NeXus workshop

[Mark Koennecke]

On Tue, 27 Jun 2000, Ray Osborn wrote:

> on 2000/06/27 4:44 AM, Mark Koennecke at Mark.Koennecke at psi.ch wrote:
> 
> > do we really need two workshops? Would enough people from America come
> > to europe to discuss such matters? Concerning SNS: can they not be
> > contended with the message that there will be a migration towards HDF5
> > within the next year?
> > 
> 
> The issue isn't whether we are going to migrate to HDF5.  I think SNS are
> already satisfied with the statements we've made, and the discussions you
> had with Tom at NOBUGS.  The main reason for wanting two workshops is the
> timing.  
> 
> There are a number of technical issues concerning how we start the migration
> to HDF5 + a number of other useful things to discuss such as Java support.
> The SNS wants to get going on both these issues soon - they might have some
> programming support next year.  Also, I know that you have money for a
> post-doc.  I think it would be useful for us to have a round-table
> discussion before that post-doc gets too far.  The transition to HDF5 is not
> just a rewrite of the HDF4 API because they handle some things differently,
> e.g. linking is done with absolute path names, not tag/ref pairs, there is
> no internal support for classes or global attributes.  We still have not
> resolved the issue of defining axes, or giving programmers high-level tools
> to extract plottable data.
> 
> These are all suitable subjects for just a few of us to grapple with, rather
> than involving the whole community.  However, we do need to involve
> scientists in refining the instrument definitions, which are still the
> weakest part of the project.  That will take time to organize - more like
> several months than several weeks, because I want to set up working groups
> in advance.
> 
> I repeat that I envisage the first as just an informal meeting of the NAPI
> team, the second as the first formal attempt to involve the whole neutron
> and x-ray community.  Let me know why you are not keen on two workshops.  Is
> it just the time they take up, or is funding an issue?  Concerning people
> coming from the US to Europe, we are going to seek funding support from the
> SNS so I hope that will not be such a big issue.  In any case, there are
> more functioning neutron sources in Europe at the moment  (I don't know
> about x-rays). 
> 
   I was afraid that more or less the same people would meet in a very
   short time intervall. Now I agree it is two different sets of people
   with some overlap. Also I was slow to understand that the move to 
   HDF-5 should bring a significant redesign of the NAPI. I looked into 
   this in some stage and got satisfied that it would be possible to
   reimplement NAPI in HDF-5 without changing the signature of the NAPI
   functions. That is why we have all these typedefs and defines in the 
   current NAPI. 


                             Mar