regression tests, not GDP

[Graham Percival]

As Trevor said, I'm trying to keep discussions about GDP focused.  So 
I'd like to close off a few branches of discussion:

1)  Who has a copy of web/master in git laying around?  Please make a 
webpage that links to the regression checks and the comparisons.  Link 
to it from... well, wherever you want.  Probably either devel/ or 
devel/participating.

Advanced users will be able to find it, the programmers and bug people 
will have it bookmarked, and new users looking at the main doc page 
won't get confused.  Everybody wins.


2)  the Program Reference is *entirely* outside the scope of GDP.  You 
know the phrase "... where angels fear to tread"?  That's how I feel 
about those docs.

Can we think of a billion ways to improve them?  Yes.  Do I have a clue 
how to implement any of them?  No (other than really simple things).  Is 
it helpful to list pie-in-the-sky features that likely will never be 
implemented?  ... I'd say generally no, but hey, if you enjoy doing 
that, I won't stop you.

If you _personally_ know how to modify the program reference (it's 
constructed via scheme and/or python and/or C++ and/or prayers and/or 
angel dust), and are willing to spend the time to make such changes, 
please speak up.  Otherwise, be aware that your discussions about the 
Program Reference will be as fruitful as planning a manned mission to Mars.


Again, I'm sorry to be a wet blanket.  I'm just doing this to save 
everybody's time and effort, since I honestly don't think this 
discussion will amount to anything.
- Graham