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

Måns Rullgård mans
Wed Mar 26 01:26:45 CET 2008

The Wanderer <inverseparadox at comcast.net> writes:

> Uoti Urpala wrote:
>> On Tue, 2008-03-25 at 16:56 -0400, The Wanderer wrote:
>>> Uoti Urpala wrote:
>>>> Will you please stop describing the issue in such intentionally
>>>> misleading and propagandistic terms?
>>> I am of the opinion that I have given quite enough ground already
>>> in acknowledging that other people can use the term if they want
>>> to.
>>> I am also not inclined to be particularly (lenient? forgiving?
>>> generous? all of those imply negative things about my viewpoint
>>> which I do not intend...) towards someone who brought up the
>>> subject by intentionally misunderstanding the term when there was
>>> no ambiguity.
>> That's not a fair description of how the subject came up either.
> True, and I apologize.
>> My response to the initial post that used 'k' incorrectly only had a
>> parenthesized remark "I assume you meant Ki, not k". I don't see how
>> you could call that "intentionally misunderstanding".
> Phrasing it that way does overstate the case. However, since it was and
> remains entirely plain what the intended meaning was, responding as you
> did (by making such a parenthetical comment rather than simply ignoring
> the usage with which you disagree) is something which I have difficulty
> seeing as anything other than intentionally pretending to not understand
> so as to have the opportunity to put forward a "correction" and thus
> advance the terminology you prefer. (That in turn is poorly phrased, and
> may be something of a run-on sentence, but seems somewhat closer to what
> I intended.)

The word that springs to my mind as describing Uoti's tone here is

>>>> k meaning 1000 is the original and existing usage, and by a huge
>>>> margin.
>>> Not when measuring data. It may be that k was historically used to
>>> mean 1000 in some cases there as well, but from what I can tell it
>>> has been vastly more often used to mean 1024 when referring to
>>> units of data, and that is the usage which "ki" and the like are
>>> attempting to supersede.
>> Except when talking about transmission of data etc...
> The only case I can think of where this is the case is when the units
> involved are e.g. baud, which since they are measuring bits are not
> necessarily the same thing. (If one pursues that line of argument, which
> I'm not sure I want to, it might be appropriate to restate my
> disagreement above as "Not when counting bytes".) I have further
> reservations about whether the baud is strictly a unit of data, but I
> would have to reexamine my "basics of networking" class materials from
> some while ago to be potentially able to build up an argument in that
> direction.

The baud is the unit traditionally used to measure symbol rate in data
transmission.  The amount of data represented by one symbol depends on
the modulation scheme employed.  For instance, with QPSK (widely used
for satellite broadcast), each symbol represents 2 bits, and with
64QAM each symbol represents 6 bits of data.  The term baud seems to
have fallen into disuse, having been (mostly) replaced by actual bit

>> And the base-2 units are less practical when talking about large file
>> sizes too.
> Practical for what purposes?
> It's less close to an acceptable approximation of the power-of-ten
> value, yes - but since the very use of the power-of-two units leaves the
> power-of-ten units out of the picture, the question of approximating
> them becomes immaterial.
>>> That said, unless other people jump in such that the discussion
>>> expands beyond just me being annoyed, I intend to drop the subject
>>> here rather than produce an extended thread of offtopic argument.
>> There also was a thread about the general subject already last year.
>> (http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/2007-February/023218.html
>> is where it stopped back then.)
> I was fairly heavily involved in that thread, if it's the one I recall.
> The argument in there is what has led me to reluctantly concede for lack
> of supportable argument otherwise that others are allowed to use the
> term if they want to.

I see no reason to tolerate this so-called word to any greater extent
than we would the renaming of other established words.  Suppose, for
instance, that I began insisting on calling C functions "finctions"
(perhaps to avoid "confusion" with the pure mathematical concept of
function).  Would everybody be OK with that?  I'd like to think not.

M?ns Rullg?rd
mans at mansr.com

More information about the ffmpeg-devel mailing list