Henry Lenzi wrote at 08/04/2014 07:21 PM:
I was just wondering, Neil. what your experience in "production for
pharmaceutical prescriptions/labeling/instructions" software would be.
I'd be glad to read some stuff you've published or other software
solutions, if it's not proprietary software. If it is, what companies
have you worked for in the health sector?
There have been many decades of work in software engineering for
military/aerospace/datacommunications, which gives some background that
can also be applied to, say, medical informatics.
For one example familiar to me, I worked on software tools and
methodology that were used for things like (to give one specific, public
example) designing the cockpit system for a particular model of military
aircraft. Some of my colleagues from back then later went to work on
medical devices, which share some requirements for correctness and
robustness.
The fields of medicine have ongoing tremendous accomplishments. Many of
the roles, such as physicians and pharma scientists, require immense
training and discipline. Still, as medical fields get more into areas
of informatics and of process, my understanding is that they are
currently learning from more established fields. For one example: the
widely-publicized finding that hospital staff following a "pre-flight
checklist" each time when hooking up a patient lowered infection rates
greatly. That's many decades behind what people who work on aviation
process have further found about process.
I'm not sure how to connect this general to individual computer
programming practices, except to say that, in clusters of people who
have done such-and-such kind of production work, I think that there is
often general agreement about some of the best practices and pitfalls.
For an example bringing us back to this email thread, many people on
this email list come from different backgrounds of software development,
to be enthusiasts of an especially powerful dynamic language platform,
yet it seems that the majority(?) of us arrived at a wariness of "eval"
independently, even though I'm not aware of any textbook that says
"don't use eval" (textbooks often accidentally suggest the opposite).
BTW, "Don't use eval" is just one tip of many, and I didn't intend to
spend so much text on it in this thread.
Neil V.
____________________
Racket Users list:
http://lists.racket-lang.org/users
