🚀 go-pugleaf

In article <o19ft5c@sgi.sgi.com> I said:
+---------------
| The whole 1024 versus 1000 thingy is a mistake that we in the computer
| science community made, and now can't easily undo. It was a way of speaking
| conveniently but sloppily of large numbers which were powers-of-2...
+---------------

Oops! What I didn't say anywhere was that the "mistake" was that we didn't
invent a *different* set of scale factor abbreviations for powers of 1024.
Had we done so, instead of appropriating the "K", "M", etc. from engineering.

Even something so simple as sticking a "2" after the letter could clarify
things, as in, "ttcp reports in K2B/s". Then you'd have a way to handle some
of the thorny mixed-mode things, like the fact that 1 M2B > 1 KK2B > 1 MB.

You see, that's the real trouble with the 1024-based units. It's so easy
to fall unknowingly into a mixed-mode situations, like when "ttcp" reports
"1000 KB/s" (1024000 B/s), or when "df -k" reports "1000" KB used (and "ls -l"
says 1024000).

But even "correct" notation probably wouldn't stop us from saying simply
"one meg" when we mean any of 1000000, 1024000, or 1048576.


-Rob

-----
Rob Warnock, MS-9U/510		rpw3@sgi.com
Silicon Graphics, Inc.		(415)390-1673
2011 N. Shoreline Blvd.
Mountain View, CA  94043