[kde-linux] Thought in desperation

Duncan 1i5t5.duncan at cox.net
Thu Dec 3 02:49:01 UTC 2009

Gmail posted on Thu, 03 Dec 2009 09:05:43 +0800 as excerpted:

> On Wednesday 02 December 2009 23:46:35 Esben Mose Hansen wrote:
>> Not KDE, IEC (International Electrotechnical Commission) by
>> recommandation of IUPAC and NIST. See e.g.
>> https://secure.wikimedia.org/wikipedia/en/wiki/Binary_prefix
>> One man's essentials is another man's fluff. Personally, I welcome the
>> new precision, I was too often caught by the imprecise units that were
>> previously common. 10% error is not acceptable in my opinion.
> First:
> "Binary multiples.
> The new prefixes and symbols for binary multiples standardized in IEC
> 60027-2 are not part of the SI metric system of units." 

This is for the users (not just the players).  The simple fact of the 
matter is that SI units are, by definition, powers of 10 (1000).  
Unfortunately, "borrowing" those same terms for use with powers of two, 
(2^10=1024) is now and has always been confusing, particularly when file 
and memory sizes are standardized to the 2^10 units system, while other 
things, including disk sizes, are standardized to the SI 10^3 units 
system.  My so-called 1 terabyte hard drive is only 931 GiB, because 
they're using powers of ten, but I want to know how I'm going to layout 
powers-of-two based partitions and how many powers-of-two based files 
(using 2-based sector and filesystem block sizes, and 2-based memory 
units when in actual use on the computer, etc) I can fit in them and 
ultimately on the hard drive and in memory.

The "i" suffix abbreviations and the powers-of-two units they stand for 
simply clarifies things, allowing people to just get their work done, 
instead of futzing about trying to figure out whether 10-based or 2-based 
units are involved, etc.  Over time, they'll no longer seem strange, and 
more and more applications begin using the more accurate (and already 
standardized) terms and abbreviations.

> Second:
> OK I might have misunderstood for whom the KDE is. Users here refers to
> people that want to have fun and play around with their computers. Real
> users that use the computer for productive work should use something
> else. I can understand that some very few are interested in binary
> values. But I can assure you most are only simple and plain ISO people.
> It is a pity that all those brilliant ideas and good things in KDE have
> to get lost because some wants to play around with KDE for their own
> fun. Why can they not use their own computer for that and not pollute
> the rest of the world.

The problem is that the 2-power values are standard for file and memory 
sizes, and have been for many years.  That's not changing.  All that's 
changing is that the labels used to describe the 2-power units are now 
more accurate.  For years, KB/MB/GB have normally referred to the power-2 
units when displaying file sizes, etc.  The only thing that's changing 
now is that there's an "i" inserted, so one /knows/ that's what's being 

> Third:
> Why not introduce an additional configuration option, A fun switch. Then
> those who want all this fun can get it. But the default should be For
> ordinary users who use their computers for work.

Actually, a config option for this function is a very good idea.  Much 
like the various CLI fileutils have both --human-readable (-h) and --si 
switches, graphical clients can and arguably should have the same.

Meanwhile, it's worth noting what the clearly assumed default is for 
"human readable".  From the ls manpage:

-h, --human-readable
       with -l, print sizes in human readable format
       (e.g., 1K 234M 2G)

--si   likewise, but use powers of 1000 not 1024

So "human readable" is considered powers of 2 (2^10=1024), not powers of 
10 (10^3=1000).  Why?  Because that has been how file sizes have been 
displayed for decades.  I bet if you look at the changelog, you'll see 
that the -h/--human-readable switch was available quite some time before 
the --si switch became available, because when talking about filesizes, 
powers of 1024 is simply assumed, because that is what has always been 

Now the labels are simply clarified to clear up the ambiguity.

