5 ms·

Programming Paradigms for Dummies: What Every Programmer Should Know

86 points by alrex021 17y ago

alrex021 17y ago
Named and Unnamed state. "It seems that we need to have and not have named state at the same time. How do we solve this dilemma? One solution is to concentrate the use of named state in one part of the program and to avoid named state in the rest. The bulk of the program is a pure function without named state. The rest of the program is a state transformer: it calls the pure function to do the actual work. This concentrates the named state in a small part of the program." This very closely resembles the philosophy behind Clojure.
jasonkester 17y ago
Mmm... Computer Science... 39 pages of things that have no bearing whatsoever on what I do every day while building a startup. Does anybody here really care about any of this stuff? I program computers for a living, and I couldn't even muster the energy to page-down through that mess of tables and graphs.
- marshallp 17y ago
  Programming computers is about creating abstractions, if you ignored the abstractions others have methodically thought out - well, then you'd still be programming in assembly. The author, peter van roy, has written a book, ctm, that many consider the successor to sicp, and created a programming a programming language, mozart/oz, that is the most advanced and well thought ever (it's descrined as almost like magic on it's homepage).
  - jasonkester 17y ago
    Sure, but then most of Computer Science is common sense to anybody who's been doing it since they were a kid. I'm perfectly happy not knowing the CS terms for the things that I do or the design patterns that I use. In 1995, I was asked in an interview what it meant to "deference" a pointer. I hadn't the first clue what this guy was talking about, despite having 5 years of self-taught C and C++ under my belt at that point, much of that time spent deferencing away happily without a fancy term for it.
    - marshallp 17y ago
      Well, Peter Van Roy's CTM is not an impenetrable and jargon filled, it's actually easier to read than SICP, I think all programmers should read it at least once (I don't know peter van roy personally, I'm not shilling for him or anything). Your point about c and c++ kind of illustrates my point, I think there are very few people who should ever be programming at such a low level, if more programmers were to thoroughly familiarize themselves with the 'abstractions' that have been discovered, that mistake wouldn't be made.
      - jasonkester 17y ago
        > Your point about c and c++ kind of illustrates my point, > I think there are very few people who should ever be > programming at such a low level... Certainly, one can be excused for programming in C in 1995. You realize that there weren't many alternatives back then, right? Substitute "Factory Patterns" if it helps you understand what I'm talking about. Or CTM or SICP. I'm all about working at a nice high level of abstraction. I'm just not particularly bothered to know the technical term for whatever abstraction I happen to be using at the moment or its storied history. I see something like "Figure 2, Taxonomy of Programming Paradigms", and I can't for the life of me understand the mind of somebody who'd consider that chart essential learning for a programmer.
        ryanpetrich 17y ago
        Knowing the proper technical terms for abstractions you use is helpful when collaborating with others
        dkersten 17y ago
        This cannot be emphasized enough.
        marshallp 17y ago
        "Certainly, one can be excused for programming in C in 1995. You realize that there weren't many alternatives back then, right?" I didn't mean to imply you directly, the question is why did the company you were applying to and the industry in general not 'see the light'. lisp,python,tcl,prolog,sql existed before 1995. Personally, I disagree with the pattern languages of object oriented design, they aren't necessary if you know the full range of abstractions of which CTM and SICP describe (functional, relational, logical, constraint etc). The proliferation of these difficult to use OO patterns is another symptom of programmers not knowing enough computer science.
- arethuza 17y ago
  YMMV I thought it looked rather interesting
- anonjon 17y ago
  Yes? I do care about this stuff. The first reason is that I'm kind of a wonk for this stuff. I like reading about the different languages and learning them. I have fun doing it; (I also program because I think it is fun, I could probably make plenty of money doing something else had i the desire...). The second reason is that I believe you should use the right tool for the right job. Firstly, by learning more paradigms I get more flexible with my programming in any language. I can pick the way to implement that will fit my solution best, rather than worrying about shoehorning the solution in my head into the wrong pattern. Secondly, I can make better choices at the start of a project. If I have a basic plan at the beginning of a project, I can decide based on the plan what will most likely make implementing it the most pain free. When your implementing language fits your problem domain perfectly, you can be incredibly productive.
- lisper 17y ago
  > Does anybody here really care about any of this stuff? Yes. > I program computers for a living, and I couldn't even muster the energy to page-down through that mess of tables and graphs. There are a lot of people who share your view. It is no coincidence that there is also a lot of crappy software out there. Interspersed among "all those tables and graphs" that you could not muster the energy for are things called "words." Those words convey meaning. You might want to try to understand that meaning before dismissing "this stuff" as not worth caring about.
DannoHung 17y ago
This is Oz's set of primary datatypes: http://www.mozart-oz.org/documentation/tutorial/node3.html#label14 http://www.mozart-oz.org/documentation/tutorial/node3.html#l... This is a god damned nightmare if you want to actually use the language. Why? Because each of these types has separate methods and syntaxes that are used for manipulation. Go on, try to write something that is generic, I dare you. Oz is, without a doubt, the best example of why you should not try to be everything to everyone that I have ever seen.
etherealG 17y ago
what every programmer should know: publishing their works in a pdf will make it hard for other programmers to read it.
- vorador 17y ago
  Could you tell me why ? After all, pdf is an iso standard.
pragmatic 17y ago
Anyone used Mozart/Oz for anything? This chapter is fascinating. Or rather has me fascinated with Oz. However, I see that Mozart was last updated in 2008.

