Shawn Smith
Smokebot
Lab 8

    The creation and implementation of the test suites was a lot easier than
anticipated.  I am glad to see that, because it does make it a lot easier to
accept that testing is a worthy thing to invest time in.  When I thought of
testing before, I thought it would be a lot more complicated than a series of 
"ok" statements.  Obviously, I didn't go too deep into the testing suite,
but it obviously is a quick and painless way to make sure critical issues
aren't broken.  

    I'm still not completed sold, however.  I mean, when I test to see if
command line input is bad, I don't see that being something that breaks later.
For example, if your program catches when you pass it zero and fails
gracefully, you can test for that once and there is little probability that it
will fail later.  You don't mess with the command line much.  But there are
other tests that it would be really handy for.  Just making sure math always
comes out right and stuff like that.  Overall, I'm still not 100% sold that
testing deserves a huge amount of time, but I agree that when it is
implemented, it sure is handy.  The only question is if you can afford the time
to build it.

   As for the group project itself, I can see testing working for certain
aspects.  For example, tests to make sure that the XML files generated by our
scrapers is always in the format accepted by our database populator is handy.
Also, it would be nice to test that the polylines we generate are always within
the viewport window.  This is an example of a test that takes time so should be
limited to testing occasionally, not every time it is used by the viewer.  We
could certainly make a smokebot for that!