i took an ML course, i thought pattern matching would mean data-driven programming like in SICP and PAIP but it turned out to just be a syntactical gimmick somewhat like a switch statement. there was no metalinguistic abstraction to be seen either, but lots of type checking which is supposed to magically solve all programming errors.
i don't know if i should be skeptical of this stuff but i guess it's the future if the powers that be says so and i should just shut up and do my homework
Name:
Anonymous2015-10-04 22:33
>>10 No, type checking is good. Python invented it this year.
I'd like to see you unwrap three layers of constructors in one line, while binding 4 variables and guarding against 2 other variables. All in one line and without pattern matching.
>>21 There are more types in heaven and earth, son, than are dreamt of in your philosophy.
Name:
Anonymous2015-10-09 15:27
ML was basically TypeScript for Lisp, but only covered a small subset of the language.
Name:
Anonymous2015-10-09 20:06
If Haskell is cancer, ML is the carcinogen.
Name:
Anonymous2015-10-09 20:47
>>25 Can't argue much with that except to point out that it's not a strict subset, and the lambda calculus was placed in the type system instead of the run-time parts of the language.
Name:
Anonymous2015-10-10 10:10
When your shitty toy language allows for useful type inference and your compiler for your shitty toy language does whole-program optimisation and your shitty toy language has a proper module system, then you can wipe your dirty penis on ML you son of a bitch
Name:
Anonymous2015-10-10 12:37
>>28 "ooh look at me, I use MLton I'm such a fag" - u
Name:
Anonymous2015-10-14 18:16
If you listen to ML fans, ML itself isn't an ML family or ML style language.
Name:
Anonymous2015-10-14 18:20
>>30 yeah you can't even do machine learning with it
Name:
Anonymous2015-10-16 10:59
Everything boils down to Machine Language in the end.
ML ``programmers'' have a dangerous mix of arrogance and ignorance. They're Dunning-Kruger personified. I wouldn't trust those punks with my garbage.
Name:
Anonymous2015-11-03 8:01
Robert Harper
Name:
Anonymous2015-11-03 9:06
Here is a picture of the MLton team at a meeting in Chicago in August 2003: http://mlton.org/Developers .. The first dude is actually getting paid Jane Street Capital levels of money so that's why he is actually smiling + going to this meeting means he is done with his PIP for the year.
The rest is smiling like they are posing for an autists anonymous emotion recognition seminar.
Not trying to be shitty about 1ML btw, which is a fine thing. The module thing is just a DSL in scheme but the types are where it's at. The types are _why_ ML got the module system it did, scheme didn't have that perspective so it landed somewhere else. It helps to remember: MLs have the lambda calculus too, it's in the type system.
That's kind of the problem. RnRS is taken more as a suggestion, and each Scheme does everything in its own subtly different ways. The result is that the most portability you get is about thirteen RFIs. This module DSL would therefore only work in one Scheme; as part of a RnRS it would at least have more luck as a standard that more than one implementation could use.