[plt-scheme] Re: PLT-Scheme object system (was re: Image Snip initialization)
On Aug 4, 2006, at 1:33 PM, Nicholas Chubrich wrote:
Thanks for the best coherent argument against parentheses that I have
ever heard. (It's quite possible that someone has said this before; I
am saying that it clicked now.)
> I realize that PLT's mission does not include modifying Scheme itself.
Indeed, we don't see ourselves at all as the guardians of Scheme. As
you may very well know, we have pushed Scheme's boundaries very much.
We have added structures, exceptions, modules, custodians, a rich
syntax system, various inspectors, and a parenthesized version of C
-- to the core. At the periphery you get units, classes, matching,
and a host of libraries. Why? When I speak about the original
DrScheme project to academic/research audiences I mention it a
"project to determine the usefulness of Scheme" as much as "a device
for introducing novices to programming (HtDP language levels)." My
stock answer is that "Scheme isn't useful. PLT Scheme is."
As you may also know, this doesn't endear us to those who are
guardians of the true standards of Scheme. Some of them are on this
list and may want to speak up.
In the end, it is all about the tension between wanting a research
language that a few hundred people use and a useful language that is
potentially interesting to the vast majority of programmers. Both
grow from the same core and similar principle. But they do want to
grow into very different directions.
This tension exists within PLT thought to a lesser extent. All of us
have realized that there is no difference for us between a useful
language and a research language. At this stage, these are the same
things. The only point of contention is how simple and orthogonal the
"evolving language" is to be. Time will tell. Our learning experience
with DrScheme and HtDCP and HtDC will flow into this evolution.
Whether this is Scheme or not, none of us really cares. Just stay
tuned. You will be the first to know.
-- Matthias