[NeXus-committee] NeXus base class discussions

[Mark Koennecke]

  High everybody,

  there are two things I'am getting concerned about in the current 
  base class definition process:
 
  - The first one is the voting procedure. Currently we get to vote on the
    complete class once only. I think this does not represent the issues
    coming up in the discussion phase good enough. I therefore suggest a two 
    step procedure:
    * In a first step Peter summarizes the controversial points of the
      discussion and the choices we have for the issues raised. We
      then vote on these choices.
    * In a second step, the DTD gets edited taking the results from the
      first step into account and then we get to finalize the definition
      in a vote.  
    A good example where this would be fruitful is the issue if the
    counts go into NXdetector or NXdata and whether we link or not.
    We have several proposals here. 
  - The second thing is a more general issue which covers many of the 
    base classes. I'am worried that the definitions become to complex.
    The discussion currently is dominated by people who have to describe
    very complex instruments. I absolutely agree that it must be possible 
    to describe such beasts in NeXus. And simulation results, too. But
    there are the simple instruments, as well. And I'am afraid that we create
    a major obstacle to the adoption of NeXus if we force people who
    deal with such instruments to use polar_angle, NXgeometry etc. instead
    of the well known two_theta and the like. I know this problem can be
    solved in software but I'am afraid that, given the manpower limitations 
    we all know too well, this software is never going to be written.   
    I think we should allow for an escalating complexity in the definitions.
    With the instrument editors deciding which level of complexity to
    choose. 

  What does everybody else thing about the points raised?

				       Best Regards,

					    Mark Koennecke