Re: Libraries and repositories

[Michael Fischer <address@hidden>]

On Wed, Dec 19, Tony Kimball wrote:
> The analogy is clear and simple to my perception:
> 
> Wall's language is practically irreproducible.  As a result, it
> represents a single unified target platform.  As a result, it is
> feasible to assemble a large library of utilities and systems built
> on top of that platform.
> 
> Gosling's language is practically irreproducible.  As a result, it
> represents a single unified target platform.  As a result, it is
> feasible to assemble a large library of utilities and systems built
> on top of that platform.
> 
> Contrast Scheme:
> 
> The language is trivially reproducible.  As a result, it represents a
> vast array of divergent platforms.  As a result it is infeasible to
> assemble a large library of utilities and systems built on top of that
> platform.
> 
> I oversimplify, for clarity and brevity.

And possibly get hung up in the fact that each of the above projects
is a computer language.

Consider the above as general programming tools, without the baggage
that languages have.

Consider that Java was and still is a square in a chess game between
multi-billion dollar companies, and almost as a side effect, a computer
language of considerable repute and utility. It could have simply been
a better toaster, and the dynamics as far as the corporations were
concerned the same. Language, other than legal, issues, never entered into
it (other than as marketing points, but that could be said of the design
of any commercial product).

Perl was a highly domain-specific utility, which grew. No one forked
the project because there wasn't any money in it, it did its job
well enough (better than the competition of shell/sed/awk), and if
you wanted more, you could send Larry patches. Open source projects
from the UNIX world seem to be fairly immune to that sort of forking,
BSD squabbles notwithstanding.

Scheme... a highly successful academic exercise? Doing an implementation
which went sufficiently beyond its much ballyhoo'd standard (and made
paying projects easier) wouldn't have been scheme then, would it?
The inventors made some nice points with it, and good schools use
it to teach valuable lessons. Next?

CPAN is there because people using/implementing Perl were and are 
doing it because it helped them do what they wanted to do, and the
attitude was about sharing. Java has plenty of third party libraries
out there too, just that they are for sale, and not grabs, because
Java is primarily about people making money. Scheme seems not to 
fit either of the above. Care to invent an imperative which suits
the Scheme community capable of driving them to implement that CSAN?

Michael
-- 
Michael Fischer                         7.5 million years to run
michael@visv.net                        printf "%d", 0x2a;
                                                -- deep thought