the right language to teach a course like this in.
I agree with your first sentence, but the second one surprises me. If the practical language and the impractical one are that similar, why pick the impractical one?
sml is a smaller language, and a "cleaner" one in some ways. if your goal is to teach students the elements of functional programming, you want them to be able to concentrate on the core concepts, and not on the incidental details of the language.
http://adam.chlipala.net/mlcomp/ is a very good overview of the differences between sml and ocaml, which lets you see some of the tradeoffs each language has made.
having done the proglang course by dan grossman long ago, they use sml too, i knew lisps, ocaml, and others, but sml smallness was really really appealing
Albeit little, installing SML still requires some anount of gymnastics, google searching, and copying stuff from SO, GH Gists, etc.
Yes, I agree broadly with the line of thinking that says- if you are learning FP, you must be willing to do these, but a beginner might be disheartened and turned away- especially as someone learning alone.
Both languages encourage a concise, functional programming style but with different flavors and toolsets. They are comparable, in terms of verbosity.
I think these are correct implementations of the tower of Hanoi.
OCaml
let rec hanoi n source target auxiliary =
if n > 0 then begin
hanoi (n - 1) source auxiliary target;
Printf.printf "Move disk from %s to %s\n" source target;
hanoi (n - 1) auxiliary target source
end
SML
fun hanoi n source target auxiliary =
if n > 0 then (
hanoi (n - 1) source auxiliary target;
print ("Move disk from " ^ source ^ " to " ^ target ^ "\n");
hanoi (n - 1) auxiliary target source
)
function definition, if expressions, recursion are more concise in SML, string interpolation is nicer in OCaml
You SML code is incorrect, shows that you actually never coded anything in SML. SML requires that "if" always contains "else" clause (which is the norm for many functional languages). And this kind of stuff which makes SML unnecessarily verbose (OCaml has for operators, single line let definitions that do not require you to make use val and fun for different type definitions etc.).
I did Programming Languages, Part A, (to learn FP semantics) many years ago. It doesn't show I never coded anything in SML, it shows I made a mistake :/ .I no longer have Standard ML on my machine.
I thought having the `let` keyword encompass `fun` and `val.` was needlessly confusing. It's not concise. if `let` can mean so many things why not just do what Haskell did.
Again.. it not "so much more verbose." which was the initial point.
Functional programmers tend to use recursion instead of explicit loops with implicit state, and it's wise to regard for loops with suspicion since it can't return anything.
Standard ML as the language of instruction is a vehicle for teaching FP semantics, not imperative ones.
Also, Tower of Hanoi is a classic problem that's well suited for recursion.
You do not need to school me about why "for" is not pure functional - as soon as you understand that it is syntactic sugar, you are fine; IO is also impure, so if you insisting in purity then the only option is Haskell.
If you want to be pedantic, in functional programming there is no such a thing as "return value", as there is no such a thing as statement, only expressions. And expressions always _have_ values, they do not _return_ them.
I agree that SML is better for teaching than OCaml no doubt about it.
I mean, it is pointless to continue this conversation; you did not use SML for a long time - I did use it last time last year, and I can attest, it is extremely impractical compared to any other FP language today.
Slightly off-topic but what's a good forum to seek help on FP practices outside of the courses like this online?
Every winter break I get back into trying to learn more FP (in Haskell) and in the past several years I have been practicing algo problems (codeforces, advent of code, leetcode).
I always get stuck on more advanced graph algorithms where you traverse a and modify a graph, not a tree structure - it gets particularly tricky to work on circular data structures (I learned about "tying the knot" but it's incredibly challenging for me) and usually the runtime perf is sub-par both asymptotically and empirically.
Many graph algorithms are designed for imperative programming. It's safe to say that functional graph programming is still in its infancy. Alga[0], a system for algebraic graphs only came out in 2017. And efficient algorithms for graphs may yet to be discovered (even something as simple as reversing a list that's both efficient and elegant only came out in 1986!)
That said, as a beginner in functional programming, it would probably be good enough if you just focus on directly translating imperative graph algorithms to functional programming. You simply solve the problem by programming at a slightly lower level of abstraction.
I don't know if [0] would be any help, it doesn't talk about graphs in particular but does talk about functional-focused approaches to data structures. This note[1] on the wikipedia page for the book says it better than I could:
> [...] consider a function that accepts a mutable list, removes the first element from the list, and returns that element. In a purely functional setting, removing an element from the list produces a new and shorter list, but does not update the original one. In order to be useful, therefore, a purely functional version of this function is likely to have to return the new list along with the removed element. In the most general case, a program converted in this way must return the "state" or "store" of the program as an additional result from every function call. Such a program is said to be written in store-passing style.
Oh yeah, a pure function that accepts previous state, and returns the new state is the pattern I use a lot.
The issue is that it is hard to do on complex graph structures in an algorithm where incremental changes happen to the graph O(n) times - it ends up creating complex code and complex execution that might be slow to pass the time limit on Codeforces, let's say.
