[Jmol-developers] wrapping jmol as microsoft .NET J# class

[Miguel <mi...@jm...>]

> Just an update on this: the porting is turning out to be non-trivial.
> The =22upgrade wizard=22 actually produces C=23, not J=23, which is fin=
e with me
> since I know C=23 but not J=23.

OK

> Overall, I've started a very targeted port, just doing the classes I
> absolutely need, and the classes they absolutely need, etc, starting
> with org.jmol.viewer. I'm working on a NET3d class, parallel to Swing3d=

> and Awt3d, to do .NET specific rendering.

OK, that makes sense.

> A few problems and my current solutions:
> * The =22upgrade wizard=22 does not do a good job with the javax.vecmat=
h
> classes; it gets over-excited about converting get/set methods to
> =22properties=22, which have a slightly different api; I have to
> un-property-ify them by hand. To minimize work, I'm trying to port only=

> the classes that are actually used by Jmol.

What a disaster.

Per my other message, I expected that even a simple converter would get
that package right. The code is very simple, the structures is clean,
there is no IO.

No doubt .net has its own 3D path package. So one *could* consider
converting to that. But I suspect that the port would be a better
solution.

> * The =22upgrade wizard=22 just gives up on anything related to UI or
> ImageProducers/ImageConsumers. I'll be adapting a DoubleBufferedPanel
> colleagues at Brown wrote, to actually hold the rendered images, and
> using a built in Microsoft Image class. I really hope this gives good
> enough performance, but it will be a few days before I can tell for sur=
e.

within org.jmol.g3d you should not try to port Swing3d or Awt3d. Just
rewrite them.

All you need to do is:
 1. turn an int=5B=5D into an image
 2. render fonts into an int=5B=5D

> * For a first approximation, I'm not planning to port much of the user
> interface, since most of what we need is at the API-level. We've got
> very naive users and we don't need to give them all of the flexibility
> of real Jmol.

I do not think that you should port any of the UI of the Jmol application=
.

You should try to get the viewer running. Then put your own code around i=
t.

> I am concerned that I haven't been able to get it to be a turnkey
> process. Perhaps as I learn more about the process, I will be able to
> capture the post-=22upgrade wizard=22 steps in a perl script.
>
> Anything stick out as a bad idea or worrisome here?
> -sascha

Per my other message, I think that to get started you should simply work
on getting org.jmol.g3d to work.

Write some a small program that puts up a sphere and a cylinder.


Miguel