86 points by alrex021 17y ago

Named and Unnamed state. "It seems that we need to have and not have named state at the same time. How do we solve this dilemma? One solution is to concentrate the use of named state in one part of the program and to avoid named state in the rest. The bulk of the program is a pure function without named state. The rest of the program is a state transformer: it calls the pure function to do the actual work. This concentrates the named state in a small part of the program." This very closely resembles the philosophy behind Clojure.

jasonkester 17y ago

Mmm... Computer Science... 39 pages of things that have no bearing whatsoever on what I do every day while building a startup. Does anybody here really care about any of this stuff? I program computers for a living, and I couldn't even muster the energy to page-down through that mess of tables and graphs.

marshallp 17y ago
Programming computers is about creating abstractions, if you ignored the abstractions others have methodically thought out - well, then you'd still be programming in assembly. The author, peter van roy, has written a book, ctm, that many consider the successor to sicp, and created a programming a programming language, mozart/oz, that is the most advanced and well thought ever (it's descrined as almost like magic on it's homepage).
- jasonkester 17y ago
  Sure, but then most of Computer Science is common sense to anybody who's been doing it since they were a kid. I'm perfectly happy not knowing the CS terms for the things that I do or the design patterns that I use. In 1995, I was asked in an interview what it meant to "deference" a pointer. I hadn't the first clue what this guy was talking about, despite having 5 years of self-taught C and C++ under my belt at that point, much of that time spent deferencing away happily without a fancy term for it.
  - marshallp 17y ago
    Well, Peter Van Roy's CTM is not an impenetrable and jargon filled, it's actually easier to read than SICP, I think all programmers should read it at least once (I don't know peter van roy personally, I'm not shilling for him or anything). Your point about c and c++ kind of illustrates my point, I think there are very few people who should ever be programming at such a low level, if more programmers were to thoroughly familiarize themselves with the 'abstractions' that have been discovered, that mistake wouldn't be made.
    - jasonkester 17y ago
      > Your point about c and c++ kind of illustrates my point, > I think there are very few people who should ever be > programming at such a low level... Certainly, one can be excused for programming in C in 1995. You realize that there weren't many alternatives back then, right? Substitute "Factory Patterns" if it helps you understand what I'm talking about. Or CTM or SICP. I'm all about working at a nice high level of abstraction. I'm just not particularly bothered to know the technical term for whatever abstraction I happen to be using at the moment or its storied history. I see something like "Figure 2, Taxonomy of Programming Paradigms", and I can't for the life of me understand the mind of somebody who'd consider that chart essential learning for a programmer.
      - ryanpetrich 17y ago
        Knowing the proper technical terms for abstractions you use is helpful when collaborating with others
        dkersten 17y ago
        This cannot be emphasized enough.
      - marshallp 17y ago
        "Certainly, one can be excused for programming in C in 1995. You realize that there weren't many alternatives back then, right?" I didn't mean to imply you directly, the question is why did the company you were applying to and the industry in general not 'see the light'. lisp,python,tcl,prolog,sql existed before 1995. Personally, I disagree with the pattern languages of object oriented design, they aren't necessary if you know the full range of abstractions of which CTM and SICP describe (functional, relational, logical, constraint etc). The proliferation of these difficult to use OO patterns is another symptom of programmers not knowing enough computer science.
arethuza 17y ago
YMMV I thought it looked rather interesting
anonjon 17y ago
Yes? I do care about this stuff. The first reason is that I'm kind of a wonk for this stuff. I like reading about the different languages and learning them. I have fun doing it; (I also program because I think it is fun, I could probably make plenty of money doing something else had i the desire...). The second reason is that I believe you should use the right tool for the right job. Firstly, by learning more paradigms I get more flexible with my programming in any language. I can pick the way to implement that will fit my solution best, rather than worrying about shoehorning the solution in my head into the wrong pattern. Secondly, I can make better choices at the start of a project. If I have a basic plan at the beginning of a project, I can decide based on the plan what will most likely make implementing it the most pain free. When your implementing language fits your problem domain perfectly, you can be incredibly productive.
lisper 17y ago
> Does anybody here really care about any of this stuff? Yes. > I program computers for a living, and I couldn't even muster the energy to page-down through that mess of tables and graphs. There are a lot of people who share your view. It is no coincidence that there is also a lot of crappy software out there. Interspersed among "all those tables and graphs" that you could not muster the energy for are things called "words." Those words convey meaning. You might want to try to understand that meaning before dismissing "this stuff" as not worth caring about.

DannoHung 17y ago

This is Oz's set of primary datatypes: http://www.mozart-oz.org/documentation/tutorial/node3.html#label14 http://www.mozart-oz.org/documentation/tutorial/node3.html#l... This is a god damned nightmare if you want to actually use the language. Why? Because each of these types has separate methods and syntaxes that are used for manipulation. Go on, try to write something that is generic, I dare you. Oz is, without a doubt, the best example of why you should not try to be everything to everyone that I have ever seen.

etherealG 17y ago

what every programmer should know: publishing their works in a pdf will make it hard for other programmers to read it.

vorador 17y ago
Could you tell me why ? After all, pdf is an iso standard.

pragmatic 17y ago

Anyone used Mozart/Oz for anything? This chapter is fascinating. Or rather has me fascinated with Oz. However, I see that Mozart was last updated in 2008.