Subject: lab director duties
From: Tom Lord <lord@regexps.com>
Date: Sun, 21 Oct 2001 04:09:39 -0700 (PDT)



	eichin chirps in with:

       > But that may have suffered from the "all scheme projects are
       > doomed" syndrome


Heh.  Doomed to provide an eventual source of embarrassment and
failure to those who oppose them for stupid reasons, perhaps.

Until then, us believers in the church of lambda will just work on
learning to Take Our Time and Get It Right.  It's *that* good.



	> a random example from the net:

	> New OSKit 0.96 released and a "Scheme machine"

That points to a trend: new technologies for very high level
programming that sneak up on established paradigms and then plow them
under a snowbank of Superior Abstractions.  Compiler technology is
another example where high-100s of K of hours of effort on old
technology is ripe to be blown away by low-10s of K of hours of effort
on higher level expressions of the same ideas (lcc, using awk no
less(!), was a stern warning sign; the antlr tools, too.)

All of this points to another reason why FSBs have to (1) focus on
engineering process rather than products and (2) create R&D
organizations that keep organized track of what's going on at the
bloody bleeding edge, leading little experiments to differentiate the
purely theoretical from the all-to-practical.

Hey -- it's the weekend:  I'm not supposed to be reading mail.

-t