begin  quoting [EMAIL PROTECTED] as of Thu, Feb 03, 2005 at 08:03:34PM -0800:
> ----- Forwarded message from Gabriel Sechan <[EMAIL PROTECTED]> -----
> >Are you saying that experts in any language pretty much develop things
> >at the same rate?
> 
> Yes.

I was under the impression[1] that LOC/day didn't really vary from language
to language for a given programmer.  Thus, the argument for high-level
languages.

> >Are you saying that it is impossible to
> >have a language automate some tedious task (memory management,
> >variable typing) without introducing new problems that are just as
> >bad?
> 
> Yes.  I'd also point out that typing is a bad example.  There's a damn good
> reasons for types- it supplies information to the programmer/maintainer
> about what the variable is.  Languages that eliminate type also eliminate
> this information.  A bad tradeoff for a half dozen keystrokes.

I've been playing in typed languages and untyped (well, lacking compile-time 
type checking) languages, and I see the benefit of both.  When I'm trying
to figure out what's going on, or track down certain sorts of bugs, I 
really like types.  When I'm noodling around or trying to write generic
code, I like typeless.

I don't think there is an "optimal" choice. Both approaches have their
upsides and their downsides.

> As for memory management-  it just isn't a problem.  I can't remember when
> the last time I had a memory pointer problem.  You eliminate a negligible

s/had/discovered/

;-P

> percentage of bugs at the cost of a lot of flexibility.  Worse, in my
> experience, the problems are only partially eliminated at best, as explained
> previously.

Well, asserted, anyway.

> Experience and observation.  About all you can.  The tricky part of all this
> is its a very hard thing to test empirically.  You can't have one person
> write the code in 2 languages, he learns how to write the program on trial
> one.  And having two separate experts write in their prefered language
> leaves questions as to thge relative capabilites of the two programmers, as
> well as the environment they program under.

I think you have to approach it statistically; lots of programmers, lots
of different problems, over a significant period of time.  Repeatedly.

-Stewart "Or figure out how to inspect parallel worlds" Stremler

[1] From reports and summaries of various studies.
-- 

KPLUG-List mailing list
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list

Reply via email to