OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.

 


Help: OASIS Mailing Lists Help | MarkMail Help

office-comment message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]


Subject: Re: [office-comment] Re: improvements for interpolation in charts


Andreas,

Thanks!

Are there any lists you can think of where we could appeal for help with specific items like this one? Content would have to come in over the comment list but nothing says we can't appeal for help outside the TC.

Hope you are having a great day!

Patrick

On 11/10/2011 07:49 PM, Andreas J. Guelzow wrote:
Hi Patrick,

On Thu, 2011-11-10 at 16:54 -0700, Patrick Durusau wrote:
Will you have time to review the proposal Regina references in the
message with subject line: "polynomial regression curve in charts"?
The proposal referenced is quite incomplete and refers to an earlier
version of ODF than 1.2. It would be easy to adjust the proposal to ODF
1.2 but it is missing the specific calculation details. For example in
polynomial regression it needs to be more specifically stated in which
sense the "best" fit should be obtained. There are various ways this
could be defined leading to different regression curves. I don't think
we should include this in the standard unless we can be precise about
which regression curve is intended.

Since Gnumeric implements versions of this, I am obviously interested in
having it included in ODF (preferably ODF 1.3), but since this is not
just a question of reviewing the proposal I can't guarantee that I will
have sufficient time.

Sincerely

Andreas

Thanks!

Hope you are having a great week!

Patrick

On 11/10/2011 03:29 PM, Andreas J. Guelzow wrote:
Hi Regina,

On Thu, 2011-11-10 at 12:32 -0700, Regina Henschel wrote:
Hi all,

I have noticed, that you are discussing improvements for interpolations
in charts. Currently both kind of interpolations are parametric. That
gives nice curves, but for some functions the result can be better with
non-parametric methods. So I want to see non-parametric interpolation
methods in addition. For example my textbook "Engeln-Müllges, Gisela:
Numerik-Algorithmen" suggests the "Akima Subsplines" for not smooth
functions. It should be possible to use such interpolations too.
In fact they should be relatively easy to implement. I'll put them on
the list of interpolation methods to define (hopefully for 1.3).
If the already made suggestion "fixed derivatives at both ends" includes
"non-parametric" too, I would second that kind.

I do not find the term "parabolic limits" besides in the description of
the application "goffice". What does it mean? Is it the same as
"parabolically terminated"? And what does "cubic limits" mean?
[I'm no professional mathematician, but will implement those
interpolations. So please excuse me, if my questions are stupid.]
I really don't yet what specifically these methods are. I still have to
analyse the goffice code to see what they are calculating, determine
whether these are mathematically reasonable (and standard (in the sense
of "accepted in the field") and then write them down in a way that
non-mathematicians can implement them. I'll let you know when I have a
first draft.

Andreas

--
Patrick Durusau
patrick@durusau.net
Chair, V1 - US TAG to JTC 1/SC 34
Convener, JTC 1/SC 34/WG 3 (Topic Maps)
Editor, OpenDocument Format TC (OASIS), Project Editor ISO/IEC 26300
Co-Editor, ISO/IEC 13250-1, 13250-5 (Topic Maps)

Another Word For It (blog): http://tm.durusau.net
Homepage: http://www.durusau.net
Twitter: patrickDurusau




--
Patrick Durusau
patrick@durusau.net
Chair, V1 - US TAG to JTC 1/SC 34
Convener, JTC 1/SC 34/WG 3 (Topic Maps)
Editor, OpenDocument Format TC (OASIS), Project Editor ISO/IEC 26300
Co-Editor, ISO/IEC 13250-1, 13250-5 (Topic Maps)

Another Word For It (blog): http://tm.durusau.net
Homepage: http://www.durusau.net
Twitter: patrickDurusau



[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]