[Cross-posted to the Tucows Developer Blog]
I suppose it’s fitting that this article is being published on Friday the 13th: it’s SoftwareDeveloper.com’s Ghosts in the Machine: 12 Coding Languages That Never Took Off. The 12 languages covered in the article are:
- ALGOL 68
- brainfuck
- befunge
- REBOL
- CFML (as in ColdFusion)
- Java2k
- INTERCAL
- VRML
- SMIL
- Haskell
- Delphi
- PowerBuilder
Some of the languages in this list were never meant to take off. Only a total sadist, masochist or Perl programmer would ever want languages like brainfuck or befunge to take off; the first two for obvious reason, the last one because then s/he’d be able to say “You think Perl looks like line noise? I’ll show you line noise!”
Some still have loyal aficionados: I know a few PowerBuilder and Delphi developers, and I’ve had the pleasure of meeting David Intersimone, who’s got the world’s toughest developer relations job: he works for CodeGear, the company that took over all of Borland’s developer tools. I’m certain that there are PowerBuilder and Delphi apps running today, performing yeoman service in small and enterprise businesses. Still, they’ve been eclipsed by Java, and .NET.
The one inclusion in the list that might draw some ire is the inclusion of Haskell, one of the current darlings of the Reddit set. I myself wonder about this choice — five years ago, a little-talked-about language named Ruby occupied the same mindshare niche where Haskell is today.
It’s an interesting list. Check it out!
10 replies on “Dead Languages: 12 That Never Took Off”
What moved Ruby off of this was was a killer app (Rails) and so far Haskell lacks that. Right now Haskell is on track to become the Lisp of the 21st century: interesting, mind-warping, and the wrong tool to choose for delivering a real application or service. I predict a long future for Haskell in academia and not much future outside of it. [Now to go hide from the oncoming mob of Lisp weenies…]
Well, there’s always a chance that the Haskell equivalent of David Heinemeier Hansson is cobbling together a Haskell-based killer app between supermodel shoots!
And for those readers who don’t get the reference to Lisp, what Jim (who I know from the P2P boom in 2000/2001) did is the programmer equivalent of standing near a tree in a lightning storm, waving a golf club and screaming “Storms suck!” as loudly as possible.
I’ve nothing against Haskell, but the difference between Ruby and Haskell is that Ruby found its killer app/mojo/community in a comparatively short period of time.
Haskell has been around in various forms since 1990 and it still hasn’t taken off. That’s 17 years of not taking off.
Ruby, on the other hand, emerged in 1995. If you count the growing buzz around Rails in 2005 as Ruby’s taking-off point, that’s ten years from concept to community.
It doesn’t take long for good ideas to catch on.
Leaving aside my original snarky answer, I think that what will end up dooming Haskell is that it is just too hard for most coders to use effectively, or at least to do something that they can’t do 80-90% as well in another language. This is what I think distinguishes Haskell from Erlang (another reddit fave) — Erlang is different that just about anything you have ever used before but the learning curve is not as steep as Haskell and the niche it occupies (concurrency) is served incredibly poorly by every other language out there and it becoming a hot topic.
Another factor that may help Erlang adoption: the book. As with Dive Into Python, there’s nothing like a well-written book to help give a language a boost. If Programming Erlang is at least on par with other “Pragmatic Programmers” titles, it should help.
People outside Ottawa would be shocked at the amount of PowerBuilder code that is run and written in the Canadian government. I think it’s the entire reason that Sybase hasn’t killed it yet.
[…] in merito qui: Seth’s Blog, ongoing, Smalltalk Tidbits …, Epeus’ epigone, Global Nerdy, Listics, Message, CrunchNotes, seattleduck, Eric Rice, Josh Bancroft’s …, Things […]
Mentioning Delphi under this title is ludicrous.
Haskell is, in a very specific technical sense, the most expressive programming language available. And it isn’t going to lose that title unless there are major changes in industry. In particular, Haskell was designed to be a **research language**. A platform for academic research in type theory, compilers, and so on. It will “always” be the state of the art from which lesser compilers and languages copy.
It doesn’t matter if Haskell “takes off”. The vast majority of new CS graduates know either it or ML. There is a glut of programmers willing to take a pay cut to do Haskell. Hiring for Haskell is easier than for other languages.
If you want to start a new project using the best general purpose tool available, there’s no reason not to.
Brainfuck is world’s smallest programming language but is very unpractical. It is very difficult to program even a simple program in it.