TODO

SVG-import - FIXED!
There is a problem with importing svg-files created from dvisvgm. 
They are not automatically arranged in a suitable order. 

There are some possible workarounds:
	- Create your own importer of svg-files. This would be the 
	  safest way, but it also requires most work. - DONE!
	- Use dvisvgm but determine which letter is which by looking into
      the coordinates of the curve. 
	- Use pdf2svg: This seems to put the objects in the right order. 
	  However, this program does not outline the strokes (instead it
	  uses stroke thickness for e.g. the overbar of sqrt). So, in this
	  case, we should consider using inkscape to convert all objects to
      paths. This might not be future safe because the uncertainty of
      pdf2svg and inkscape functionality which may both change. 


Naming problems
- When adding an object with a name that is already taken, Blender adds .001, .002, etc to the object name. We must somehow catch this so that we can keep a reference to the correct object. 

Restructure: 
Currently the svg importer is responsible for creating the geometry on screen. 
It is better to keep everything in Python for as long as possible. 
This means that we should probably store the results of the SVG-importer is separate phobject (phovie objects).
For example, a LaTeX text object should include the following: 
	- A list of all the curves that need to be created. 

A phobject should be able to contain subphobjects. 
	- E.g. a LaTeX text can be divided into multiple subparts. 
	

LaTeX text
- In order to be able to handle parts of different LaTeX text separately, we must do the following: 
	- LaTeX the full object creating an SVG file which is then parsed into separate curves. 
	- LaTeX the first subpart of the text and count the number of curves it contains. Store this number of curves from the first object into subobject 1. 
	- Continue like this until all parts have been stored away.