Autocoding project proposal.

[Courageous]

>The same applies to programming languages.  Each language has roughly
>equal expressive power in the sense of the applications that they can
>be used to develop (give or take). Some people would claim that some
>languages (say, assembler) are much more difficult than others (say,
>SQL), but that has been mostly a religious discussion.

It depends on perspective. Expressive power for _what_? When one frames
the problem in a particular context, it certainly is true that in
those contexts some languages have vastly more expressive power (i.e.,
specifically the character-to-effect ratio) than others _for that
context_. For example, internally at the company I work for we use a
gui markup language which I developed, specifically designed to be
syntactically compact when compared to its general-purpose programming
language alternatives. An attempt to replicate the functionality of this
language ("uiml: user interface markup language") in another environment
would result in _many_ more lines of code than the equivalent UIML
expression. 

However, in UIML you _can't_ express imperatives, conduct conditionals,
and you certainly can't manipulate bits. For those, its expressive
efficiency is zero percent.

The same could be said of trying to replicate SQL in pure C or some
such. Clearly, SQL is _designed for expressive efficiency_ for the
database query and update problem. Anything outside of this domain
will do poorly in comparison.

This also begs the question: is expressive efficiency in every context
the absolute determiner of what you want?

Obviously, no. Sometimes other things intrude. Operating systems aren't
written in C with inline assembly for the expressive efficiency of
the solution. Other factors intrude.

C//