Re: compatibility of plotting with sparse matrices

[John W. Eaton]

On 13-Dec-2012, Ben Abbott wrote:

| I don't think it necessary to copy the behavior Matlab here (maybe
| only for --brain-dead).

I think it would be bad to have this "feature" depend on that option.

| Fixing the gnuplot backend should be fairly
| easy, by converting the properties to full matrices when needed.

I don't know.  It seems to me that the only reason to preserve
sparsity is to save memory.  So if you are just going to convert to
full then you might as well do it when storing the property values.

The case that brought this to my attention was a report about things
like

  loglog (sprand (200, 1, 0.1), 'o')

failing with gnuplot.  It works with opengl because there, the log of
the sparse array is computed which produces a sparse array with many
-Inf values.  Then those are ignored when creating the actual plot.
It seems like it would be much better to avoid computing and storing a
lot of useless values.  But that will be a lot of work, and I'm not
sure that the optimization is worth it.

I think I'll just fix things so that the stored data values are full.
That's a nice simple fix.  I don't think it is worth the effort to
worry about the efficiency now.

jwe