15 replies to this topic

[Pally]

according to the wikki page on F#, its actually a multi-paradigm language; so it can be adapted easily to many situations (i guess.)<br /><br />Still, without thinking about a specific domain of problems its hard to really nail what a productive language is. Some problems lend themselves to functional style, some to OO, some to procedural etc. A language with the best fit to the problem will give you a productivity boost.

Upon further thought into this discussion I think your right, 

​the only way to really remain productive is to use a domain specific language that does the job most productively for the task at hand...

When I started this discussion I thought of a general purpose language... I suppose any general purpose language will run into some conflict between various tasks and have trouble creating productive solutions... productive seems conflicting term as well because productive low level manipulation of memory could be your idea of "productivity" one day and tomorrow your idea of productivity may be getting a similar program working even though its sloppy in memory consumption. In both cases your trying to be productive but with separate tasks at hand which conflict with another end goal.

 

 

So ironically this discussion of the most productive language seems to have led to the same outcome the popular "whats the best programming language" topics always lead to.

Edited by Pally, 07 March 2013 - 07:37 AM.

[LinkChef]

Although I've not been a programmer for a very long time, I did look into programming languages long before I actually programmed.

 

Although my first opinion on the matter is simple, it's one that should be said.  Obviously, the best programming language is the one YOU know the best.  Not surprisingly, you can be super efficient in anything that you know really well when you compare yourself to others.  Plus, if you've written code before that would be handy for the program you have now, you know where that code is, and copy, paste, and adapt into a program.

 

However, if we get rid of that entire idea, although I can't say with experience from other languages, C# with the .NET Framework has to be one of the most efficient out there.  .NET is the entire reason though.  C# in itself is great, and is very good for people familiar with C++ or Java, but .NET framework allows coders to be a tad lazy.  One of the biggest things I've learned is that computers math-wise can only add, subtract, multiply, and divide.  To do powers, or logarithms, etc., you have to be smart with the basic operators and making a method.  That's why we have the Math.Pow() method in .NET.  This takes a lot of time out of the coding when you know you can do more complex operations without working through how to do it with basic operators.

 

Anyway, that's what I've seen, but I've read a few others, and they're all really interesting.  Have a good day everyone!

[WingedPanther]

LinkChef, I've been looking into ASP.NET recently with MVC 3. It's easy to say it allows a programmer to be lazy, but it also allows a programmer to easily integrate unit testing into the project, practice the discipline of separating logic from display of content, etc. In that sense, .NET can also be said to make it really easy for a programmer to be disciplined. At the same time, I'll be having to think carefully about how to use LINQ with my database queries to optimize the efficiency of my code while maintaining maintainability

 

Everything requires balance

[mbcev]

I want to throw another point in for LISP's. I've spent quite some time now learning Clojure to augment my usual Java tomfooolery and I have to say that after some practice I've been finding myself able to write more with less. But before I get into that:

I've played with C, C++, C#, JavaScript, Racket, Python, heck even MIPS and x86 for a time too. The biggest reason I stuck with Java over adopting any of those alternatives (except where specific situations warranted their use, and in fact my first language was C so I still hold a soft spot in my heart for it) was specifically because they didn't improve my productivity at all. In fact they slowed me down simply because I didn't know the standard libraries. Although I don't believe this truly counts as a legitimate "slow down" because if I were more experienced in any of those other options then I wouldn't be any slower with them than I would be with Java (except maybe the assembly languages ). 

 

I mean I could write up my own instances of the data structures I needed, or parsing libraries, networking and io libraries too, but it kind of defeats the purpose of having "standard libraries" right? And since I have become very familiar with Java and it's countless libraries, why spend the brain power hacking away in those other languages? My problem solving methodology remains entirely unchanged so 90% of the time, the code I wrote in one language I could almost completely transcribe into another (this is especially true with Java and C# because even their syntax and choice of keywords is almost identical haha). As far as this family of languages is concerned, my opinion pretty much is that you're going to be the most productive in the language with the most tools for you to be productive with (read: lazy). Typically this is going to mean "who's got the most useful libraries for me to take advantage of". 

 

Back to Clojure; what's so special about it? For those who may recall my frustration with the language when I originally picked it up, it was worth the headaches and here's why:

 

Languages, both computer and human/linguistic, as well as Mathematics are ways of expressing human thought. Human languages specifically allow us to transmit ideas to one another in very complex and abstract ways which can be difficult to express in terms of Computer Languages or Mathematics. One of the issues with Human languages though is that they're rather static. You generally can't take expressive shortcuts like say recursion in CS/Math. This is where Mathematics has come in. Ultimately mathematics is very much just the short hand abbreviations for what was once written out in sentence form. There are more shortcuts that you can take to encompass wider sets of ideas this is what makes it so useful still today. Next came programming languages which are so immensely powerful in their own right because they allow us to go so far as to automate human thought! You can now string together lines of Mathematical expression, which was already a short hand for spoken languages, in order to produce machines that can just flat out replace a human. The efficiency gains are enormous and we see them every day! Dealing with ATM's is way faster and way cheaper than dealing with human tellers in a brick and mortar bank as a for example.

 

So what does this have to do with LISP's? More specifically Clojure? So I'll admit, LISPs are hard to pick up at first... especially if you've become ingrained in the ways of Declarative and/or Imperative programming (as I most definitely was)... but what makes them special is that they let you do things that you would otherwise have to do "the long way" much like what mathematical symbols did for written word and later programming for mathematics. The more pure functional languages like Common Lisp, Scheme, or Haskell in a sense allow you to automate your automation. This is why LISP's have stuck around for as long as they have, but they do still have there share of problems, which is why they never gained enough traction to ever take over the world. Clojure gets a lot of slack from the "pure" Lisp community for being so middle of the road in terms of "Lisp purity" but I think that's what makes it so much more productive to use. It's flexible. It's got most of the fluff that you have in a "pure" Lisp while still allowing you to escape from those shackles when you need to. I believe this is why Ruby, which was mentioned a few times already) is so popular too. It has a lot of functional programming influence in it but it isn't "pure". Neither Clojure, nor Ruby, are tied down to their religion like Haskell is, lovely a religion though that it may be. Clojure is also convenient because of the interop with Java means I can use libraries when I need/want to instead of writing it myself after finding out another Clojurian hasn't written it for me. And Ruby is popular enough that it doesn't have that much so much anymore (at least I presume so; I can't speak from experience on that one). Basically, you have your productivity gains from libraries/tools that you'd see in any of the other previously listed languages, as well as from the "magic" that is Functional Programming.

 

Long story short: my vote goes to Clojure in terms of productivity. But then, that's just my opinion. To each their own  .

Edited by mbcev, 14 March 2013 - 01:15 PM.