OASIS Open Document Format for Office Applications (OpenDocument) TC

Dear TC members,

I am forwarding the following three proposals created by the
OpenOffice.org-Chart development team concerning chart axes
for inclusion into OpenDocument 1.2...

----------------------------------------------------------------------

Chart: Reverse Axis

Between Section 15.31.3 and 15.31.4 a new section should be added (or 
section 15.31.3 could be expanded):

Reverse Direction

The chart:reverse-direction attribute determines where the minimum and 
where the maximum values are displayed on an axis. The default direction 
of an axis is the one used in standard mathematics. 3D Cartesian 
coordinate systems are left handed per default and polar coordinate 
systems are counter clockwise. If x and y axes are swapped as in 
horizontal bar charts (when property chart:vertical equals true), then 
the unswapped coordinate system is standard mathematics and left handed.



----------------------------------------------------------------------

Chart: Right-angled Axes

Expand section 10.5.1:

Right-angled-axes

The attribute right-angled-axes can be used for 3D charts. If set to 
true it determines that two of the three axes in the cartesian 
coordiante system of the 3D scene must form a right angle in the 
projection on the screen. One axis projection is vertical and one is 
horizontal. Thus the chart is not drawn in a perspective look.



----------------------------------------------------------------------

Chart: show bars attached to different y axes side by side

The section 15.30.5 about Bar Chart Properties needs to be expanded:

group-bars-per-axis

With bar charts, the property chart:group-bars-per-axis can be used to 
display bars side by side even if they are attached to different y axes. 
The property can be used within the style at the plot-area element. If 
false, all bars on the same x axis are handled as one group and are 
displayed side by side. If true, only bars attached to the same x and y 
axis are handled as one group.



-- 
Lars Oppermann 

Lars,

Some comments on the first proposal: Chart: Reverse Axis

Lars Oppermann wrote:

> Dear TC members,
>
> I am forwarding the following three proposals created by the
> OpenOffice.org-Chart development team concerning chart axes
> for inclusion into OpenDocument 1.2...
>
> ----------------------------------------------------------------------
>
> Chart: Reverse Axis
>
> Between Section 15.31.3 and 15.31.4 a new section should be added (or 
> section 15.31.3 could be expanded):
>
Before getting to the proposal, a couple of comments on 15.31.3:

> If a scaling attribute is omitted, the axis is set to adaptation mode. 
> This means that the value is not set to a fixed value but may be 
> changed by the render application if data changes. However, the 
> chart:axis-logarithmic attribute is set to false.

Typo:

render -> rendering

Omit last sentence: However, the chart......

> The optional chart:axis-logarithmic attribute can be used to cause 
> logarithmic scaling on an axis. By default, proportional scaling is used.

Omit last sentence: By default, .....

Add: The default value of the chart:axis-logarithmic attribute is false. 
By default, an axis uses proportional scaling.

> Reverse Direction
>
> The chart:reverse-direction attribute determines where the minimum and 
> where the maximum values are displayed on an axis. The default 
> direction of an axis is the one used in standard mathematics. 3D 
> Cartesian coordinate systems are left handed per default and polar 
> coordinate systems are counter clockwise. If x and y axes are swapped 
> as in horizontal bar charts (when property chart:vertical equals 
> true), then the unswapped coordinate system is standard mathematics 
> and left handed.
>
I am not entirely clear on the end result that is the goal.

Cartesian coordinate systems, 3D or 2D are all "left handed" by default. 
That is to say that the zero point on each axis is located at a single 
point and most usually to the left of anyone viewing the representation.

It the intent here to be: (apologies for the poor representation):

0
|
y axis
|
|
|
____x axis_________0

?

I don't know that I would say that swapping the y and x axis is standard 
mathematics, although the statement that it results in "left handed"  
presentation.

To be honest, I found the last sentence unexpected and difficult.

If the axes are swapped:

|
x axis
|
|
|
0,0_______y axis______

so you are saying that:

|
y axis
|
|
|
0,0 _________x axis_____

Is unswapped and therefore left-handed?

Sorry, but the language leaves me wondering what the proper behavior 
would be in this case.

Hope you are having a great day!

Patrick



> 
>

-- 
Patrick Durusau
Patrick@Durusau.net
Chair, V1 - Text Processing: Office and Publishing Systems Interface
Co-Editor, ISO 13250, Topic Maps -- Reference Model
Member, Text Encoding Initiative Board of Directors, 2003-2005

Topic Maps: Human, not artificial, intelligence at work! 

Hi Patrick,

Remarks on the 15.31.3 are well taken...

