1 Due to our own lack of experience/hardware/software, we definitely need
2 volunteers (we'd be willing to help where we could of course) for this stuff:
4 o some sort of velvet-like interface for Windows
5 o all of the above kinds of things (particularly just binaries of the
6 command-line applications) for the Mac.
8 Things that maybe we'll get to someday, but if you want to try them first
9 then please feel free. In no particular order:
11 o make use of the Layout widget for geometry management of the main GUI
12 o more intelligent use of the command line ... command completion
13 and argument specification ... type node hit space,
14 get a "list of options" type delete hit space, get a prompt
15 for a number, type number hit return, etc. In other
16 words make it something along the lines of some of the
18 o improve plotting of displaced shapes. This probably means that
19 element writers need to provide a pointer to a function which
20 knows about the shape functions for an element and can fill in
21 a vector over the area or length of the element based on
22 nodal displacements which, when plotted, will reflect the true
23 displaced shape of the element ...
24 o add snap to tool capability
25 o robust quadrilateral mesh generation, 3-d mesh generation, etc, etc.
26 o additional higher-order elements -- particularly an LST and maybe
27 a quadratic 6-node axisymm element. The current linear axisymm
28 element could also be made more accurate given an actual
29 stiffness integration scheme. These last would all be fairly easy.
30 o fold Timoshenko beam theory into the standard beam, i.e., if
31 kappa or nu is given then expect a G and use Timoshenko
32 theory instead of Euler-Bernoulli.
33 o put some real effort into internationalization. The current
34 functionality is there because it was easy. More robust
35 implementation would mean maintaining a table, in code probably,
36 that mapped all possible input and as much output as possible
37 to a language determined by the user.
38 o plug in something more universal than PPM to replace GIF -
40 o fix the object apply code in velvet so that when an object
41 gets applied to a node or element its color changes as appropriate
42 (rather than waiting for a recolor request from the color
44 o an OpenGL 3D stress viewer
45 o consolidation of much of the isoparametric code: brick, CST, LST,
47 o uniform stresses, nodally averaged stresses as appropriate,
48 principle stresses, von Mises stress
49 o better, more uniform handling of distributed loads
50 o better handling of thermal analysis for the two elements that
51 we currently have, convection over the ctg surface for example,
52 rather than just the end