Things To Do to Make IPlot 1.0 Even Better!

Black and White Geometry Plots

This is because, although we can fill with any pattern, it is difficult to get a nice border around the entire region. Because the region is sometimes composed of many polygons (one for each boundary) even if we could get a nice border around each polygon, what we really need is a border around the whole region. It should be possible to extract the points that lie along the exterior of the region and draw lines along them for a nice border in black and white, but I didn't have time to do this.

Black and White Contour Plots

Much can be done to improve black and white contour plots. Mostly for these we need a bunch of bitmaps which will make dotted and dashed lines with good distinction between the configurations of dots and dashes. Symbol plots could be good too (diamonds, circles, squares, etc)

Contours

Contours are only made at present. They are never destroyed. I'm not sure that freeing up the memory would actually do any good anyway, but it would be more prudent. See contour.c where the delete part is commented out.

Dialogs

A lot of the dialogs need simple things to make them more user friendly. Pressing return in a text field often inserts unwanted characters instead of accepting the entry. The tab key can also be programmed to advance to other fields.

File Input

Currently, the functions to read in an HDF Vset file assume that there will both a Geometry and a Mesh structure within the data frame. It is furthered assumed that the Geometry will be the first structure and that the Mesh will be the second structure. For files that only have Mesh and no Geometry this poses a problem.

One issue is finding a way within VGeMesh to determine whether or not a given structure is a Geometry or a Mesh structure.

A more important issue, as far as IPlot is concerned is that some of the routines in IPlot depend upon there being a Geometry structure. One example is the routine FindMinMax which calculates the minimum and maximum extents of the data structure. There is no reason why this can't be done on Mesh data, except that it takes longer than doing it with Geometry data.

Multiple Data Sets

As the user adds a new data set, a global variable noof_data_sets (defined in globals.h ) is incremented and the new information occupies the next spot in the IPlotDataSets array. For convenience, the noof_data_sets variable is linked to the TCL variable ipl_noofdatasets. (More about Variable Linking) Ideally, what is needed is a way to delete data sets and then add new ones in the first available spot, instead of after the last spot.

Data sets aren't removable at present as this was not a priority feature. Ideally, what is needed is a way to delete data sets and then add new ones in the first available spot, instead of after the last spot.

Multiple Materials at a Single Point - Sorting Input Data

Right now, IPlot expects that when multiple materials meet at a single point in physical space, the data will contain a copy of that point for each material that occupies it. This is defined by the HDF Vset requirements. However IPlot further expects that these duplicate points will lie contiguously (all right after each other) in the data set.

The data needs to be sorted or else this requirement needs to be eliminated. The current reasons for this requirement follow.

• The point's need to hold their material type. The only way that they can get this information is from the element material type. When the data comes from FOREST, the elements will only indicate the point number of the first point in a set of points with the same material. In other words, no matter how many materials meet at a single point in space, all elements of all material types which contain that point, will list that point as the same number.

  This can be changed. If it is not changed, and we must assume that the duplicate points lie randomly throughout the data set, then we must search for points by their coordinates when assigning a material number. This will take much longer. Currently, we go directly to the first point in a set of duplicates and then search the following points by coordinates.

• Secondly, when we use the Node Plot functionality and click on a point, it is very easy to check the next point in the list for the same coordinates, as opposed to searching the entire list for points with the same coordinate. This is really the only reason why the points have to know what material they are to begin with.

Printing

See the section about printing (off the main technical documentation page). Printing is done rather as a hack and could be greatly improved.

Velocity Plots

It would be very cool to add another plot mode. I had an idea for a velocity plot which would plot velocity vectors at each node. (None should have velocities unless oxidation was done) There could be a different colored vector for each material - choosing the color automatically might be tough since you'd probably want to superimpose the vectors on the geometry plot - so you couldn't just use the region color because it would be the same. The vector would point in the direction of motion and be of a length proportional to the magnitude.

IPlot Home Page Technical Doc's User's Guide