Glad to hear it, I subscribe to your RSS and look forward to your thoughts. Since I have another shot at making myself clear (accidentally posting whilst editing didn’t help my case), I might as well use up a bit more of your space and attention. I’m referring to problems where there’s very little exploratory programming to be done, so the inherent agility of various dynamically typed languages doesn’t help very much.

I pretty much disagree with everything you wrote there. I’ll probably elaborate in a future blog post.

The examples I can think of fall into three categories; interacting with multiple poorly designed legacy systems, algorithmic research and limited platforms (embedded systems, working on GPUs, parallel computing back in the days when you had to start with sockets, etc). The difficulty arises from the operations necessary to complete the tasks, fundamentally complex algorithms. Once you’re in that space everything else (containers, flow control, syntax) adds a relatively small fixed cost, you pay it once and then get on with the actual work.

There are plenty of actually hard problems to solve, that’s the basis of Computer Science in my mind. You can find a plethora of examples in the “The Art of Computer Science” series. Yes, much of that work is now available in just about any decent language, but I think my point is plain.

I think that this is the point the author of parent article was trying to get across. He diluted the message by being needlessly inflammatory and occasionally wrong. It would have been a much better piece if it was shorter.

At the risk of making the same mistake, I can’t believe you (Jared) linked to that function point analysis as support. I don’t think much of the technique to begin with and that was a great example of why no one should. They included HTML (Why not include photoshop while you’re at it?) and SQL Forms beat everything else by nearly a factor of 2. Ugh.

Since the programmer, the problem and the best way can vary the best language can vary.

I completely agree with this statement.

However, in many (most) corporate environments, the programmer doesn’t have a free choice of language. I think we have around 8 acceptable languages at my firm, and my team considers 3 of these to be acceptable. If I used one of the other 5, I wouldn’t be popular

I think I am getting side-tracked from one of my main points. And that is, sometimes you will be working on a project that doesn’t have a library available such as a device driver for a brand new piece of hardware (yes, I finally thought of one).

The other 99% of the time, more important than the language is the libraries available. Even C has data structure libraries, socket libraries, event libraries, threading libraries, xml libraries, database libraries, etc. And because everything else is written in C, they have access to all of the C libraries, plus all of their own libraries.

Since the programmer, the problem and the best way can vary the best language can vary.

Best can mean a lot of things. Fast (C, Assembly), safe (Ada, BitC), shortest (Perl among others), readable (depends on the programmer), portable (C (because a C compiler is available virtually everywhere) or interpreted languages (because it runs, where the VM/interpreter runs)) code to name a few examples.

There is a reason, why there are multiple programming languages.

In the examples you gave, you’re forced to use a particular language. But the libraries provided (at least in Linux Kernel and Emacs, I’m not familiar with the others) are still more important than the language itself.

I mean, with pure emacs lisp sans libraries, you can do… not a lot, at least without a huge amount of effort. With emacs lisp + emacs core libraries, you can effortlessly write an amazing text editor (:-)), or a great interface to Sybase, or a stock subscriber, or…

Anonymous seemed to imply that cases exist where it doesn’t matter if you have no libraries at all. Trival examples such as hello world and fibonacci aside (and the C implementation of hello world uses stdio), I struggle to think of any.

iPad, iPhone where you must use Objective-C (at least to make native apps)

SAP where you have to use ABAP

The Linux Kernel where you have to use C.

and so on…