Lambda the Ultimate

inactiveTopic Calling LtU editors
started 8/4/2003; 3:16:41 AM - last post 8/5/2003; 9:47:17 AM
Ehud Lamm - Calling LtU editors  blueArrow
8/4/2003; 3:16:41 AM (reads: 1149, responses: 9)
Calling LtU editors
It's been awhile since I last prodded the LtU contributing editors, so I figure it's time...

Many of you have been silent for quite some time, which means we haven't heard anything new about type theory for natural language semantics, Scheme DSLs, Erlang and the like.

Come on, you know who you are. Language lovers, stand up and be counted!


Posted to admin by Ehud Lamm on 8/4/03; 3:17:07 AM

Luke Gorrie - Re: Calling LtU editors  blueArrow
8/4/2003; 9:02:49 AM (reads: 627, responses: 1)
One thing I wish I could post is Knuth's The IBM 650: An Appreciation from the Field, which appeared in his Selected Papers on Computer Science. Alas, those unhelpful folk at the IEEE are charging $19 for a copy of that single article in PDF. It is really good to read about the hackerish side of Knuth. (I only recently learned that he's an Emacs hacker too.)

Ehud Lamm - Re: Calling LtU editors  blueArrow
8/4/2003; 9:15:03 AM (reads: 651, responses: 0)
You could tell us why you find it interesting though.

Luke Gorrie - Re: Calling LtU editors  blueArrow
8/4/2003; 9:44:55 AM (reads: 617, responses: 4)
It talks about his early programming experiences. Staying up all night debugging, squeezing featureful programs into tiny memories, competing to write programs that fit on a single punched card, being amazed by the first programming languages. And of course, his affection for the IBM 650. I enjoyed this because I can relate to it much better than most of his other writing (after an s/IBM 650/Amiga 500/ at any rate).

He also mentions that he tends to think that machine code makes the best first introduction to computer programming. I'm not sure if he ever said this explicitly in TAOCP.

Generally the tone is much like his turing award lecture, which has been covered previously.

Ehud Lamm - Re: Calling LtU editors  blueArrow
8/4/2003; 9:47:24 AM (reads: 640, responses: 3)
He also mentions that he tends to think that machine code makes the best first introduction to computer programming. I'm not sure if he ever said this explicitly in TAOCP.

It's his preferred approach, obviously. Kind of silly, but then he is from a different generation.

Marc Hamann - Re: Calling LtU editors  blueArrow
8/4/2003; 10:42:26 AM (reads: 661, responses: 2)
Kind of silly, but then he is from a different generation.

I'm not sure I agree it is silly; impractical perhaps (as will be familiar to anyone who has looked through TAOCP for an algorithm they needed right away and had to wrestle with MIX to get at it ;-))

I think that everyone who wants to have a deep understanding of programming would benefit at some point from the experience of assembler programming. I think it demystifies the compilation/interpretation process and clarifies the fundamentals of "what is going on" when a program runs.

Ehud Lamm - Re: Calling LtU editors  blueArrow
8/4/2003; 12:02:50 PM (reads: 667, responses: 1)
I think that everyone who wants to have a deep understanding of programming would benefit at some point from the experience of assembler programming.

Obviously. But this is nither the only nor even the most important aspect of programming.

Worse, there are so many different machine models (and VMs) that's it's much better to teach this sort of thing via a formally defined abstract machine (coupled with an executable implementation, for practical practice). But this part should come well after getting used to what I call "algorithmic thought."

Marc Hamann - Re: Calling LtU editors  blueArrow
8/4/2003; 12:14:11 PM (reads: 693, responses: 0)
much better to teach this sort of thing via a formally defined abstract machine

I think this is one of the strengths of SICP: they introduce abstract machines within the framework of a higher-level language, giving the best of both worlds.

I wasn't really a beginner when I read it though, so I can't tell how effective it really is in this regard. ;-)

Kimberley Burchett - Re: Calling LtU editors  blueArrow
8/5/2003; 7:27:28 AM (reads: 533, responses: 0)
I have to say that it wasn't until I learned assembler that I finally understood pointers, or the vague claim that "code is another kind of data", or why compilation was important.

If I hadn't learned assembler, I might have plateaued at a much lower level of appreciation than I currently have.

However, I wouldn't say that it should be the first programming language -- that would probably turn off a lot of people before they even got started. But for those people who can stick it out, or who have to stick it out (e.g. because of course requirements), assembler might not be such a bad way to start. They do include assembler in the first semester of CS at Harvard...

Luke Gorrie - Re: Calling LtU editors  blueArrow
8/5/2003; 9:47:17 AM (reads: 517, responses: 0)
Of course, Assembler has turned lots of people on to programming as well. It has been a major hobbyist programming language throughout computing history as far as I know, and was still the favourite for discriminating hackers when I was a teenager (in the early 90s.)

Your perspective may vary :-). But no way would I have traded in my Amiga 500 for an abstract machine (no matter how elegant its instruction set!).