[courtesy cc of this posting sent to cited author via email]
email@example.com (Paul Prescod) writes:
:Sun is pushing Java more than ever. The reason for their interest in TCL
:seems to be this dichtomy between systems programming languages and scripting
What dichotomy? Repeating something doesn't make it true. Perhaps you
missed this brief article I posted last week in response to an article
in comp.lang.python by firstname.lastname@example.org (Jim Phillips).
Jim Phillips wrote:
:This was an interesting paper, and I agreed with a lot of the comments on the
:advantages of scripting languages. I do, however, feel that the bashing of
:object-oriented programming was unjustified for two reasons:
:1) False separation of "scripting" or "typeless" languages from
:object-oriented languages. Objective-C and Python both use a single type for
:objects and do run-time member lookup allowing any object which can handle
:the necessary calls to be used with any package.
(As does Perl, Tom notes tactlessly.)
To me, the whole notion that there exist two kinds of languages:
TYPE A TYPE B
rapid-development product-oriented deliverables
is a facile oversimplification loaded with false dualism. Reality would
seem to be that items in those columns float freely back and forth
depending on many factors, and pretending that they are two sides of a coin
ignores many shades of grey as well as unrelated orthogonal axes.
Cannot compiled languages glue things together? Cannot interpeted
languages be used for systems programming (think of BASIC-PLUS and RSTS/E
for a non-Unix example)? Is there some reason why interpreters should
not have objects, or compilers dynamic/fluid types? I can compile awk
or Perl into C, and thence to assembler and machine language. Or I can
run C or Pascal in an interpreted environment. Does that all of a sudden
change what they are?
I don't think so.
Many of the other points are valid, but it just seems too much a case
of black and white.
Tom Christiansen email@example.com
X-Windows: Even your dog won't like it.