A bit of trivia: what does usr stand for?

[Bruce R. Montague]

I don't know anything about the derivation of the name
"usr", but it is of mild interest that one reason 3-letter
filename suffixes are so common (.exe, .jpg, .com, .txt, 
etc..) is that many old DEC machines (including the PDP-11)
used a character set called "rad 50" (or RADIX-50, Radix    
50, etc..). Rad 50 had 40 characters (50 in octal), and 
could store 3 characters in 16 bits.  A number of filesystems
used a single 16-bit word for the filename suffix in dir 
structures and such, and thus the 3 char maximum on suffix
lengths.

If you ever had to work with rad-50 at low-level, you are
liable to remember it, because characters were not aligned
on bit boundaries but had to be inserted and extracted
using a process similar to that used for converting between
base 10 numbers in ASCII text and binary numbers.

DRI used 3-letter suffixes for CP/M, likely because the
RT-11 system at the NPS that seems to have strongly 
flavored CP/M used RAD-50. MS DOS, of course, inherited
the convention from CP/M. None of this still has anything
to do with Unix's "usr", but 3 letter name limitations
were common at one time. RSX-11, another PDP-11 OS, had
a one word rad-50 3 character maximum on command names;
this (or other similar systems) may have influenced the
use of common 3-character names such as "ddb"... or maybe
not.

There were a lot of similar character sets - Univac
Fielddata, CDC Display Code.... These were all different
of course. Dealing with char conversion in mainframe days
was a mess.

Of course, today these days 3 char suffixes (.htm) are 
probably just common because the statistics work out nice.


 - bruce