> > I would love to have some help! There is much more work than I can > do. > > > > At a high level, my request would be that they: > > - take the job seriously > > - conform to some general coding standards regarding > > * indentation style and naming > > * efficiency > > * dynamic memory allocation > > - be open to constructive criticism and code review/peer review > > > > I will strive to do the same ... particularly the last point > > I am glad you feel that way about a possible co-operation Miguel. > Your requests (above) are the same as I would suggest if I were in your > place so I see no problem there. Good
> The [EMAIL PROTECTED] project is actually more of a science project than > anything else. I/we expect serious people working on it. As far as I can > tell at least some of our forum members are university students and > graduates. Due to the nature of the project I think others (read: less > educated) people will be discouraged if only by the problems we will be > facing of which only a minority will be on programming. > I can not tell you at the moment how many people we will be able to put > on helping with the 3D engine. That is fine. > We are only in the start-up phase at the moment. We are currently > working on setting up mailing lists for our teams. Only if this is done > we will accept developers. I am also working on a very crude todo list > for each team so that people get an idea about the work involved. > > In fact I am not 100% sure if we will be using Jmol at all. It is at the > moment a proposition made by me, and the initiator of the project > Robert Bradbury is inclined to go along with it. At the moment I am > only the (humble) acting co-ordinator of the project. It would be helpful for me to to understand what type of potential applications you are thinking of ... Are you thinking of Jmol for viewing molecules? Or, are you thinking of the using the graphics engine for rendering other types of objects? Whatever you are thinking of using, is it in the context of an applet or of a standalone application? > I would be grateful Miguel if you could give me more details about this > new 3D engine you are working on. It would help me defending our > possible co-operation to the other critical project members. You said 'defending' ... what else are you looking at? > Especial points of interest for us are: > 1. Speed! The, as ever, big Java question. Will the engine be able to > handle thousands even tens-of-thousands of atoms? Well ... it is not too bad. It is a function of: - CPU speed - quality of Java (old & interpreted or new & compiled) - OS platform ... painting an image on Win32 is faster than on X - window size ... double the number of pixels, double the rendering time All these times are using a window size of approx 300x300. On a 333 Mhz Mac OS9 running NS 4.78 (java interpreter) it is *very* slow. Maybe 10 frames per second with a small molecule in a 300x300 window. On a 700 Mhz Win32 machine running with compiled Java (MSFT or Sun JVM) you can get 20 frames per second with hemoglobin, 4500 atoms. I have a large file, a membrane section, with 40K atoms. On both a 700 Mhz Win32 box and 1 Ghz Celeron Linux box it takes 350 ms to render a frame. If the target machines are less than two years old then you won't have any problems. If they are more than 4 years old (as many seem to be in the academic/educational community), you may be disappointed. Other random thoughts: - Java is within 10-20% the speed of C code - *if* you are careful about how you allocate object and don't lean heavily on the garbage collector - Java systems calls are slower, especially the graphics calls - performance could be boosted significantly if I wanted to restrict to 256 colors ... but I don't want to do that - performance could be boosted significantly if Sun would fix some graphics performance bugs, but I am not optimistic about them doing anything > 2. How far is the development of the new 3D engine? The engine itself is basically done. Other features I am working on are the addition of new rendering shapes. Things like curved ribbons to represent an alpha helix in a protein molecule. The biggest issue holding up a release is some major changes that we made to the file IO system ... moving to the CDK libraries. > 3. A list of capabilities would be helpful, present and possible future > ones. I'm not exactly sure what you mean to encompass by 'capabilities', but ... - this is *not* a general-purpose OpenGL-style engine - it was designed specifically for rendering molecules - it does a very good job of rendering * spheres (atoms) * cylinders (bonds) * and lines (dotted and solid) * text (using the Java fonts) * triangles & quadrilaterals * spline curves - color shading of solid colors is quite good - it is a z-Buffer implementation, so shapes intersect properly. - there are no transparent colors (alpha channel) - there is no support for textures - currently no support for sprites - antialiasing is not yet fully functional (and is slow) - currently there is no mechanism for rendering tessellated polygon surfaces. This is planned. But it will certainly *not* have graphics-hardware-like performance - runs on old browser JVMs * NS 4.7 * IE - runs on newer Sun JVMs Not directly related to the engine, but perhaps of interest: - the applet runs as an untrusted applet * loads automatically without requiring user confirmation - the applet can be driven from JavaScript - it implements most of the RasMol/Chime scripting language for manipulating molecule rendering, especially .pdb protein files > I would expect that in the beginning the co-operation would be a full > one until the engine is complete and running bugfree. Well ... I would say that the core engine *is* complete and *is* running bugfree ... at least I hope it is bugfree :-) What remains is implementing polygon lists for implementing iso-surfaces (and I want to do some experimentation with 'spline surfaces' too) But it would be good to have someone else take a critical look at it, or to use it heavily. > After that we will > probably only co-operate on parts that are of mutual interest. I hope that much of this is of mutual interest :-) As I mentioned before, I think it depends upon the kinds of things you are looking to do with it. Check out these web pages with any web browser: http://jmol.sf.net/demo Miguel > > Val�re > > > > > > ------------------------------------------------------- > This SF.net email is sponsored by: SF.net Giveback Program. > SourceForge.net hosts over 70,000 Open Source Projects. > See the people who have HELPED US provide better services: > Click here: http://sourceforge.net/supporters.php > _______________________________________________ > Jmol-developers mailing list > [EMAIL PROTECTED] > https://lists.sourceforge.net/lists/listinfo/jmol-developers ------------------------------------------------------- This SF.net email is sponsored by: SF.net Giveback Program. SourceForge.net hosts over 70,000 Open Source Projects. See the people who have HELPED US provide better services: Click here: http://sourceforge.net/supporters.php _______________________________________________ Jmol-developers mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jmol-developers
