[PRL] Fwd: Three Schools of Expression

Mitchell Wand wand at ccs.neu.edu
Mon Jul 23 15:46:42 EDT 2007


Here's a message Dan Friedman wrote me a few days ago.  I'm curious to see
whether the patterns of communication he discusses are consistent with your
impressions.

Can "schools of communication" be usefully defined for other languages?

--Mitch


---------- Forwarded message ----------
From: Daniel Friedman <dfried00 at gmail.com>
Date: Jul 16, 2007 8:37 PM
Subject: Three Schools of Expression
To: Mitchell Wand <wand at ccs.neu.edu>

Mitch,

What really distinguishes three kinds of communication.  Let's
call them schools for want of a better name:

Haskell School
Scheme School
Prose School

The Haskell School uses very short names and gets away with it
because the types carry so much information.

The Scheme School uses rather long name to make up for the lack
of types

The Haskell and Scheme Schools treat the long names and types as
a kind of a poorman's documentation.  It is what you would expect
from those who find writing unnatural (very few of us write novels
for a reason).

The Prose School, which I believe in, is a compromise between the
Haskell and the Scheme School.  The Prose School knows that there
are types, but would rather spell out the types while writing, we
we might say that member takes a list of values and asks of a value
is a member of that list.  The types of values for the Haskell
School all have to be the same and the value looked up as to be that
same type, but that is all rather obvious.  The Prose School feels
that the structure of the program is as important as what it does,
so making sure the structure of the code is intact and swallowable
is critical, so although the Prose School uses decent names for
variables, they believe that what is said about the code, including
the names of critical variables matter.

So in the Prose School, the program is both an incidental and the
major artefact.  This is a difficult attribute to maintain, but
one worthy of the effort.

This is me urging shorter names.

... Dan
-------------- next part --------------
HTML attachment scrubbed and removed


More information about the PRL mailing list