Patrick Durusau wrote:
>> Reverse Direction
>> ...
> I am not entirely clear on the end result that is the goal.
> 
> Cartesian coordinate systems, 3D or 2D are all "left handed" by default. 
> That is to say that the zero point on each axis is located at a single 
> point and most usually to the left of anyone viewing the representation.
> 
> It the intent here to be: (apologies for the poor representation):
> 
> 0
> |
> y axis
> |
> |
> |
> ____x axis_________0
> 
> ?

With the help of the chart developers, I was able to get some 
clarifications...

For 3D Cartesian coordinate systems we need to indicate whether they 
should be left handed or right handed as both variants are widely used.
OpenOffice.org uses left handed systems for 3D charts per default:

y max
|
|         z max
|       /
|    /
| /
xyz min------x max

> To be honest, I found the last sentence unexpected and difficult.
> ...

Chart properties are applied in the following order:
By default the coordinate systems are constructed as described above. 
When the chart's chart:vertical style property is set to true, the x and 
y axes are swapped. After this, the chart:reverse-direction style 
property is applied to individual axes.

default:
ymax
|  zmax
|/___xmax

vertical=true:
xmax
|  zmax
|/____ymax

vertical=true AND y-axis specifies reverse-direction=true
xmax
|  zmax
|/___ymin


-- 
Lars Oppermann 

Lars,

Lars Oppermann wrote:

Patrick Durusau wrote:
>> Right-angled-axes
>>  ...
>>
> Does this have the same effect as the dr3d:projection attribute found in 
> 9.4.1 Scene?
> 
> Noting that 9.4.1 is cited by the current 10.5.1.

Here is some more explanation from the chart team...

It's different in effect but related. Having dr3d:projection set to
parallel makes 'parallel' lines. Setting chart:right-angled-axes to true
should produce right angles.
But in fact if setting right-angled-axes to true the values for
dr3d:projection would be ignored.
So one could think of expanding the values for dr3d:projection offering
an additional value 'right-angled' or similar.

dr3d:projection is parallel or perspective, the basic viewing modes for
3d projection (there are some more in the parallel tree, see Foley et
al).
  When You look at a 3d scene from the front and use non-rotated 
objects, the Lines of the shapes which are parallel to X and Y 
respectively will build a right angle in parallel and perspective 
projection mode. I think the 'right-angled' has nothing to do with the 
perspective projection mode, but with the orientations used in the scene 
geometry itself. This is no new mode for perspective projections. 
Remember: this is the mathematical projection between the camera 
coordinate system and the
2d view coordinate system.


-- 
Lars Oppermann 

Lars,

I appreciate the 3d's team attempt to explain what they meant but I am 
afaid I am still lost. The explanation seems to make sense but it is 
dangling just out of reach.

Can you provide the citation for Foley et. al that is mentioned below? 
Perhaps if I can review some of the literature I will have a better 
understanding of the explanation.

Hope you are having a great day!

Patrick

Lars Oppermann wrote:

> Patrick Durusau wrote:
>
>>> Right-angled-axes
>>>  ...
>>>
>> Does this have the same effect as the dr3d:projection attribute found 
>> in 9.4.1 Scene?
>>
>> Noting that 9.4.1 is cited by the current 10.5.1.
>
>
> Here is some more explanation from the chart team...
>
> It's different in effect but related. Having dr3d:projection set to
> parallel makes 'parallel' lines. Setting chart:right-angled-axes to true
> should produce right angles.
> But in fact if setting right-angled-axes to true the values for
> dr3d:projection would be ignored.
> So one could think of expanding the values for dr3d:projection offering
> an additional value 'right-angled' or similar.
>
> dr3d:projection is parallel or perspective, the basic viewing modes for
> 3d projection (there are some more in the parallel tree, see Foley et
> al).
>  When You look at a 3d scene from the front and use non-rotated 
> objects, the Lines of the shapes which are parallel to X and Y 
> respectively will build a right angle in parallel and perspective 
> projection mode. I think the 'right-angled' has nothing to do with the 
> perspective projection mode, but with the orientations used in the 
> scene geometry itself. This is no new mode for perspective 
> projections. Remember: this is the mathematical projection between the 
> camera coordinate system and the
> 2d view coordinate system.
>
>

-- 
Patrick Durusau
Patrick@Durusau.net
Chair, V1 - Text Processing: Office and Publishing Systems Interface
Co-Editor, ISO 13250, Topic Maps -- Reference Model
Member, Text Encoding Initiative Board of Directors, 2003-2005

Topic Maps: Human, not artificial, intelligence at work!