In the OCaml world maybe this is the place where you say "screw it, this abstracted function does some stateful temporary business, but it looks pure from the outside" but in Haskell it's a lot harder to pull off without going deep into monads (and I forget how those work every time).
Oh definitely go deep into monads. If you use the ST monad to do mutations, some people think Haskell provides nicer syntax to do imperative programming than traditional imperative languages like C. But of course such nicer syntax only comes from understanding and using important abstractions like monads, foldable, or traversable. Then there are niceties like automatic SoA/AoS transformation using type families.
Now if you understand the type family GHC extension, all this will be easily understood. If not, basically the idea is that you can automatically transform Array of Structs into Structs of Array by defining that transformation using a data family instance that you write yourself. For common types like tuples, the library already defined polymorphic instances for you.
Of course this transformation isn't always desirable and you don't have to use this. If your code isn't performance sensitive I wouldn't even bother with unboxed vectors; I'd just use regular vectors.
If you expect perfect factual accuracy from your teachers, yeah. On the flip side, they don't have the curse of knowledge yet i.e. it's still fresh in their mind what was difficult in the beginning, so they can probably explain very well. Just keep in mind who's teaching you, and it's just like if a co-student teaches you.
And if you feel you have to take what you hear with a grain of salt, that's probably good for your learning too.
I just mean in terms of experience. If you deviate from textbook accuracy and go into providing practical advice for real-life scenarios, someone with only a bachelor's and no work history is someone who can only give you canned anecdotes from others. Looking at his resume, it looks like he's had some internships so he has a bit of experience, and that's probably worth something
I went to CMU and was in AEPI fraternity with Brandon Wu. Brandon Wu is exceptional when it comes to functional programming. He was head TA of 15150 for a year, I have 100% confidence in him.
He took a break from TAing for some time, and when he returned he decided to have some fun with his application. He wrote a transpiler from a C like syntax to SML the day before his interview, and used it to joke that they should transition to teaching functional programming in C.
Yeah idk anything about him in particular. I just mean that, on paper, someone who goes straight from bachelor's to teaching is someone who doesn't really have any experience. I notice that a lot of the people who extol the virtues of FP to me are a lot of people straight out of college and have little real-world reference for why FP is useful in a prod environment
I do wish functional programming were more taught. we're getting to a point where I almost think OOP should be taught briefly and the rest of the focus on C/C++ for low level stuff (OS, data structures, some algorithms), and something functional or pseudo-functional for high level stuff. Most of the newer codebases in startups now require functional concepts to understand what's happening. For example, try writing modern JS without understanding .map, .reduce, et al, and function passing, etc.
Regardless I think it's important that students get exposed to more than just Python, which seems to increasingly be the only thing students come out knowing.
> I do wish functional programming were more taught
My first CS bachelor's semester in Germany in 2017 taught functional programming using Haskell (as well as C and NASM assembly in another course on computer architecture).
OOP using Java and Python was only introduced in the second semester.
> Regardless I think it's important that students get exposed to more than just Python, which seems to increasingly be the only thing students come out knowing.
In my B.Sc. studies I used C, C++, Haskel, Assembly, Java, Python, and Swift.
> My first CS bachelor's semester in Germany in 2017 taught functional programming using Haskell (as well as C and NASM assembly in another course on computer architecture).
Also worth mentioning that since this was in Germany not only was it a great education, but there’s no crippling student debt either.
Same here. Had Scala in my first year. Clojure and Racket later. On top of that I had C, ASM, Java, Python and R (had a focus on machine learning). So I got plenty of education in that department.
There are definitely some good schools out there that expose students to a broad field. For anybody who is a student and looking, this is somethign I would strongly recommend considering.
My complaint with FP:
Sometimes I just want to do something silly, like adding a log somewhere.
If I choose to add said side effect, now all my functions are marked with an io signature (so there might be _other_, nastier side effects hiding there as well - mainly an issue if you have multiple people contributing to the same project).
If I don't add the side effect, and choose to refactor multiple layers of code, I will need to make all my functions return multiple values and later fold over all the accumulated strings and... life is too short for that.
The principles really resonate with me, but maybe we are limited by the current tooling, because the development experience is quite clunky in its current stage.
This is Haskell-specific, it sounds like. I agree, the IO monad is really quite inconvenient sometimes.
I work in OCaml, which is also a functional language, but prints can be added in single lines. I address this point in Lecture 19 (Imperative Programming), actually, but my perspective is -- we invented immutability and purity to serve us, but we need not be fanatically beholden to it. In my opinion, I think Haskell goes in that direction, when every usage of IO now needs the IO monad to get involved.
A little mutability is OK. Functional programming is about the avoidance of side effects, more than simply forbidding it.
In Haskell you have a lot of options to type your functions in a more granular way. Consider the type class MonadIO, which lets you specify that your function works on any monad that can do side effects, not just IO specifically:
-- Before
captureAudioDuration :: DeviceID -> DiffTime -> IO WaveData
-- After
captureAudioDuration' :: MonadIO m => DeviceID -> DiffTime -> m WaveData
You can build the same thing, but for logging!
class Monad m => MonadLog m where
log :: String -> m ()
-- In IO, just log to stdout.
