[NeXus-tech] missed the telecon

Tue Sep 17 10:16:35 BST 2013

Hi,

just some more stuff from me.

On 09/17/2013 11:04 AM, Tobias.Richter at diamond.ac.uk wrote:
> I'll just throw in some information to prepare for the teleconference.
>
> * I do not fully understand the changes for #266 yet and may have some comments when I do, but this looks like good progress. Thanks Pete!
>
> * The 2014 NIAC meeting would traditionally be associated with the 2014 NOBUGS meeting, at J-PARC/KEK next year. I did ask Jiro about dates and whether he was willing to host NeXus meetings as well. He confirmed that they have formed a local organising committee.
This is my first guess as well. But are our american colleagues able to 
travel to Japan? If not, would we be able to meet them
someplace in america?
> * I agree with Pete that geometry (mostly documentation thereof) and multidimensional arrays are a problem. And I am also not sure how the inheritance helps in practical terms.
IMHO:
- The interface stuff I sent out is meant to help document and organise 
our base classes better.
- The geometry stuff I consider solved with the adoption of the CIF 
approach. If I am wrong, please
   give an example of what is not covered or coverable.
- Multidimensional arrays stay a problem. Where the biggest problem 
seems to me that we want different
   axes for different use cases of a data array. Use cases can be scan 
intent, the actual scan, after some conversion to
   Q(E) or whatever and more.  CIF has axissets (a collection of axes) 
for this purpose. I wonder if we can use this idea.
   I am not sure yet how to encode this in NeXus though.

Regards,

                 Mark
> ** for the geometry we agreed on the CIF geometry in 2010. We are using that for the CIF->NeXus converter now (obviously) and I have beamlines writing that directly. We really need to get the information out there so other facilities can make use of that. I will send out some some details about changes/additions required (to the standard or documentation) due to the converter shortly.
>
> ** For the multidim scans an NXscan base class would help (not sure what the NXscan application definition does for us). It could capture
> *** rank + dimensions of scan
> *** devices actively driven
> *** timestamps arriving at each point
> *** demand values?
> That would help to assign dimensions of many datasets that need a per point value.
>
> Tobias
>
> ________________________________________
> From: nexus-tech-bounces at nexusformat.org [nexus-tech-bounces at nexusformat.org] on behalf of Pete Jemian [prjemian at gmail.com]
> Sent: 16 September 2013 19:29
> To: nexus-tech at nexusformat.org
> Subject: Re: [NeXus-tech] missed the telecon
>
> Can't attend tomorrow's telecon so here is my 0.00000002eu.
>
>
> About the different rules for app defs and base classes, see TRAC #266
> (http://trac.nexusformat.org/definitions/ticket/266) which is now ready
> on a branch for examination.
>
> Will the validation procedure need to be modified to account for the new
> files: applications/nxdl.xsd, base_classes/nxdl.xsd, and
> contributed/nxdl.xsd?
>
> The cmake packaging needs to be modified to include the new files listed
> above.
>
> I agree that a 2014 NeXus Code Camp would be in order to make
> substantial progress on an implementation.  Where and when will that
> code camp be held?
>
>
> About the NeXus interfaces, what Mark suggested should be implemented
> within the nxdl.xsd XML Schema rules for NXDL files.  I'm sure there is
> a way to document and describe the rules.  We will probably want a
> shorthand, as Mark has shown, to make it easier to discuss.  It's not
> yet obvious how this improves the situation in a practical sense with
> verbose declarations of nomenclature in the base classes.
>
> Seems to me the biggest issues (that cannot be solved by interfaces) to
> improve NeXus are with how NeXus describes geometry and with the
> multi-dimensional arrays.
>
> Thanks,
>     Pete
>
>
> On 9/4/2013 10:05 AM, Pete Jemian wrote:
>> I, too, missed today's telecon and will miss the next since I have a
>> standing conflict at that time on Tuesday.  Keep me informed as I would
>> like to attend again.
>>
>> I see that we have matters to resolve with a difference in the rules for
>> application definitions and base classes.  Since we are considering a
>> change in our VCS (version control system) infrastructure, I will create
>> a new ticket for those changes and examine offline.  Once the VCS path
>> is clear, we can propose the new rules and commit.
>>
>> Pete
>>
>> On 9/4/2013 9:41 AM, Wintersberger, Eugen wrote:
>>> Thanks for the update
>>>
>>> On Wed, 2013-09-04 at 16:37 +0200, Mark Koennecke wrote:
>>>> Eugen,
>>>>
>>>> On 09/04/2013 04:31 PM, Wintersberger, Eugen wrote:
>>>>> Hi folks
>>>>>     unfortunately I missed the teleconference today. I was too busy
>>>>> with a
>>>>> couple of other things and obviously missed the time.
>>>>>
>>>>> Was there any discussion about the GIT migration ;)?
>>>> there was. Git was seen rather favourably. But no one is suffering from
>>>> svn so much as to invest
>>>> to much time... Stuart is kindly being asked to give the conversion from
>>>> svn/trac to github a try.
>>>> You could also try; but talk to Stuart then. Further decisions will be
>>>> made based upon  the
>>>> results of the trial run.
>>>>
>>>> Regards,
>>>>
>>>>                  Mark
>>>>
>>>>> regards
>>>>>      Eugen
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> NeXus-tech mailing list
>>>>> NeXus-tech at nexusformat.org
>>>>> http://lists.nexusformat.org/mailman/listinfo/nexus-tech
>>>> _______________________________________________
>>>> NeXus-tech mailing list
>>>> NeXus-tech at nexusformat.org
>>>> http://lists.nexusformat.org/mailman/listinfo/nexus-tech
>>>
>>>
>>> _______________________________________________
>>> NeXus-tech mailing list
>>> NeXus-tech at nexusformat.org
>>> http://lists.nexusformat.org/mailman/listinfo/nexus-tech
>>>
> _______________________________________________
> NeXus-tech mailing list
> NeXus-tech at nexusformat.org
> http://lists.nexusformat.org/mailman/listinfo/nexus-tech
>