Menu
▾
▴
Re: [Jmol-developers] Question about coordinates in Jmol
|
From: Jonathan G. <gu...@uw...> - 2011-06-30 13:50:20
|
I try to answer your questions below. In summary, now that you've changed the isosurface slab command so that it can be done after the surface is defined I am mostly concerned with building a GUI that is intuitive enough that people don't have to understand the details of the isosurface and slab commands to look at a section through the displayed object. My first goal is to get this working well for isosurfaces, but I don't think it will be hard to extend it to do everything.
On Jun 30, 2011, at 1:07 AM, Robert Hanson wrote:
> well, the molecular coordinates are just the molecular coordinates. This "slicing" is by them, right? Not the screen coordinates? There are three coordinate systems:
>
> -- molecular coordinates
> -- relation of those by rotation to the original view (the orientation quaternion/matrix)
> -- translation of that into screen coordinates (the transformation matrix)
>
> I'd sure like to hear more about what the specs are going to be on this "slice" command. Before you go too far, Jonathan, can you remind us of what exactly it is going to do?
>
> Q: Are we slicing in molecular coordinates (like isosurface slab plane... and slab plane...) or screen coordinates (like slab/depth and isosurface slab nnn)?
The first draft will be slicing in molecular coordinates, although I think some people might also like a GUI for slab/depth like behavior.
>
> Q: Does the slab change when you rotate the model, or is it fixed? (Basically the same as the previous question.)
People wanted to be able to look at the slice from different angles so that's why molecular coordinates.
>
> Q: What functionality does this add that does not already exist in the other commands?
Other than a ghost feature, which presently requires building a duplicate surface, I cannot think of anything.
Anyway, the issue is that people want the ability to specify the slice in a way that is logical relative to the view. Tentatively, I am going to make the controls relative to the origin of the boundingbox. If they are relative to the view orientation, things change on rotation and translation. I think that would be confusing. I think there should also be an option to do everything relative to the absolute coordinates (molecular). I'll just set a switch. So I am still at my original question, which is will the boundbox coordinates always be parallel to the molecular coordinates?
>
> There is no point in adding a new command unless it makes *scripting* far easier. This sounds a lot more like a GUI thing -- which would use the already-existent commands, possibly with a few modifications to suit.
I agree. Initially as I began looking at what Jmol could do, I thought it would help. Now I'm not sure.
>
> It seems to me a popup window is a nice idea -- at least for the application. Not convinced with respect to the applet -- we don't do that, so far, except with the signed applet for file read/write. A good dialog can take lots and lots of code to produce and work with. If possible, with the applet, we should do that in JavaScript, not Java. (My opinion.)
I agree for another usability reason as well. When you have more than one applet on a page it is difficult to tell which one the pop-ups go with. A javascript tool that is anchored on the page in the vicinity of the applet is much easier to understand. That is one of the reasons, I thought about a special command. The idea being that the smaller the scripts that have to be passed the better the potential for speed.
>
> By the way, you asked about the possibility of duplicating an isosurface so there was a translucent "shadow." Since rendering is a two-pass system, a better way
I like your idea and have sent a comment in a separate message.
Jonathan
Dr. Jonathan H. Gutow
Chemistry Department gu...@uw...
UW-Oshkosh Office: 920-424-1326
800 Algoma Boulevard FAX:920-424-2042
Oshkosh, WI 54901
http://www.uwosh.edu/facstaff/gutow
|