BOOT pin for WMP100

My company has been shipping WMP100 based units for about 10 months.
From time to time,we have units returned that reset every time WIND:15 is
encountered.We can temporarily fix the problem by downloading OpenAT 7.3a again
with application,application data and objects erased.We then download our application
again and run our AT command configuration program.We shipped a unit to SierraWireless
that was exhibiting the WIND:15 problem.The SierraWireless support person suggested that there might
be something wrong with our BOOT pin implementation even though we followed the simple scheme in
the WMP100 hardware manual.One side of a single pole single pole switch is connected to 1.8V.The
other end of the switch is connected to the BOOT pin of the WMP100.When the switch is not closed,
the input to the BOOT pin is open.The support guy thought this to be bad even though that is the
specified scheme.Has anyone had trouble with this?
I should add that DWLWIN loads code ok on 95 percent of our units without the switch being closed.
The input to the BOOT pin probably,in 95 percent of the cases,acts like a logic one(1.8V) when it is open
thus making the BOOT pin see 1.8V.I think the boot switch is probably redundant.It would probably be
ok to tie BOOT to 1.8V.

We also ship units with the WMP100. We have the BOOT pin only coming to a test pad. When we program the OS onto the device we temporarily connect the BOOT pin to 1.8V, but for the rest of its life we leave it floating.

Same. No problems to dateā€¦

Same method, 10k programmed all fine bar for usual BGA smt issues.
Also consider getting the External EEPROM memory chip batch programmed by a programming house prior to mounting/smt if your code base is stable.