Re: [Phys-L] Supercooled Water , A Seven Year Physics Tiff

[brian whatcott <betwys1@sbcglobal.net>]

For high-reliability code (as in avionics) it is not unknown for two 
coding teams to each execute a version, preferably using a different 
language. Comparison of outputs provides a pointer to coding errors. 
Here, starting from the same app, two teams produced discrepant 
results.    As often happens in academic contexts, the code looked 
somewhat ragged and needed beautification before being let out for code 
comparisons, and at this point, the paradigm failed with physicist-style 
hurt feelings and resource restraints. So it took seven years to bring 
the comparison to bear, with a by no means obvious problem with a 
conceptual model code stream-lining for speed.

Reminds me of a development moment in a  point of sale terminal 
development for a major convenience store chain, with which I was 
involved as systems manager. A prototype section I wrote just would not 
play nicely with an interacting section carefully written and tested  by 
the joint design chief. We were using a newish language-vehicle at that 
time; DEC Pascal on a PDP1132 development system. The Yale trained 
developer decided he needed to recode my section, because he had 
carefully checked each paragraph of his section in debug mode for 
correct results (I had not).
When he reran a test sequence with his fresh code, it still failed in 
the same way, and the realization finally dawned that it was DECs debug 
code that had a bug!   :-)
Sadly, when it came to the 100 terminal demonstration test for the 
customer, it came out that the market for online entertainment ticket 
sales; theaters, shows, sports events, et al had been all but cornered 
by a leading ticket sales organization called Ticketmaster - so that the 
whole engineering effort, comparable to airline ticket sales systems 
(which are notoriously fraught)  - though successful in engineering 
terms, was a marketing failure.

Brian W


On 8/23/2018 1:44 AM, bernard cleyet wrote:
> Would neutron diffraction also be suitable?
>
>
> bc
>
> > On 2018/Aug/22, at 22:38, brian whatcott <betwys1@sbcglobal.net> wrote:
> >
> > Oh dear!
> >
> > <https://physicstoday.scitation.org/do/10.1063/PT.6.1.20180822a/full/>
> >
> > Brian w