[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

picking names for functions



It seems petty, to me, to fight over whether
"CHAR" describes a low-level thing or a 
high level thing -- I mean to argue over
the names of things to such a degree.

If one insists that there is such a principled
importance to names that a minority opinion
should be extended indefinitely in arguments
then one ought to consider if they are fighting
the wrong battle:

With so many interests and perspectives around
the table, perfect agreement on names will always
be unlikely.

A different open-agenda item -- the sense that
a module system is needed, especially one that
allows programs to create customized top-level
environments in a standardized way importing
procedures and data but not necessarily names
from other environments -- might be more relevant
to the pedantic concerns of naming.

Given a portable module system, advocates of
one naming aesthetic as opposed to another
can compete for mind-share:  if one naming
aesthetic emerges as best, it can later influence
the lower-level parts of the standard.  By that
time, the superiority of one aesthetic over others
will be closer to objectively observable (if
there is such a thing, which I doubt).

In the meanwhile, when there are disputes about names
and any of the editors are convinced -- they 
can end debate by taking a quick vote among themselves.
That's about as good as it'll ever get, imo.

At some point (and I'm seeing this in my recent 
sampling of the thread), discussions like this FAIL
in the specific way of devolving into little more 
than informal rehearsals of "highlighted axioms and
theorems of standard Unicode" -- people wind up 
doing nothing other than reciting well-known dialects
and explaining to one another, yet again, that 
unicode programming is tricky to do well and that
there is no single, simple, universal Right Thing.

It's very enlightening,
I'm sure, for people unfamiliar with Unicode but who are, 
for some reason, closely following these arguments -- but I
don't see how it contributes much to SRFI or R6.

(A simple "these names suck and I suggest X because..."
and leaving it there in spite of replies in kind 
would have had as much information conent.)

-t