What's your opinion on Lisp/Scheme?

HI I AM NEW TO THIS INTERNET THING HOW TO I PICK UP MY E-MAIL??

Useless

>>61775897
Why?

>>61775704
Very interesting. It's a shame that JS wasn't just Scheme in the first place, as it was supposed to be.
It's so fucking easy to compile to Scheme compared to JS, languages could without any problems run in browsers.

Why is Lisp said to be good for AI but is never actually used for it? It's the same for Prolog.

>>61776284
Peter Norvig who's an expert on AI used to shill it in the 80's and 90's. But now we got all these new types of machine learning that it's not the best anymore desu.

>>61776306
What's the best now then?
And what made Lisp good for AI back then?

File: sicp1450651227519-1.gif (9KB, 511x494px) Image search: [Google]

9KB, 511x494px

>>61776284
doesn't matter how good a language is for something when most people are dumb as rocks
>>61776323
short answer is that data and code are indistinguishable
lisp macros too, like nothing else
go watch sicp

>>61776165
no libs, community dead

>>61776670
So what are the go-to languages for AI now?

>>61776284
spoiler alert, what is shilled as AI, ML, NLP is literally a couple of matrix multiplication and more retarded than coding websites

>>61776670
That's not true at all, Guile is quickly becoming \emph{the} GNU configuration and go to API language.
And I couldn't be more happy about this.

>>61776657
Is sicp really that good or is it a meme?
I've read the other memebook (K&R) and I was unimpressed.

>>61775704
Not my thing. I prefer FP languages to be more like Haskell or (impure but useful) like Scala.

Lisp syntax gets on my nerves.

>>61775704
Useless in fact. But because it's used in SICP and is a functional programming language in a way, some pajeets in /dpt/ consider it cool alongside with Haskell and C.

>>61776702
In a way, but picking the best methods and then running them fast and in parallel is still both difficult.

Then again, I go with Scala on for the latter part:
https://spark.apache.org/mllib/
https://spark.apache.org/graphx/

Finally something nice.

>>61776702
I wouldn't say that. They use matrix multiplications to process large amounts of data.

The essense of ML and sometimes NLP is probability theory and statistics.

File: sicpLecture-1a - 00:00:52.png (105KB, 320x240px) Image search: [Google]

105KB, 320x240px

>>61777106
It is certainly a meme, but that doesn't mean it's not also great.
It remains the best introduction to the subject even after all these years (the fundamentals of programming having themselves not changed), but also has much to offer any level of programmer coming from any background.
https://archive.org/details/MIT_Structure_of_Computer_Programs_1986/

>>61775704
Scheme is my favorite language.
I hate Common Lisp.

>>61777204
Is it a good first step towards becoming an actually good programmer?
I know my way around C and have a basic idea of how stuff works but sometimes I see people on /g/ talk about the theory behind CS, data structures, complex language features and design patterns, and I feel retarded because I don't understand shit. I'm searching for books to get better.

>>61775704
I've barely stuck my foot in it. I'm finding it hard to read, the shape of my programs is hard to see.

>>61777106
I myself don't think it's "great".

Most suffer the same "1-4 authors did it in at most a few years each" syndrome. [Sucks that our governments don't copyleft teaching materials for continuous improvement.]

Also it might not be what you want. Honestly, for many, learn lisp the hard way plus CLRS / the algorithm design manual / WP+SO or whatever is more to the point.

>>61777253
SICP will get you straight to the heart of some of those thing better and deeper than anything else I know.
Watch a lecture or two, and if it's not for you, whatever, try something else.
Don't feel bad though, neo/g/ mostly just spews buzzwords and doesn't actually know about or understand any of those things.

>>61777296
What about the little schemer?

>>61777253
>Is it a good first step towards becoming an actually good programmer?
I'd instead recommend reading:
> CLRS / the algorithm design manual / WP+SO
Because they're more succinct and to the point.
Oh sure, knowing more is better (I might honestly say that for example, Kolmogorov Complexity would be interesting to know, or all stuff about any kind of knapsack problem), but you probably don't have time for that.

>>61777349
I actually have the book so I'll just read it instead of watching lectures.
What are some other good books about the subjects I mentioned? Things that'll make me a good programmer as opposed to a code monkey

>>61777365
Not entirely my kind of book. Then again, why not.

>>61777379
No man, you can't skip they lectures, they're so great.
Sussman is my waifu.

>>61777378
>CLRS
>WP+SO
What are those?

>>61777190
>probability theory and statistics.

valid

ML and AI are memes to divert VC money, they mean nothing of what average people would think watching scifi

>>61777162
running an executable is not hard

>>61777754
CLRS can just be found with any search engine, you'll figure out which result is correct. Try it.

Wikipedia and StackOverflow are the other two.

>>61777993
> running an executable is not hard
If it doesn't need to work on 4 cores or maybe 40x 16 cores on a network, maybe it isn't.

But once it needs to run in parallel and maybe distributed across a network, you're facing a like 50 year old aresepain of a problem that is more current than ever now that individual machines get more and more cores.

>>61775704
Scheme is good to academic purposes, nothing else.
In my school the scheme advocates convinced (or implemented it themselves) a webserver in scheme to run some administrative processes as course subscription.
Needless to say it was a flop, the server didnt stood up to the stress of a couple hundreds of students requests.

>>61778717
What about Lisp in general?

>>61777378
Does CLRS teach design patterns and programming principles or is it only algorithms?

>>61778096
basically you say, don't do dumb shit and distributed programming is hard, which has nothing to do with what is sold as ML today mind

>>61778717
Lisp is a programmable programming language. You can do anything with it

>>61777204
I agree. Too bad MIT dropped it entirely from the curriculum some time ago.

http://www.posteriorscience.net/?p=206

>>61779364
Why would you care? It's not like if you're an MIT undergraduate.

Financial programs use Haskell and OCaml but not Lisp. Why?

>>61776284
You can do basic linear algebra in any programming language now.

>>61780628
Clojure

>>61780777
Who uses Clojure in finance though? Never heard of that.

>>61778846
It does teach design patterns as well

>>61777253
Read C++ Primer