When should software applications be programmable?

Paul Robinson <tdarcos@mcimail.com> writes:
>How do we determine when something is a "real" interpreter of a "real"
>language, and when it doesn't quite reach that point?


With all due respect, I think that you are asking the wrong questions, at
least from a practical perspective. A more intersting line of questions is
this:


When should "conventional" software applications be programmable? What
sort of objects and opertaions should the "programming language" of the
software provide to the user? Should the language be highly specialzed
(like a database query language) or very general (like a C interpreter for
an editor)? How should the interpreter be structured? To how much of the
basic software should the language have access.


>Here's the reason why: at what point does a programmable application cross
>over from just a means to assist the application in doing whatever it
>primarily does, and becomes a programming language in and of itself?


I think that you will find that, in general, there are no hard lines of
meaning. Sure, you can apply the Turing equivalence test but even many
programmable hand calculators can pass that. File i/o is another
reasonable criteria but that, too, is easy to add at a trivial level
without meaningful support.
--

Return to the comp.compilers page.
Search the comp.compilers archives again.