[FFmpeg-devel] GSoC with FFMpeg waht a combination!

Uoti Urpala uoti.urpala
Wed Mar 26 03:55:14 CET 2008


On Wed, 2008-03-26 at 01:43 +0000, M?ns Rullg?rd wrote:
> Uoti Urpala <uoti.urpala at pp1.inet.fi> writes:
> > I wouldn't be OK with it because practical ambiguity is rare. However if
> > such ambiguity was common then the terminology would need to be changed.
> > Uses of base-10 and base-2 units are not separated clearly enough that
> > you could rely on context to disambiguate them.
> 
> It worked fine until hard drive manufacturers saw a way to screw their
> customers.  Looks like you swallowed their bait.

I think that is a stupid argument. It should be obvious enough that the
people who moved to fix the usage of the units had better motivation
than the actions of HD manufacturers. Plus talking about "swallowing
their bait" is nonsense - changing units from base-2 to base-10 is not
beneficial to the manufacturers overall; it was only doing the change
before competitors that could give a temporary marketing advantage.

> > If for example 'ls -l' shows file size as 183297384 bytes then I'll say
> > it's 183 MB. Would that really be obvious from context if I also used
> > 'MB' when talking about base-2 units? Or do you claim that I must
> > convert the value to base-2 units?
> 
> Of course you should.  That's what ls -lh (GNU) does.

"Should" why? What's the advantage compared to base-10? Nowadays the
underlying filesystem implementation is abstracted far enough away that
I see no reason to consider base-2 units in normal usage. Why should I
treat file sizes differently from other areas, which use steps of 1000?

As long as we write the numbers in decimal 183,297,384 B / 183,297 kB /
183.3 MB / 0.1833 GB is more practical than the base-2 183,297,384 /
179,001 / 174.8 / 0.1707. The latter is harder to convert when using
decimal, and also has the ugliness of "extra large" values between 1000
and 1024 (1023 before moving to 1 of next unit).

So can you give a reason why it would be desirable to use base-2 units
when talking about file sizes? I mean independently of what you call the
units or terminology issues in general. The only benefits in ordinary
use I can think of are in special cases like files around the FAT size
limit which is a power of 2. All commonly visible differences favor base
10.





More information about the ffmpeg-devel mailing list