But about that config option...  I'd strongly suggest that if you feel as 
strongly about it as you seem to, you go put in a request for it.  KDE 
has an idea-storming site just for such ideas, with users voting on 
them.  It's possible you could even get it in for 4.4, to be released in 
February, if you hurry, as this one should be relatively simple to 

All that said, regulars here will certainly confirm that I don't believe 
kde4 is ready for ordinary use either.  There's waaayy to many bugs, etc, 
remaining.  I routinely run beta and even scm/alpha versions of various 
applications, so I have a reasonable level of experience with pre-release 
software on which to rely.  The kde 4.2 that KDE first claimed to be 
ready for ordinary use... was more like a late alpha.  There were way too 
many even basic features still missing to be even beta, let alone ready 
for ordinary use.  kde 4.3 is certainly vastly improved, but it's not 
ready for ordinary use either.  Still some features missing, tho mostly 
they are there, but broken in various ways when you actually try to use 
them.  That's normally considered beta.  Based on the rate of improvement 
and various reports, 4.4, scheduled for February, will probably be 
getting to release candidate level -- usable by ordinary people in most 
cases, but still some frustrating bugs that should be fixed before 
release.  Some software projects release their .0s at that level -- the 
observation behind the common wisdom to wait for the .1 releases before 
actually trying to depend on something.  If 4.4 indeed qualifies as rc 
level, then 4.5 should be the first real generally usable kde4.  Given 
the 6-month release cycle, 4.5 should come out in August.

Again, this is nothing new to regulars, as I've written the same thing 
many times, but it's worth repeating.  The FLOSS community has a saying, 
release early, release often, and I 100% support that.  Thus, I don't 
have a problem with the state of the kde4 releases per se, and in fact, 
appreciate the opportunity to use them and get a good jump-start on 
what's coming (I DID say I regularly run beta and even development 
sources!).  There are, however, two problems.  The first isn't the 
quality of the releases, it's that the quality of the releases doesn't 
match what they're claiming.  The second is that they dropped support for 
their very well polished and still very good previous version, 3.5.10, 
the one that actually works, WAAYYY before the new version is even CLOSE 
to sufficiently well working to properly replace it.  Altho... there have 
been some recent developments in that area, and it seems kde3.5 will 
continue to be supported, at a very basic/minimal level, but one hopes 
well enough to continue to work against modern versions of various system 
libraries and compile with a reasonably new gcc, plus get security 
updates, etc, for another year or so, leaving some months to switch to 
kde4.5, which looks to be the first reasonably comparable to kde3.5 kde4 
we'll get.

But as of 4.3, there's absolutely no way I could with integrity and in 
good conscience recommend kde4 for normal use.  And honestly, given the 
issues with such basic things as printing, I must seriously doubt the 
judgment or integrity, if not both, of anyone trying to make that claim, 
yes, including the KDE folks that ARE making it.  And what's worse, 4.2, 
when they /started/ making that claim, was even worse, so much so one 
must wonder if the claim was intended for comic effect!  Either that, or 
they simply haven't the foggiest what it is to actually /use/ the 
software, not just play around with it.  Too bad people actually trying 
to /use/ the software actually believed them!  So I certainly agree with 
you in that regard!

Never-the-less, while 4.2 was certainly a stretch to try to work with, 
even for this guy accustomed to running betas, 4.3 is much improved and 
reasonably easy (and fun) for me, /as/ a guy used to running betas, to 
work with.  I've been watching the changes since before 4.0's release 
(which I mistakenly thought I could run, as a .0 should at /least/ be 
beta quality, boy was I wrong about that!), and every release *IS* vastly 
improved.  It's getting there.  It's just taking some time, and all these 
premature claims about how it's ready for ordinary use aren't helping 
things.  But it WILL and IS getting there, and along about 4.6, kde4 
should be getting very impressive indeed, as it should have surpassed 3.5 
and everything else out there, and it won't be slowing down or looking 
back! =:^)

Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman

More information about the kde-linux mailing list