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

The Wanderer inverseparadox
Fri Feb 16 13:37:27 CET 2007


Uoti Urpala wrote:

> On Thu, 2007-02-15 at 17:49 -0500, The Wanderer wrote:
> 
>> Panagiotis Issaris wrote:

>>> Why? Because you dislike the sound of it?
>> 
>> Because it's attempting to say "no, you can't use the terminology
>> which has been established as standard, you have to use this
>> instead or people will be confused". There's a brief comment in my
>> other post on a possible alternative.
> 
> No, the Ki prefix does not "attempt to say" such a thing. People who
> try to get you to understand why the Ki prefix should be used might
> say that, though they probably wouldn't say "has been established as
> a standard" either. But whether people say that or not, it's an
> obvious logical fallacy to use that as a justification for the claim
> that the Ki prefix itself "is an ugly abomination".

...I don't recognize the fallacy involved offhand...

My attempts to respond to the rest of the above have all bogged down
fairly quickly, and I've got limited time before I need to leave for
work, so I'll leave that there. (FTR, I wrote this post from the bottom
up, which is just about a first for me. Note please also that I'm still
only partly awake, so this is almost certainly not the most
well-considered response I've ever written.)

>>> Long? How long? AFAIK kilo meaning "1000" has been in use for a
>>> _much_ longer time than kilo meaning "1024". And, furthermore,
>>> the only scientific field where k has ever had the meaning of
>>> 1024 is computer science. So, I consider it _wrong_ usage.
>> 
>> I never claimed that "kilo" == "1000" had not also been around for
>> a long time, or even that it had not been around for a
>> significantly longer time. All I claimed (implicitly) was that
>> "kilo" == "1024" had been around *long enough* to have become
>> established, such that
> 
> It's been around long enough that people working closely with
> computers can be expected to at least recognize the usage, but hardly
> long enough to justify making the existing units officially
> ambiguous.

On that, we appear to disagree.

>> attempting to force it out is Really Really Bad Form at best.
> 
> Attempting to make the existing units which are much more established
> ambiguous is much worse form.

But that attempt is not being made. Those units *have already become*
ambiguous; at most, those who oppose the imposition of "kibi" and its
like are attempting to prevent them from becoming unambiguous again.

> You also don't see the difference between good and bad ideas: the
> ambiguity is bad and fixing it is an improvement whatever the degree
> of current standardization. Even universally accepted customs are
> worth changing when they cause problems.

On the latter statement, I would tend to agree - which, rationally
(though I don't claim to be exceptionally rational about all this),
would seem to mean that I would have to agree with the former as well.
It's just that I do not think that the current proposed fix is
acceptable. I've proposed one alternate possibility, which you rejected
immediately, but I am not necessarily closed to the suggestion of
additional other possible fixes.

(And, yes, I do see the difference between "good ideas" and "bad ideas". 
That's why I'm so far willing to compromise as to acknowledge that there 
is, indeed, a need for unambiguous terms and that it can perhaps be 
grudgingly acceptable to use these terms as options in some contexts. 
The fact that they fill that niche, however, does not reduce the problem 
with the terms themselves.)

>>> On the contrary: To give an existing and much used prefix an
>>> additional meaning (superset of meanings) was beyong the
>>> authority of whatever group that actually did this.
>> 
>> Ah, but I don't think it was consciously done by any group; I
>> believe it was done "in passing* by simply using the term that way
>> because it was close enough, and became a de-facto standard usage
>> thereby. Nobody arrogated to themselves the authority to redefine
>> the existing term.
> 
> It became a "de-facto standard" for some people, but most people have
> never accepted a change in the meaning of k, M etc. As the usage
> spread it also became obvious that it was a bad idea.

"Most people" also do not appear to have been working routinely with the
units in context of which the "1024" meaning was used - that being,
units of data, specifically bytes.

(For that matter, I'm not necessarily willing to accept that claim of
"most people" without some evidence.)

> It seems that you view your own personal usage as the standard that
> everyone should accept and are unable to adapt to changes. I'm quite
> sure you would not consider writing "KiB" such an abomination if
> you'd learned that from the start.

Actuall, it's closer to that I view my own personal usage as the
standard that *I* am willing to accept. (You'll note that even my
original post expressed the penalty for using the terminology in terms
of how I would feel about it.)

And no, I wouldn't consider it an abomination if I'd learned it from the
start, that's true - much the same way as people who learn it without
having known the previous usage, and hence not knowing any better, will
find people who consider it an abomination incomprehensible. But I
*didn't* learn it from the start, because *it didn't exist* from the
start - so that's not really a workable argument.

>> The invention of "kibi" and so forth, by contrast, *was* done by a
>> specific group, though I forget offhand which one and I've got
>> little enough time before class that I'm not going to look it up.
>> That group are thus attempting to claim authority which they do not
>> - and, perhaps, which any "they" fundamentally *can* not - have.
> 
> Again you're ignoring the distinction between bad and good ideas.

I'm not saying that introducing an unambiguous form is not a good idea.

I'm saying that the form which was chosen, and the attempt to force it
on the world in place of the established usage, is fugly at best.

(I've been reminded irresistibly for quite some time now in this
conversation of the fight to introduce the metric system in America...
and even though there I agree that the metric system makes a lot more
sense, and I do not have the immediate revulsion for it which I do have
for the terminology at hand here, I still don't automatically think in
terms of it.)

> Yes, the ambiguous usage wasn't really designed or actively spread by
> anyone (which is natural given that it's not such a good idea).
> Noting that it had become widespread enough to cause problems (which
> can happen even for bad ideas if there's no obvious alternative) and
> then designing a better alternative did take conscious effort from
> someone. That's not a reason to oppose it.

But it also involved that someone saying "I get to decide for everyone
else" - which is incredibly arrogant at best. (And, again, the
particular thing which was designed - which appears, at a glance, to
have been chosen in part for its lack of radical difference from the
established usage, presumably so as to be less controversial - is not an
attractive option; its very similarity seems to worsen its likelihood of
acceptance.)

This arrogance, and the associated imposition, is at or near the root of
my perception of "kibi" as an "ugly abomination" (see top).

-- 
       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