|
|
|
##### **Generel things**
|
|
|
|
|
|
A line chart is a diagram representing a certain change between arbitrary units. Generally, it consists of two interval scaled axis. The important concept behind this diagram is to show a certain change in respect to units and in respect to (maybe) other datasets. It is not intended to show absolute measurements.
|
|
A line chart is a diagram representing a certain change between arbitrary units. Generally, it consists of two interval scaled axis. The important concept behind this diagram is to show a certain change in respect to units and in respect to (maybe) other datasets. It is not intended to show absolute measurements.
|
|
|
|
|
|
|
|
|
|
##### **Axisscalings**
|
|
##### **Axisscalings**
|
|
|
|
|
|
Due to the fact that a line diagram represents change, it should concentrate to show it without deliberate bias. Therefore, the minimum tickmark on each axis should be set according to the minima of the dataset (or on the global minima of all datasets that are to be represented). It is not useful to set the coordinate origin to (0|0), neither as to any other point other than the global minima. Another trivial concept which has to be followed is the fact that the maxima of the datasets should also be represented on the diagram.
|
|
Due to the fact that a line diagram represents change, it should concentrate to show it without deliberate bias. Therefore, the minimum tickmark on each axis should be set according to the minima of the dataset (or on the global minima of all datasets that are to be represented). It is not useful to set the coordinate origin to (0|0), neither as to any other point other than the global minima. Another trivial concept which has to be followed is the fact that the maxima of the datasets should also be represented on the diagram.
|
|
The *LineChartRasterizer* currently uses the min and the max of the datasets to set the ticks accordingly. The ouuput and the resolution highly depend on the choosen scalings, therefore other concepts should be applicable, as long as the do not falsify the representation.
|
|
The *LineChartRasterizer* currently uses the min and the max of the datasets to set the ticks accordingly. The output and the resolution highly depend on the choosen scalings, therefore other concepts should be applicable, as long as the do not falsify the representation.
|
|
|
|
|
|
##### **Axis**
|
|
##### **Axis**
|
|
|
|
|
|
There is not much to say about the axis itself, the *LineChartRasterizer* does not use an arrow head to show continuity of the datasets because there is no continuity. It ends with tha max value on the x axis which was present on the dataset. As for the tickmarksize, the *LineChartRasterizer* currently uses 2.
|
|
There is not much to say about the axis itself, the *LineChartRasterizer* does not use an arrow head to show continuity of the datasets because there is none. It ends with tha max value on the x axis which was present on the dataset. As for the tickmarksize, the *LineChartRasterizer* currently uses 2.
|
|
|
|
|
|
##### **Line rasterizing algorithm**
|
|
##### **Line rasterizing algorithm**
|
|
The *LineChartRasterizer* currently uses the **bresenham** algorithm to draw a line between two datapoints as nearly to the ideal line as possible. Because **bresenham** assumes an equal distant grid, there is a conversion to be done between the geometrical position (for example in milimetres) and the dot position, measured in braillecell/ dots. The current implementation looksup the nearest point, which is why there are sometime aritfact points.
|
|
The *LineChartRasterizer* currently uses the **bresenham** algorithm to draw a line between two datapoints as nearly to the ideal line as possible. Because **bresenham** assumes an equal distant grid, there is a conversion to be done between the geometrical position (for example in milimetres) and the dot position, measured in braillecell/ dots. The current implementation looksup the nearest point, which is why there are sometime aritfact points.
|
... | | ... | |