>         All indications in the errata point to this flaw
> easily being caught in
> "proper" testing, therefore testing which DIDN'T catch the
> flaw is obviously
> flawed itself.

They suggest 'statistical' testing with some 100 units for an
unspecified number of hours, presumably with an (equally unspecified)
perfect failure detection mechanism.

Now how does this apply to the project I just finished, where the
customer wanted 10 units, so I build 12 (one to get wrong, one for me to
keep for testing code changes, 10 for the customer). Am I required to
build 111 instead, build a perfect test rig, and test for a week or so
after each code change before releasing to my customer?

For a project like this (where development cost is much more important
than unit cost, at least initially - the customer might want 50 or 1000
units lateron) the 18F's are the PICs of my choice because they are much
easier to work with. But when unexpected trouble can pop up that offeset
all advantages.

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: www.voti.nl
consultancy, development, PICmicro products

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.