-- Other implementations might be a state/writer monad
-- or a library/application-specific monad for business logic.
instance MonadLog IO where
log msg = putStrLn ("log: " ++ msg)
-- Before: Bad, doesn't actually do any IO but logging
findShortestPath :: Node -> Node -> Graph -> IO [Node]
-- After: Better, type signature gives us more details on what's happening.
-- We can still use this in an IO context because IO has a MonadLog instance.
-- However, trying to capture audio in this function using either
-- of the functions above will lead to a type error.
findShortestPath' :: MonadLog m => Node -> Node -> Graph -> m [Node]
As you can imagine this can get quite verbose and there's other patterns one can use. Feel free to ask any follow-up questions :)
yeah, biting the bullet seems to be the way to go. As you mention the lots of "man-hours that have been spent on trying to perfect logging", when doing research the usage of that time might be ok, but when building a product you might need to make concessions.
You might benefit from a more pragmatic functional language. Erlang is broadly functional, but you can output from anywhere if you want to. It's probably one of the least pure functional languages out there, but it's super handy.
> If I choose to add said side effect, now all my functions are marked with an io signature
You got the wrong idea. You're supposed to write FP in a way where the side effect is highly layered and segregated away from pure code. IO are singularities within your chains of pure function compositions. As soon as you hit a singularity you have to break out of it as soon as possible.
The main idea is the meat. Keep your bread tiny and keep it very very separate from the meat.
The pattern is called Imperative Shell, functional core. Think of your IO as two pieces of bread in a sandwich and your pure code is the meat that connects the read to the write.
The game you're playing with haskell is to avoid letting the IO monad pollute any of your logic as much as possible.
Anyway that being said in applications where IO is all over the place... this pattern becomes largely ineffective. You basically have more bread than meat.
Others have mentioned having the same problem with this issue. One post I particularly like about the subject is this one on function colouring (how lagnuages with async/await syntax have a similar "infection"; this is a response to the original post on function colouring and not the original post).
https://www.tedinski.com/2018/11/13/function-coloring.html
It's not about monads, it's about effectful code, which is represented by special types (e.g. IO in Haskell, Eff in PureScript). Effectful code can call pure code, but not vice-versa. Since a program will have to do something, the main function is always effectful, i.e. it returns an effectful special type. So you're right that pure code isn't evaluated until some effectful code is ultimately returned by the main function and executed (by a runtime or equivalent). However, in purely functional languages most code is pure, even though it's ultimately called by effectful code.
Monads and side-effects aren't intrinsically related. Simplifying, a monad is something with flatMap() - in JavaScript, Array and Promise are monads (kinda). What flatMap() gives you is the ability to chain things, which is useful to sequence side-effects so that they can be performed by a machine in a given order. That's why IO and Eff are monads.
To me the issue is even simpler, not even about effectful code or monads, it's about the "issues" of lazy evaluation.
It doesn't really matter if you made a horrible uncomputable mistake deep inside a pure function, if you either discard or never use that mistake you would never notice it happened.
For example, nixpkgs (in nixlang) for sure isn't "strict pure FP" - you can do effects whenever you want - but will often have uncomputable evaluations in it. And you will only notice they're uncomputable evaluations when you try to use them. There's even a warning left for future wanderers for that:
### Evaluating the entire Nixpkgs naively will fail, make failure fast
AAAAAASomeThingsFailToEvaluate = throw ''
Please be informed that this pseudo-package is not the only part of
Nixpkgs that fails to evaluate. You should not evaluate entire Nixpkgs
without some special measures to handle failing packages, like those taken
by Hydra.
'';
I see that I formulated my thesis badly, conflating strict evaluation with monads, when as you pointed out they're not strictly related.
It seems that the fundamental problem with the functional paradigm (in its pure form) is that the real world - including the architecture of the computer that is used to run the programs on - is full of side effects, i.e. is essentially "imperative," and with this impedance between them the idea creates more problems than it solves.
Any time an HN comment uses the phrase "fundamental problem" chances are the problem is not only not fundamental, but also rather fun to solve for the right kind of people.
Exactly. These languages come up with more and more strategies and abstractions (from monads to modern effect systems) to help you manage your side effects well. It then raises the level of abstraction in your programs to become simpler and more concise.
just in time for nobody to really care about programming because LLMs are so good at translation and all computer code and programing are a subfield of linguistics...
Programming runs along the back of mathematics. A field famous for spending centuries wasting time on silly games until those same silly games end up being the foundation on which modern society functions.
Even if LLMs completely remove the need for programming (a rather big IF), this is not time wasted.
Functional programming is a perfect pairing with AI-generated code because the type systems are generally more expressive than non-functional languages, which means the compilers can catch all sorts of errors.
Bob Harper (a CMU professor with a focus on PL theory) also has a really good SML reference that is closer to a textbook than lecture notes.
http://www.cs.cmu.edu/~rwh/isml/book.pdf