Re: glib uses wrong prefix for base-2 units



Benjamin Drung wrote:
Hello,

g_format_size_for_display uses the wrong prefixes for units that are
counted in power of two. The SI defines following prefixes:

k = 1000
M = 1000 k
G = 1000 M
...

Please use the IEC standard for binary prefixes:

Ki = 1024
Mi = 1024 Ki
Gi = 1024 Mi
...

Oh god, let the flames begin...

* It is the correct usage of the standards.

True, but just because there is a standard, it doesn't mean it's a good standard.

* It would avoid ambiguity and consumer confusion:
http://en.wikipedia.org/wiki/Binary_prefix#Consumer_confusion

Arguably, using KiB etc. in the user interface could confuse the user just as much.

* The users want it. E.g. look at brainstorm:
http://brainstorm.ubuntu.com/idea/4114/
http://brainstorm.ubuntu.com/idea/17839/

You mean: some vocal users of a single Linux distro want it.

* The Linux kernel uses it (man units).

No, the Linux man-pages project, which just happens to be hosted on kernel.org, uses it.

I'm not completely against this kind of change (aside from the fact that I think "kibi", "mebi", and "gibi" sound retarded), but why change now? There are previous discussions about this:

http://bugzilla.gnome.org/show_bug.cgi?id=301838
http://mail.gnome.org/archives/gnome-i18n/2007-May/msg00109.html

... that resulted in favor of leaving the prefixes as they are. What has changed in the meantime (these two discussions are 4 and 2 years old, respectively) that warrants revisiting this? Has there been new rampant widespread adoption?

	-brian

P.S. As an aside, one could also argue to dispense with the use of the powers-of-two interpretation in user interfaces entirely. If Seagate/WD/etc. sells you a 500GB hard drive (500 billion bytes), wouldn't you expect to see "500GB" in your file manager as the full capacity (well, minus filesystem overhead)? A non-technical user wouldn't expect to see 476, regardless of whether or not it's followed by "GB" or "GiB".

Aside from the (mostly irrelevant, slight) speed efficiency of doing division by 1024 via bit shifts rather than division by 1000 by actual division, what use does the powers-of-two interpretation have today?

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]