units update?

[Matthias]

Teo,

On Mar 27, 8:57 am, Teodoro Laino <teodor... at gmail.com> wrote:
> Matthias,
>
> I personally find the idea of the --codata2006 quite silly.  MAYBE --
> codata1998 instead!
> The newest the better (as usual)!

That was just an example.

> Few notices:
> Let me just point at your attention that we are not back-compatible
> (at the moment not even with cvs tree of 4 months ago)
> and hardly I think we will ever be (maybe only with a release version
> tree)..

Again, a mechanism appropriate for released versions that's what I was
referring to.

> Moreover what will be the impact of this change on the numbers?
> 10^-10? 10^-8? 10^-5? 10^-3? 1?[Hartree]
> Are these changes within the error of the methodology you're using or
> larger?
> (I hope they didn't discover in 8 years that the speed of light was 1
> order of magnitude wrong..)

That's not the point. In CP2K some care was taken from the very
beginning that mathematical and physical constants are really
"constant" and unique within the code and in my opinion that should
also be the case for different code versions. I would just like to
recall this.

> Please, note as well that there are few people spending  several
> hours in cleaning and getting rid of duplications and now we
> are planning for it.. Can you please explain me that?

Do you really want to discuss the issue of 20 more code lines for a
code consisting of 500k lines?

> So I would really like to know what is the "real reason" for having
> two sets of physical constants since, sorry,
> I cannot really understand your justification:

Alternatively, we could just accept one set of constants and ignore
any minor update of these constants in the future. I have no problem
with this policy.

Matthias

>
> > Numerically strict regtests and pushing the release
> > issue, but being careless concerning the employed constants does not
> > match very well in my opinion.
>
> > Matthias
>
> I'm not careless.. I'm realistic.. ;-)
>
> Teo