W3C home > Mailing lists > Public > www-smil@w3.org > October to December 2007

Re: [SMIL30 LC comment] rounding in 17.4.4 and 19.4.4 ( LC-1801)

From: <tmichel@w3.org>
Date: Mon, 12 Nov 2007 11:34:45 +0000
To: Dr. Olaf Hoffmann <Dr.O.Hoffmann@gmx.de>
Cc: www-smil@w3.org
Message-Id: <E1IrXYz-0003sS-3M@wiggum.w3.org>


 Dear Dr. Olaf Hoffmann ,

The SYMM Working Group has reviewed the comments you sent [1] on the Last
Call Working Draft [2] of the Synchronized Multimedia Integration Language
(SMIL 3.0) published on 13 Jul 2007. Thank you for having taken the time to
review the document and to send us comments!

The Working Group's response to your comment is included below.

Please review it carefully and let us know by email at www-smil@w3.org if
you agree with it or not before 20 Nov 2007. In case of disagreement, you
are requested to provide a specific solution for or a path to a consensus
with the Working Group. If such a consensus cannot be achieved, you will
be given the opportunity to raise a formal objection which will then be
reviewed by the Director during the transition of this document to the
next stage in the W3C Recommendation Track.

Thanks,

For the SYMM Working Group,
Thierry Michel
W3C Staff Contact

 1. http://www.w3.org/mid/200708101159.48572.Dr.O.Hoffmann@gmx.de
 2. http://www.w3.org/TR/2007/WD-SMIL3-20070713/


=====

Your comment on 17.4.4 Animation Module:
> Hello SMIL working group,
> 
> this does not look like an accurate method to round:
> 
> 'The mathematical definition of rounding is:
> coerced-integer-value = Math.floor( interpolated-value + 0.5 )'
> 
> -> There are different mathematical definitions for rounding.
> This is a biased rounding because number=integer+0.5 is always
> rounded up (for other cases the rounding definitions are the
> same). This is not an accurate rounding method.
> This means, over a larger number of rounding the average 
> rounding error is not zero.
> For an unbiased rounding, for example for even
> integers number is rounded down, for odd up...
> 
> example:
> value, top eq. 'rounding' , unbiased rounding (round to even)
> 0.5	1	0
> 1.5	2	2
> 2.5	3	2
> 3.5	4	4
> 4.5	5	4
> 5.5	6	6
> 6.5	7	6
> 7.5	8	8
> 8.5	9	8
> 9.5	10     10
> 
> average
> 
> 5	5.5	5
> 
> 
> I would like to suggest to use the unbiased or
> 'round to even' method ...


Working Group Resolution (LC-1801):
Originally the working group had relaxed the definition of rounding so
that it became implementation-dependent how ties were to be rounded.
However, the working group has decided to revert its decision.  Rounding
is now once again defined as floor(x+0.5).  The reason for this are:
- It is important for interoperability that rounding is well-defined (i.e.
that implementers know what to do);
- This method of rounding was already widely deployed in existing
players;
- There is indeed a bias, but the working group feels that the bias is
small;
- This method of rounding is very easy to implement on small devices,
whereas the method of rounding ties to even is much harder to implement.

The motto here is "practicality beats purity".

----
Received on Monday, 12 November 2007 11:34:52 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:34:28 UTC