Final Comments

Last modified September, 1996 - Armel Le Bail

Content

• Advices
• Improving STRUVIR
• Bugs

Have Fun and Good Luck

Advices

• At least, respect the limits : 1000 polyhedra, 6000 atoms, 20 colors, 150 atoms around one atom in distance calculation... Non respect may lead to crash in some cases although the program should stop by itself with an explanation for the user.
• Care to put the end of the .dat file not on the same line as the final atom in the list because this atom will be ignored. The end of file (EOF) should be at the beginning of an empty line after the last atom.
• In option CIRC, prefer small sizes like CIRC 0.03 CS instead of CIRC 1.30 CS (the struplo.hgl file will give Cs atoms as points but the .wrl files will be OK, otherwise nothing may work because STRUPLO has problems -apparently- with superposed circles...).
• Due to a bug, the HPGL or POSTSCRIPT drawings may look uncomplete. This occurs generally when the 2D projection is made along a crude axis (by View 1 0 0 or 0 1 0 or 0 0 1, or by ROTA 90 0 0 and so on). A possible way to avoid this is to replace too direct projection axes by neighbouring ones : for instance replace ROTA 90 0 0 by 85 5 0.
• Sometimes polyhedra may appear uncomplete in the VRML output, when the range selected by the XYZR option is too small. Enlarging the box size should overcome the problem. However, sometimes you really want a small part of the structure. There is a simple way to deal with this bug in VRML. The .wrl ASCII files may be edited and the uncomplete polyhedron may be easily recognized and deleted. Data concerning each polyhedron in po.wrl begin by a comment line like the following :

# POLYHEDRON 46 YB
    Material{
      diffuseColor .000 1.000 1.000 
      specularColor 0.5 0.5 0.5 
      shininess 0.2 } 
    Coordinate3{
      point[ -.101 5.419 2.710, 
             1.762 4.065 4.133,
.......etc

         2, 6,-1, 1, 7,-1, 4, 7,-1 ]}
# POLYHEDRON 228 YB 
    Material{
      diffuseColor .000 1.000 1.000
.....etc

# POLYHEDRON 39 YB
......etc 

So that you have just to delete the undesired polyhedron. Numbers 46, 228, 39 etc are those easily found in the atom list in the output file struvir.imp

Improving STRUVIR

Implementation of more efficient capabilities will be done only if there is some interest from users (and if I am able to do the job!). The Windows 95 executable was created by MSVC++2.00 from the C source obtained by applying f2c to the FORTRAN source (using libi77.lib, libf77.lib and f2c.h).

STRUVIR has a long history as a public domain software, mainly under the STRUPLO name. Many people have contributed to improvements. Sources are still available for free. So why not to improve the program by yourself. If so, please continue to propose your version as a freeware with sources (you may not like colors I have selected, for instance...).

Interesting changes could be :

• -automatic recognition of other coordinations >12
• -different sizes for spheres as it is possible for circles
• -ball and stick and other rendering
• -optimization of VRML 1 enabling more viewers to show something
• -introducing VRML 2
• -allowing labelling
• -making a better manual in HTML
• -.....

Bugs

Possibly there could be different kinds of problems. The .dat files delivered with STRUVIR.ZIP have been tested and visualized successfully with the VRweb and Live3D VRML viewers. This is unsufficient for claiming bug absence. Particularly, the coherency of shading according to light sources has not been checked : the drawing is sometimes 'bizarre' as stated by some pre-beta testers. Viewers like Cosmo Player from SGI and the Microsoft VRML Add-in will not give good results essentially because the STRUVIR VRML output is not well optimized. Only those viewers enabling two-sided face lighting in VRML 1 should be used. Also, you may find that 'flat shading' renders better the VRML STRUVIR files than 'smooth shading'.

Bugs correction :

• March, 01 1996 : was not possible to obtain more spheres than atoms involved in polyhedra.
• March, 27 1996 : Live3D under Windows 95 loaded the .wrl files then Netscape 2.0 crashed ! Nothing such happened under Windows 3.11. A ",-1" was lacking somewhere.
• July, 11 1996 : A. J. Holland during his work for including executables of STRUVIR for various platforms in the CCP14 suite found that some data files produced crash on DEC ULTRIX, ALPHAVMS an VAXVMS only. The 2002 FORMAT in the subroutine WRITE produced sometimes lines exceeding 132 characters.
• July, 26 1996 : added one line in the documentation. ALL CHARACTERS IN UPPERCASE Care to respect exactly the way the commands are defined in the documentation. Thanks to Mark Koennecke.

HOME