Re: [Cdk-devel] RFC #12 common API for CDK renderers

[Peter Murray-R. <pm...@ca...>]

At 12:15 26/02/2003 +0000, Egon Willighagen wrote:

>--------------------------------------------------------------------------
>  RFC CDK #12
>
>  Name: common API for CDK renderers
>  Proposed: 2003-02-26
>
>  PROPOSAL
>
>A API to be implemented by all CDK renderers and possible others, e.g.
>Jmol's 3D renderer. The API would define how the renderer can be instantiated
>and how a molecule can be passed to the viewer that should be displayes by
>the viewer.

<snip/>

I think this is very important. At present I use Jmol as my 3D renderer for 
JUMBO (after an unfortunate excursion into Java3D).

There are some fundamental questions:
- does the renderer support events? At present Jmol supports atom selection 
(does it support click?). I would like Jmol not only to render my chemical 
objects but to interact with the calling program or library.
- what is the minimal data model that has to be passed? 
molecules/atoms/bonds? Does the renderer modify these in any way (if so it 
becomes an editor). I would suggest it should NOT at this stage. However it 
could be valuable to return a transformation matrix (e.g. so the user could 
transform the molecule and apply this to their data)
- does the renderer have any chemical perception (e.g. join-the-dots, 
aromaticity?) I would argue not and that it should render exactly what is 
passed to it.
- What control should the user have over the rendering? should it all be 
interactive through a GUI or passed as an argument).
- should there be different configurations of the renderer? I would like a 
simple applet for embedding in HTML and a more complex application with 
pulldown menus etc. etc.
- should the renderer deal with non-molecular objects (boxes, ellipses, 
etc.) as well as molecules?
- if there are different renderers for the same data should they be able to 
access the same data structure. For example should 2 and 3 D representation 
of a single molecule be viewable simultaneously. What happens if the data 
in the model is changed?

I suspect that it may be useful to exclude some of these at this stage :-)

P.