I've been thinking about a 70cm antenna made by using copper tape on a
window. It seems that I can get a 50 ohm match by drooping the two
radials about 38 degrees below horizontal. Of course, I don't know how
to include the effects of glass on the antenna nor rectangular coppper,
but it is a starting point. The azimuth pattern looks very good
according to EZNEC.

Any thoughts?

Yes, it's been done successfully... it's a basic ground-plane antenna.

There will be some amount of loading from the glass, which will
probably require you to shorten the elements slightly to achieve
resonance where you want it. The use of wide conductors will have a
similar effect - EZNEC may be able to model this. If you have an
antenna analyzer you could just cut-and-try by making a prototype with
the copper strips attached to a piece of cardboard, trim until you get
a match you like, and then cut a second set of copper strips and
attach to the glass.

You may want to add a ferrite bead-balun to the coax just below the
feedpoint. Not strictly necessary but it can help keep RF off of the
outside of the coax (from which it would radiate on the indoor side,
probably squinting some power in useless directions).

One caution, though: before you start this, consider whether your
window uses a "low-E" energy-saving glass. This sort of glass has a
thin film of vacuum-deposited metal on one side, and it's likely to be
conductive and lossy enough to make your antenna not work at all well.