This isn't currently needed anywhere else, so the list of filenames
is just hard-coded into the script. Instead of making substitutions
in the source data, this just generates the full data.
Fairly straightfoward in Python.
Also, fixed the sidenav references, which should be using the file
in the sgtutorial subdirectory.
Also, tweaked index.html a bit.
Added a SourceGen tutorial with lots of screen shots. Uses
"responsive web design" so it works well on mobile devices.
This version is using jQuery load() calls to pull in pieces, but
that causes a lot of blink when loading because the loads are
asynchronous and may not complete until after the initial page
render has finished.
Tutorial prev/next links not yet working.
