Stopping cows
There's a scene in John Wyndham's marvellous Chocky where the main character asks his father "Why does a cow stop?"
What he means by this is why does a cow's intelligence stop: why are cows smart enough to escape out of an open gate, but not smart enough to see that they could lift up the latch with their noses and escape whenever they want? Why do cows hit a point where their problem-solving ability just stops dead?
I was reminded of this as I read Matthew Huntbach's essay What's Wrong with Ruby this morning. He points out that "Ruby, and the whole scripting language phenomenon is a slap in the face" for the academic language-design community of which he is apparently a part. We have been used to complaining that the programming languages we had developed were so much better than the mainstream programming languages, the only reasons they hadn’t taken off were that they weren’t developed and promoted by big companies, and that the inherent conservatism of industry restricted commercial programming to tried and trusted languages.
Yet a whole stream of new languages: Perl, PHP, Python and now Ruby, each initiated by one person as back-room projects, have been adopted for serious use and achieved many thousands of users. Unlike academically-derived languages, they have given the impression of being thrown together to meet a need without any strong underlying theoretical basis.
That's right. Your languages weren't unsuccessful for the reasons you thought they were, and languages that don't meet your standards of elegance have flourished in actual use. Something you thought was necessary for a language to succeed (corporate backing) turns out not to be necessary after all. Now, let's apply your insights to the rest of your essay (which, by the way, is a perfect example of the kind of thing I was talking about in my post Tightrope Walking). Ruby et al don't have static typing, formal semantics, standards documents, or the guarantee that each object will have a class that determines its behaviour - in short, they're not academic languages - and yet thousands of bright users use them and like them. Could it be that perhaps the lack of these things isn't such a problem as you think? Might they, in fact, have advantages that you're too blind to see?
See that bit of metal attaching the gate to the fence? You might like to think about what you can do with that.
What he means by this is why does a cow's intelligence stop: why are cows smart enough to escape out of an open gate, but not smart enough to see that they could lift up the latch with their noses and escape whenever they want? Why do cows hit a point where their problem-solving ability just stops dead?
I was reminded of this as I read Matthew Huntbach's essay What's Wrong with Ruby this morning. He points out that "Ruby, and the whole scripting language phenomenon is a slap in the face" for the academic language-design community of which he is apparently a part. We have been used to complaining that the programming languages we had developed were so much better than the mainstream programming languages, the only reasons they hadn’t taken off were that they weren’t developed and promoted by big companies, and that the inherent conservatism of industry restricted commercial programming to tried and trusted languages.
Yet a whole stream of new languages: Perl, PHP, Python and now Ruby, each initiated by one person as back-room projects, have been adopted for serious use and achieved many thousands of users. Unlike academically-derived languages, they have given the impression of being thrown together to meet a need without any strong underlying theoretical basis.
That's right. Your languages weren't unsuccessful for the reasons you thought they were, and languages that don't meet your standards of elegance have flourished in actual use. Something you thought was necessary for a language to succeed (corporate backing) turns out not to be necessary after all. Now, let's apply your insights to the rest of your essay (which, by the way, is a perfect example of the kind of thing I was talking about in my post Tightrope Walking). Ruby et al don't have static typing, formal semantics, standards documents, or the guarantee that each object will have a class that determines its behaviour - in short, they're not academic languages - and yet thousands of bright users use them and like them. Could it be that perhaps the lack of these things isn't such a problem as you think? Might they, in fact, have advantages that you're too blind to see?
See that bit of metal attaching the gate to the fence? You might like to think about what you can do with that.