[Nexus] Correct way to specify multiple signals
V. Armando Solé
sole at esrf.fr
Thu Jan 11 15:17:35 GMT 2018
On 11/01/2018 16:01, Benjamin Watts wrote:
> Please be careful to fully consider the situation, Armando!
> You are asking for a default plot - Ray is saying that the description
> of that "default" plot shouldn't have arbitrary limits and that this
> might make it too complicated to be practical.
> In order to provide new options to specify default plots beyond what
> is currently provided by NeXus, we are discussing changes that have
> consequences beyond the use-case you are using as an example.
It has no consequences. Every program currently respecting current
NXdata definition keeps working.
> Therefore, your answer to Ray of "Nobody is asking for it" is not
> really true. Many of us who use NeXus will not be happy with a change
> to NeXus that solves your issue without having enough flexibility to
> also resolve other, similar issues and so the ability to "specify all
> other possible non-default plots" is something we *should* consider as
> part of the requested change as part of keeping NeXus functioning well
> in the future.
There I fully agree with Ray and disagree with you.
You will not be able to define all other non default plots for the very
simple reason that if you have saved 80 motors, 10 counters and 2
detector images you will not able to provide all the possible
combinations the user might want: plotting a counter versus two motors,
plotting a counters versus a motor and a counter, ... It is
responsibility of the programs reading the files to provide a system for
the users to choose. PyMca does that since 2010. However, if I would
have to foresee a way for the user to provide labels, uncertainties,
etc. via user actions I would end up with an unusable monster when NeXus
already provides an elegant solution.
> Please don't shut down such discussion with dismissive statements - we
> all want to make NeXus better.
Of course, but somehow I feel I am proposing things to do exactly and
what I find is not particularly encouraging.
More information about the NeXus
mailing list