[Ffmpeg-devel] [PATCH] Print KB for 1024 bytes

The Wanderer inverseparadox
Thu Feb 15 23:42:35 CET 2007


Panagiotis Issaris wrote:

> Hi,
> 
> M?ns Rullg?rd schreef:
> 
>> Panagiotis Issaris said: [...]

>>> I do not think that the meaning of units should be context
>>> dependent.
>> 
>> They are.  Just accept it.
> 
> Units such as "s", "m", "g"

"m" is obviously "meters", and "g" could be "gallons" or "the standard
unit of gravity", but what's "s"?

> and prefixes such as "M", "G", ... are _not_ context dependent.

Actually, yes, they are; "M" means either "1000000" or "1048576",
depending on context, just as "K" means either "1000" or "1024"
depending on context.

I agree that this is ambiguous, and that there are some situations in
which ambiguity is fundamentally unacceptable. That is the only reason
I'm even remotely uncomfortable rejecting "kibi" and "mebi" and so
forth: they do fill that potential need. However, that doesn't change
the fact that I still reject them.

I would hazard a guess (I haven't spent excessive amounts of time
thinking about this) that the "correct" way to fix it would be to devise
*separate* prefixes which have unambiguous meanings, which could *not*
be conveniently mistaken for the current usage, and then *allow but not
require* those prefixes to be used - leaving people free to use the
existing ambiguous form if they want to. (By contrast, the current
proposed unambiguous terminology works only if people stop using the
ambiguous form entirely, because one of its "unambiguous" terms is
exactly the same as the ambiguous one.)

However, this has not been done, and what *has* been done is
unacceptable. So, here we are.

-- 
       The Wanderer

Warning: Simply because I argue an issue does not mean I agree with any
side of it.

Secrecy is the beginning of tyranny.




More information about the ffmpeg-devel mailing list