[ale] sloppy coding in breath tester

[tom]

On Fri, 15 May 2009, Jim Kinney wrote:

> On Fri, May 15, 2009 at 9:16 AM, tom <tfreeman at intel.digichem.net> wrote:
>
>> I only scanned the link, so ...
>>
>> There is another possiblity - the authors were never CS trained at all.
>> I'd guess that many of the people writing code for the breathalyzer were
>> originally engineers or scientists with a dominant background in
>> instrumentation. Coding would have been pickup activity which started
>> taking over their lives like kudzu. Hence sloppy code due to a lack of
>> background.
>>
>
> I disagree. I would suspect the bad code to be from a coder with lousy
> math/science skills not a science person with lousy code skills.
>
> There is a big difference between teh average of an array of values
> and the average of an array of values plus a new value.
>
> But without knowing details of who the bonehead was who didn't do due
> diligence verification this all just blowing smoke.

Well, we definitely agree about the blowing smoke part here.

So, as you expect, here is some more smoke...

While not directly coding, I have seen some PhD chemists do absolutely 
horrible things to a simple algebraic expression in support of their 
notions. One of those was also a major "supporter" of using statistical 
methods - like performing a regression on three variables which had the 
constraint a + b + c = 100. Einstein wondered why some of us laughed 
behind his back.

Basic ignorance isn't limited to any field of endevor. Which is sort of 
what I was trying to imply earlier on