It's a must read because it's so far outside our normal experience. First the languages are derivatives of APL that do one thing very well and nothing else. Don't try doing a web page in APL. Second, the programs efficiently process billion and trillion line databases in a tightly constrained environment. Third, Whitley throws away everything every few years and reimplements from scratch, even the parts that work. He talked about millions of hours of running time without an error. Pretty amazing.
It got me to thinking about the virtues of simplicity instead of the increasing complexity that we are getting into.
[Whitney's] languages are derivatives of APL that do one thing very well and nothing else.
I don't think that's fair. They are general-purpose languages. It's true that they've mostly been used on certain classes of problems so far, but that could just be a historical accident. Who's to say what people would make with them if they found a broader audience?
Whitney is not someone who thinks in special cases.
Edit: I agree with everything else you said, though.
There's a reasonable argument that APL and derivatives are more expressive than lisp and its derivatives, at the expensive of having a much larger startup cost in terms of things you have to know.
I suspect that in another few years the old-school APL alphabet will be positioned to start making a comeback, via in-software, touch-screen 'keyboards' (thereby eliminating the expense of a very small run of 'funny symbol' keyboards).
I really enjoyed reading that. As russell said, it is far outside my experience, but it is more like I wished my experience would have been.
I read an article on APL back in the early 80s and it sounded wonderful, but I never had an occasion to learn or use it. The closest thing for me now are list operations in Python and the collection functions I'm learning in Clojure right now.
Consider that Arthur's software sells for 25K per core and 8 cores is pretty much a minimum. And that the sole executable is < 300K. Now re-read the article :D.
What an absolute treat. Whitney has done things which, if he hadn't done them, would be called impossible. It's such a shame that his work isn't widely available. There's a whole world inside those 300 kb. It's the only thing I've seen that got me as excited as Lisp.
It's a must read because it's so far outside our normal experience. First the languages are derivatives of APL that do one thing very well and nothing else. Don't try doing a web page in APL. Second, the programs efficiently process billion and trillion line databases in a tightly constrained environment. Third, Whitley throws away everything every few years and reimplements from scratch, even the parts that work. He talked about millions of hours of running time without an error. Pretty amazing.
It got me to thinking about the virtues of simplicity instead of the increasing complexity that we are getting into.
[Whitney's] languages are derivatives of APL that do one thing very well and nothing else.
I don't think that's fair. They are general-purpose languages. It's true that they've mostly been used on certain classes of problems so far, but that could just be a historical accident. Who's to say what people would make with them if they found a broader audience?
Whitney is not someone who thinks in special cases.
Edit: I agree with everything else you said, though.
Thanks for posting this.
There's a reasonable argument that APL and derivatives are more expressive than lisp and its derivatives, at the expensive of having a much larger startup cost in terms of things you have to know.
I suspect that in another few years the old-school APL alphabet will be positioned to start making a comeback, via in-software, touch-screen 'keyboards' (thereby eliminating the expense of a very small run of 'funny symbol' keyboards).
It's a bit hard to read in the web format, so you might want to download it as a PDF: http://mags.acm.org/queue/20090203/templates/download_offlin...
I really enjoyed reading that. As russell said, it is far outside my experience, but it is more like I wished my experience would have been.
I read an article on APL back in the early 80s and it sounded wonderful, but I never had an occasion to learn or use it. The closest thing for me now are list operations in Python and the collection functions I'm learning in Clojure right now.
Consider that Arthur's software sells for 25K per core and 8 cores is pretty much a minimum. And that the sole executable is < 300K. Now re-read the article :D.
What an absolute treat. Whitney has done things which, if he hadn't done them, would be called impossible. It's such a shame that his work isn't widely available. There's a whole world inside those 300 kb. It's the only thing I've seen that got me as excited as Lisp.
Tip: Click on the image to zoom in. This crucial functionality is so well-hidden you could call it an easter egg. Horrible usability.