Re: [office-comment] chart:axis-type="date"

[Regina Henschel <rb.henschel@t-online.de>]

Hi Andreas,

shouldn't this go to office@lists.oasis-open.org ?

An essential additional feature of a "date"-scale is, that it allows a 
major grid in month, so that each grid line can be on the 1st of the 
month, independent of the length of the month. Thereby still each month 
has its own amount of days.

Kind regards
Regina

Andreas J. Guelzow schrieb:
> Hi,
>
> I think I finally understand the problem chart:axis-type="date" is
> suppsed to address:
>
> Some charts, for example the chart:bar class, have an axis that is
> predefined to be either a categories axis (i.e. category labels are
> evenly spaced and the corresponding values are shown at these evenly
> spaced locations) or a value axis (i.e. values are shown at their
> relative locations).
>
> For example for a barchart with x-values that are dates (years ), e.g.
> 1950,1980, 2000, 2010, with a category type axis the 4 bars would be
> evenly spaced as if at values 1,2,3,4 with these points on the x axis
> labelled 1950,1980, 2000, 2010. If we had the same chart type but with a
> value type x-axis the bars would not be evenly spaced since the
> difference between 1950 and 1980 will be three times as large as the
> difference between 2000 and 2010.
>
> I believe Regina's suggestion is to obtain the second chart by labelling
> flagging the second axis as chart:axis-type="date".
>
> To me all of this has nothing to do with dates. These values 1950,1980,
> 2000, 2010 simply differ from other categories like a,b,3,d that they
> are all numbers and so can be interpreted as relative locations on the
> x-axis.
>
> So since there is nothing special about dates being involved (except
> that apparently Excel allows the second type of chart only for date
> values) I believe we should not make this date specific, but simply flag
> axis with their type (if it differs from their default type), e.g.
> chart:axis-type="values" or chart:axis-type="categories" or
> chart:axis-type="value-preference" where the last type means that if all
> values/categories given are numbers then the axis behaves like a value
> axes (placing items in their relative position) and like a categories
> axis otherwise.
>
> Andreas
>
>
> --
>
> Andreas J. Guelzow, PhD FTICA <mailto:andreas.guelzow@concordia.ab.ca>
>
> Registrar & Director of Enrolment Services
>
> Professor of Mathematical & Computing Sciences
>
> Direct: +1 780 479 9290
> Toll-Free: +1 866 479 5200
>
> Concordia University of Edmonton <http://concordia.ab.ca/>
>
> concordia.ab.ca <http://concordia.ab.ca/>
>
>
> Concordia University of Edmonton Twitter <http://twitter.com/CUEdmonton>
> Concordia University of Edmonton Instagram
> <http://instagram.com/cuedmonton/> Concordia University of Edmonton
> Facebook <http://facebook.com/CUEdmonton>