Who's minister of propaganda this week?

Paul Prescod paulp at ActiveState.com
Sun Mar 18 03:05:49 CET 2001


Michael Chermside wrote:
> 
> ...
> 
> Yes, I can see that. But sometimes it's awefully nice to have the
> compiler do it for you automatically. 

I absolutely agree that compiler checks of code can catch thing in
statically typed languages earlier and this is a good thing. 

> It certainly doesn't hurt.

I think it does hurt because you end up writing more code and that code
is prone (like any code!) to the subtle semantic errors that are really
tough to find. So you are trading trivial, relatively easy-to-find
errors for subtle, buried in the code errors. It's like driving twice
the distance to avoid a particularly busy intersection. You feel safer
because you can look at the news and say: "Whew, I missed a car crash at
that intersection today" but you have no idea how much you increased
your risk of an accident just by driving farther.

> (But see Alex's response, where he argues that perhaps it DOES hurt,
> for other reasons.)

I think it takes a lot of work with both kinds of languages to recognize
that you lose something significant in exchange for the earlier error
messages. And I think that slowly over the next few years, programmers
are going to to come to that realization.

It is highly akin to the GC situation. A few years back, garbage
collection of any sort was considered unaffordable for real apps (even
in the face of evidence to the contrary) merely because most programmers
did not have real-world experience with it. Now, nobody designs a new
language without GC. Of course it wasn't all just education. Moore's law
helped too. The situation is the same with dynamic typing. Moore's law
and education will bring programmers over.

-- 
Take a recipe. Leave a recipe.  
Python Cookbook!  http://www.activestate.com/pythoncookbook




More information about the Python-list mailing list