Re: [Jmol-developers] Initial .cube file isosurface support

[<rka...@ri...>]

On Jul 17, 2005, at 9:33 AM, Miguel wrote:

>> The reason one may want to consider deleting a surface may be if it
>> turns out that having, lets say 10 molecular orbitals with 2=20
>> polarities
>> for each orbital (i.e., 20 isosurfaces) on one atomset, which one=20
>> would
>> like to be able to turn on/off base on a javascript....
>
> I don't understand. 20 orbitals =3D=3D 20 surface -> 20 different =
names=20
> on/off

Indeed. My concern would be, which I forgot to add to that long=20
sentence ;-),  that so many isosurfaces could put a strain on the=20
amount of memory required and the JVM has available...

On a small aside: you seem to consider a molecular orbital with a=20
negative cutoff as a different orbital than one with a positive cutoff.=20=

Chemists don't see it that way since what really has physical=20
significance is the square of the orbitals, so the negative isosurface=20=

and the positive isosurface are considered at the same time. The reason=20=

to use signs is that the molecular orbital value can be negative, so=20
the mathematical construct does have that feature. The main reason, I=20
think, is that it is easier to see where nodal planes are, i.e, places=20=

where the MO value goes through zero, which also makes it easier to see=20=

what the symmetry of the orbital is.

>
>> By the way, the isosurfaces only work on a single atomset (or frame).
>> Would it be possible, down the line, to be able to set up multiple
>> frames with isosurfaces and be able to animate them to see molecular
>> orbitals been broken or made during a reaction?
>
> I suppose that it would be possible at some point ...
>
>> At the moment the isosurfaces are present for all atomsets, but I=20
>> think
>> it may make more sense to have one associated with a single atomset
>> (since I would assume that in most cases a different atomset would=20
>> have
>> atoms in different locations and as such the isosurface of another =
one
>> doesn't make sense).
>
> Q: Are you saying that a particular surface should be tied to a=20
> specific
> 'model' ?

Yes indeed. (I wish I could be as succinct as you are :-).

Ren=E9