[GLLUG] Meeting Thursday, September 28

[Jeremy Bowers]

Your points may have some merit, but you chose some spectacularly bad
examples.

Thomas Hruska wrote:
> Microsoft,
> for example, probably had several (internal) business meetings about the
> name choice for the 'dir'ectory listing command.

Extremely unlikely. Microsoft may now be a multi-billion-dollar
behemoth, but Microsoft DOS was one of their first products, you know,
back when Bill was still programming and everybody knew everybody.

Besides, 'dir' clearly harkens from CP/M, which MS-DOS was strongly
based on:

http://www.dcast.vbox.co.uk/cpm/cpm_commands.html

You can find several other DOS commands in that list, including some
obscure ones most people have never heard of.

> But there wasn't ever a vote about the name by those who would use
> [ls].

Also a spectacularly bad example, because the people who used ls *at the
time it was named* probably considered that an unusually sensible name.
This is from the era that gave us "car" and "cdr", for instance. They
were all programmers in an environment any of us would consider actively
hostile nowadays.

My basic problem with the "Programs should have names that say what they
are" argument is summed up pretty adequately with this link:

http://freshmeat.net/browse/63/

Quick, give me 273 distinctive names that all say "Text Editor" to a
novice, yet can't be confused with each other. (Granted, they aren't all
the same thing, but the challenge holds equally with even just 10 names
for a text editor. I can testify from personal experience that there are
people who confused Microsoft Word, WordPerfect, and Lotus Word.)

Insisting that programs should all be named after what they do is like
insisting all cars should have some element of "car"-ness in the name,
lest I confuse a Dodge Stratus with a type of cloud formation, or a
Honda Civic with some sort of civil rights activist. There simply aren't
enough targetted names to meet your demands.