I was all set to dive into “No Silver Bullet” for the next Code Reads, but given last night’s news of the passing away of John Backus, father of FORTRAN, I thought I would do a quick revision of the plan.
The next Code Reads will focus on Backus’s 1977 Turing lecture, “Can Programming Be Liberated from the von Neumann Style?” It’s full of equations and math notation that, superficially at least, look daunting to this reader — but I will give it a try, and perhaps the collective expertise of all of you will help bolster me in those areas where I falter!
Post Revisions:
There are no revisions for this post.
Mark C. Chu-Carroll has an entry discussing Backus’s functional programming language.
Oh, Scott… Don’t do that to yourself.
Ironically, since one of his lasting contributions to the field was a method for formalizing programming syntax, Backus did not have the knack for inventing clean, intuitive syntax. Fortran is notoriously tricky, as in this example from the dragon book:
In
DO 5 I = 1.25, “DO 5 I” is an identifier that is assigned a floating-point value. InDO 5 I = 1,25, “DO” is a keyword introducing a loop.It seems that his FP syntax is pretty bad too. He’s assigned cryptic symbols to “combinators” that could have (and had, even then, e.g., in Lisp) natural language descriptions. E.g. “alpha” is “map”, and “/” is “fold”
That Fortran example would have worked better if the formatting had come through. Sorry.
I’m providing my assessment of it now because my Internet access will be sporadic in the near future. Warning: I pull no punches.
http://my.opera.com/Vorlath/blog/2007/03/21/code-reads-backus-1977-turing-lecture
OK, this is great — I feel like we’ve definitely hit some new level of participation here now that the discussion begins without me :-)
Seriously, thanks — I’m going to read the critique carefully, but not till after I’ve read the whole Backus paper first… Meanwhile, hope your sporadic internet access means you’re getting a vacation of some kind and not that your experiencing a telecommunications breakdown!
Hi Scott,
My wife’s parents also live in Ashland, where John Backus last resided.
It turns out that Backus’ daughter is our family veterinarian in Ashland. My father-in-law knew Backus for a few years because of this. And yet I don’t believe he knew Backus was the inventor of Fortran.
I’m guessing that the old “six degrees of separation” could probably be five or even four when talking about the software community.
Patrick
Here’s a comparison and example about what I talk in my first link above. You can check this afterwards. It has a couple diagrams that may describe things better than just words.
http://my.opera.com/Vorlath/blog/2007/03/23/why-the-black-box-analogy-does-not-work-for-function
Yeah, my Internet access is in limbo. It’ll be back up next week tho. In the meantime, I may well have a vacation of sorts with the result of another project (real estate related), but I feel kind of stupid that I can’t raise the final 5K cuz I don’t know anyone. We have a plan B tho although less lucrative. If this works, vacation day will be every day. Anyways, after that I can finally show to the world what I’ve been talking about all this time by finishing my own software development tool. If anyone has articles on IC type software, let me know. There’s one particular one I’m looking for that said the only reason we didn’t use it was because the economic model for it would not sustain companies. Would programmers still want easy and reliably built software if it meant there was no money in it? What if that was the tradeoff?
Scott, did you take my advice and throw in the towel on this one? Or have you just been busy with other things?
It’s been ridiculously long and I’m woefully late on this but I am actually finishing it up this week and expect to post soon on it. Thanks for keeping me honest :-)