[FoRK] coding at low reynolds number

Dave Long dave.long at bluewin.ch
Tue Apr 9 07:08:29 PDT 2013


Judging by the observed speed of diffusion of ideas from mathematics  
into CS, let alone general hacking, the intellectual environment is  
viscous indeed.

For example, consider that one may find a paper from the Yellow  
Submarine era[0], which not only has the following definition[1]:

let rec concat(xs1, xs2) = cases xs1:
	cons(x, xs3)	: cons(x, concat(xs3,xs2))
	nil()			: xs2

but even contains a commutative diagram[2].

-Dave

[0] Burstall, "Proving properties of programs by structural induction",
Computer Journal Vol.12 #1 (Feb. 1969) pp. 41-48
[1] speaking of viscosity: having recently had to dash out some  
android code to deal with a Dick Tracyesque watch, it would appear  
that it'll be sometime after the next decade (ca. 1966 + 2010 -  
1946?) before the equivalent of eclipse will, out of the box, allow  
me to avoid hand-translating pattern matching.  After all, if these  
IDEs can make a VAX with 10 users appear responsive, one might hope  
they'd be doing something useful...
[2] cf http://math.uchicago.edu/~wald/lit/green_eggs.txt



More information about the FoRK mailing list