Re: [h5md-user] static vs. dynamic H5MD elements

[Felix Höfling]

Am 01.07.2015, 16:14 Uhr, schrieb Pierre de Buyl
<address@hidden>:

> Hi,
>
> On Wed, Jul 01, 2015 at 10:48:21AM +0200, Felix Höfling wrote:
> > I have a suggestion which, I believe, would improve reading of the
> > specification. It does not affect any existing or future H5MD file.
> >
> > Currently, we distinguish between "time-independent" and  
> > "time-dependent"
> > H5MD elements. The difference is hard to grasp in quick reading and  
> > (even
> > worse) prone to typos. I suggest to replace
> >
> >     time-independent --> static
> >     time-dependent --> dynamic
> >
> > So we would speak of "static" and "dynamic" H5MD elements from now on.  
> > Are
> > there any drawbacks of such a change of wording?
>
> I am not convinced of the advantages (except the lesser number of  
> characters to
> type) but I see two inconvients:
> 1. work and blurring the history of the specification
> 2. possible confusion with physical quantities (the static structure  
> factor,
> etc).
>
> P
1a) the change is a simple search & replace, I can do this.

1b) With respect to the history, I believe that it is not difficult to
grasp the change from the context. For the majority of users (who will
hopefully join H5MD in the future), the history is not so relevant, but a
clear wording of the spec is.

2) This is intentional, actually the different naming was inspired by such
physical quantities. The static structure faktor S(k) could indeed be
stored as a static H5MD element (although one may define a time-dependent
version as for, e.g., the pressure—which is what I've actually been doing
in my work).

However, I'm open to other suggestions.

Felix