From: Marc Baaden (baaden_at_smplinux.de)
Date: Thu Nov 13 2003 - 10:11:44 CST

>>> John Stone said:
>> with the technique in general. The main problem with doing it for VMD is
>> in regards to trajectory data, (both preexisting and real-time data coming
>> from IMD sims) where every frame needs its own grid since the atom position
     s
>> move. [..]

But in any case, when you press the button, the nearest atom needs to be
determined. As a simple workaround/option, one could just use that existing
code to highlight the nearest atom. This will certainly work for small systems
(couple of thousands of atoms ?), and we'd eventually need to find something
more efficient for bigger systems. But often the interesting part is quite
limited, eg say you look at a protein trajectory .. most of the time you are
interested in the protein only, so one could restrain the selection to just
that part, which should make it more efficient, as usually the number of
protein atoms is only a small fraction of the total system (water/bilayer/..)

>> Anyway, the acceleration data structure is still only one issue, the other
>> is making a good GUI for the feature, deciding how the "highlighted" atoms
>> work, i.e. does the user have any control over HOW they are highlighted, et
     c.
>> The first implementation of this code was essentially hard-coded, and behav
     ed
>> somewhat similarly to the way the sequence viewer extension works in VMD,
>> highlighting the nearby atoms in yellow, etc. If I recall correctly, there
>> was actually no GUI control, no way to turn it on or off at all.

I was indeed thinking of the sequence viewer. But would it be very difficult to
let the user choose the representation ? I could just imagine an additional one
in the graphical representations, eg. defining VDW for a selection of something
like "CurrentSelAtom" or so which dynamically links to the nearest selecetd
atom .. ?

>> Here's a fun little question for the VMD user community that relates
>> to this topic:
>> Would you like to see a "palette" of atom selections in VMD?

Sounds like a great idea to me.

>> representations window. Does this sound advantageous? Do people make
>> significant use of the same atom selections over and over, and would this
>> make life easier?

Talking for my case, it's less the atom selections than using/combining
different graphical representations (for different selections also). Eg
switching on the active site, switching off the lipids, etc. With a palette
using an on/off toggle for each defined graphical representation this might
be simpler.

>> (VMD already has atom selection macros that you can use to define things
>> that you use really-long-term as a new atom selection word, but this isn
     't
>> quite the same thing that I'm asking about above..)

The atom selection macros are IMHO very useful and I make extensive use of
them. Eg for things that don't change for a given system, I usually define
them as atom selection macros (loops, active sites, ..)

Cheers,
Marc

-- 
 Dr. Marc Baaden  - Institut de Biologie Physico-Chimique, Paris
 mailto:baaden_at_smplinux.de      -      http://www.marc-baaden.de
 FAX: +49 697912 39550  -  Tel: +33 15841 5176 ou +33 609 843217