Re: Ada in Medical Devices: was RE: Inexpensive Developer Level Commercial ADA compile

["John Cupak" <Jcupak744@mediaone.net>]

Time for my $0.02!

As a software professional, I have been teaching software engineering,
process,
language, and tool courses for a major defense company for the last 4 years.
Previously, I was a software developer and occasionally taught Ada. I have
over 30 years experience, and have a MS/CS. I'm not a novice.

I have seen the shift to Ada only because it was mandated by the customer,
then
a rapid departure to C++ and Java once the mandate disappeared. Languages
are NOT selected based on their reliability, etc, but (apparently) how much
it will
cost to hire / train programmers. "Penny wise, but pound foolish"
philosophy.

A few years ago, when I was looking to change my employer, I seriously
considered
employment with a well-known medical instruments company. After a bit of
research,
however, I was apalled at the lack of standards, reliability, and
programming language
knowledge! "Ada" was an acronym - NOT a programming language! Not even the
FDA had appropriate software standards for medical companies!

As a result, rather than be put in a position where my software engineering
and language
expertise would put the patient at risk, I choose to stay in the defense
industry. I doubt
the medical instruments industry has changed much in the last few years.

An interesting sidelight - in a Software Testing Overview course that I
teach, I present
the story of the Therac-25 (garnered from the internet) to impress upon the
attendees
how the medical industry ignores standard testing techniques, places the
patient at
risk, and "protects" the software developers! Inspection and testing
techniques which
are REQUIRED by CMM, ISO, etc., etc.

I'd love for the FDA to mandate Ada, but I doubt they're going to find
anyone to
support them. In fact, I'm SURE the medical instrument companies would lobby
against such a mandate! Too bad - for all of us.

John Cupak

P.S. My opinions stated above DO NOT reflect those of my current or any past
employers.


Michael Garrett <michaelgarrett@csi.com> wrote in message
news:9q8ii6$at1$1@suaar1aa.prod.compuserve.com...
> --From: Bob Leif
> --To: Michael Garrett, Ann Brandon et al.
> --Although to most of us the utility of Ada for medical devices is
obvious,
> --the FDA has not been educated to this fact.
> --As a former professor of Biomedical Engineering, I have always been
upset
> --about the lack of instruction in the regulatory aspect of the
profession.
> --This is in spite of about one fourth of the students becoming involved
in
> --satisfying the FDA, who is the first customer for the device.
>
> I would definitely agree about the lack of training on the regulatory
> aspects
> of medical device development. Not only on the part of students training
for
> careers in the field but also the training of the regulators
> themselves.....As you stated.
>
> I am currently working with a major university in the area on developing a
> bio computing
> curriculum. We have delivered the first in a series of three new classes
> that include
> lectures on the complete software lifecycle including validation
> verification procedures and documents,
> 510K filing and follow-up GMP inspections. We will be showing examples of
> documents
> and procedures that meet the regulations. The courses are currently
> "language independent"
> with students having exposure to assembly and C.
>
> The class involves a lab where students take a project through the entire
> lifecycle, including
> mock review of documents and functional testing.
>
> Now here is an interesting dilemma. There is not enough time in the
classes
> to teach a new language
> nor is the hardware available to run Ada. ( Current hardware is MOT
HC12 ).
> But I would like to convey
> the feeling that I had before and after learning Ada writing statements
such
> as:
>
> unsigned int selected_energy; /* range checking up to anyone who decides
to
> set the energy !!!!! */
>
> or
>
> void deliver_pulse_to_patient( unsigned int selected_energy )..........
>  you get the picture };
>
> My experience of such coding was drastically changed after learning Ada,
> even though I paid lip
> service to process  process etc. I might note that the experience resulted
> in process changes
> at the company that I was with.
>
> Your statement about one fourth of the students choosing a career path of
> "Satisfying the FDA"
> is an interesting one. I have found that in many companies the process of
> regulation has become
> one of satisfying regulators with a show of documents, rather than
> developing a process that
> continuously improves quality, safety and reliability. ( The razors edge
> ....... ).
>
> --I might also suggest, that if one has an installed, well tested code
base
> in
> --C, it is quite reasonable to reuse this code. However, this can be done
by
> --encapsulating the C into the bodies of Ada packages.
>
> Long Story.......
>
> Michael C. Garrett
> Garrett Technologies Inc.
> michaelgarrett@csi.com
>
>
>
>
>