Human readable output

Tony at Tony at
Tue Nov 15 10:47:28 GMT 2005

Christoph Biedl  Tuesday, November 15, 2005 4:20 AM
> Manuel L?pez-Ib??ez wrote...
> > I vote for this feature. In du and df commands, this is invoked with:
> > -h, --human-readable
> >      print sizes in human readable format (e.g., 1K 234M 2G)
> >
> > --si  likewise, but use powers of 1000 not 1024
> >
> > Currently, in rsync, -h is the short form of --help.
> Please do not, I repeat, do not three terrible things the gnu folks did
> for ls, df etc:
> 1. Continue the prefix confusion which is using k, M, ... for 1024,
>    1024*1024 etc. Their meaning is defined to 1000, 1000*1000 ...
>    If you really want to use the 2^(10*n) quantities, use the
>    not-so-new-anymore non-ambiguous prefixes ki, Mi, Gi.
> 2. Show a bias against the SI system by dropping the short-hand notation
>    -H (now --si) for it while not touching the confusing -h (which
>    counts in powers of two).
> 3. Re-use parameter names. "du -H" greats me with
>    | du: WARNING: use --si, not -H; the meaning of the -H option will soon
>    | change to be the same as that of --dereference-args (-D)

For whatever it's worth, agreed.
There has to be something goofy about using base-10 numbers to express
base-2 quantities.
With something like rsync, you will have mixtures of bytes, thousands of
millions of bytes, billions (English or American?) of bytes.
It would be nice to be able to add such quantites easily and accurately.
Part of that is the ability to put meaningful digits AFTER the decimal

How many is 12.34MB if 1MB is 1024*1024?
I think things like 128k of memory are something like a baker's dozen.
A dozen is still 12 and methinks a k should still be 1000.

More information about the rsync mailing list