Heat does kill off batteries, that's why some cars now have heat
shielding or venting for airflow around the battery. But, that's only in
some cases and uses (one is Ford Crown Vics with the full police
interceptor package) where the underhood temperature is excessive to
start with.

I racked my brain for something solid to back this up, as I remember
seeing 150-200 deg F (somewhere in my past) as somewhat normal underhood
temperature, with "excessive" starting at 200. Unfortunately, my
Google-Fu isn't working well tonite, I couldn't find a website cite to
back that up.

But I did find this little gem:
(well worth saving on your hard drive, it's less than 6 mb)

http://www.batteryminders.com/batterycharger/press_room/starter-battery-frequent
ly-asked-questions.pdf

The author also states that maintainable batteries are preferred in hot
climates, but no definition of "hot" by a degree figure.

Same author cites a test done by Johnson Controls that found the average
life of a "good quality" battery is 37 months.

That same PDF has a section on why and how batteries fail, but it's a
little bit too technical in some ways and misses some basics.

Assuming that the charging systems is OK, most "just overnite" battery
failures are individual cell failures. When the battery gradually fails,
it's degradation of most or all of the cells. The PDF explains failure
conditions well but didn't really explain 'that' difference well.

A 12 volt battery is 6 individual 2 volt cells connected "head to tail"
(series) to total 12 volts. Each cell has to be in good working
condition for the whole battery to work. It's like those Xmas tree
lights that if one bulb burns out, the whole string's dead, or the chain
is only as strong as its weakest link.

"Mindfarts" do that. Been dere, dun dat. My (un)favorite is sitting in
the car, listening to the stereo with the key in ACC, then turning the
key to ON to open the power window to talk to someone, and forgetting to
turn the switch back to ACC. Damned DRLs (daytime running lights) do
drain the battery fast!

I have a cheap digital multimeter.  You almost make it sound like the
movie "Apollo 13" where they're trying to determine the least amount
of power they can get away with using to get back, using an identical
rig on the ground.  Being an engineer, the one thing I didn't
understand was whether or not they compensated for any temperature
differences that could lead to decreased current draw.

Again - I have the suspicion that it was just a weak battery that
finally gave up after 5-1/2 years.  My other thought was that it was a
dying starter (where a charged battery might mask the symptoms),
although that wouldn't explain the weak sound of the "finder"
indicator after the car had been sitting for a few hours.  It started
fine with a portable battery-powered jump starter (these things are
great) but after a while it gets inconvenient if you need it every
other start.  The electrolyte level was fine, but I'm not sure if that
matters much after 5+ years.

My 1990 Mustang blew an alternator 150 miles from home (at night), so
I had Ford put in a rebuilt alternator (they had no new ones - bad
idea, should have driven home in the morning without an alternator).

The new alternator had a short of some kind in one of the windings
(probably solder), so on random days my battery would be dead in the
morning.  I took it in to a local Ford dealership, and they yanked out
the perfectly good hatch light switch (there is no trunk), and charged
me $100 to not fix the problem.   A few weeks later when the short was
gone, I replaced the perfectly fine hatch switch.

Beware of things you do not expect, and things which seem to be the
problem but are not.

The fuse box Ammeter (DVM, etc) idea is GREAT, but be careful about
your assumptions.  The short may not be fused!  The alternator and the
voltage regulator for example are not fused.

After a few weeks the short burned itself out, without frying any
diodes, and all was well until that alternator blew a diode a year
later.  Replacing the factory High Output (!!!) 65A alt. with a brand
new Ford 100A alt. solved the diode blowing issue.

I suggest starting this procedure by disconnecting one battery lead,
and connecting the DVM across the gap.  Do this with EVERYTHING turned
off, and the meter set on DC Amps.  This will tell you the total draw,
which should be very small, but is probably close to 1 Amp if it kills
your battery overnight.  Then if you find a 30 mA draw on a fuse, you
know that is not the problem.  At that point I would ignore anything
below 100ma on any one fuse. (1000 mA = 1 Amp).  You must do this with
a fully charged battery.

Do not forget to reset your computer and FI system after screwing with
the battery connections.  I did once, and it trashed my injectors a
few days later with a backfire (that can't happen, I know - but it
did).