[Nexus] Basic questions about application definitions

[Andy Gotz]

Dear Ray,

I was expecting someone to ask this question i.e. I asked myself the 
same question ;-)  I understand someone has to start using the 
application definitions otherwise they will never be used. Do the 
authors of the definitions use them? That would be the minimum I would 
expect if someone proposes a definition. It would be useful to have the 
name of the principal author and which programs support them as part of 
the definition.

On our side we will see what we can do to make them work for us. We will 
send you our feedback on our experience. We encourage other sites to do 
the same. Especially those who are using Nexus extensively or have 
proposed application definitions.

Kind regards

Andy

On 17/02/2015 20:11, Osborn, Raymond wrote:
> Hi And
> I’m obviously not in a position to understand all the issues and resources that you have to juggle with, but I would encourage you to use the application definitions as they now exist, or propose modifications if they are not adequate to your needs. The only way that the application definitions will become useful is if major facilities like ESRF actually start using them. I can’t guarantee that they won’t need modification once we have a lot more real-world examples, but if we know that the ESRF is using a definition, we can try to make any changes backwardly compatible. Without early adopters, we will never make any progress.
>
> With best regards,
> Ray
>
> On Feb 13, 2015, at 7:55 AM, Andy Gotz <andy.gotz at esrf.fr> wrote:
>
>> Hi Carlos,
>>
>> thank you for this update. I was not aware of this device. We will check it out.
>>
>> The main worrying issue we have is the lack of adoption of app defs by the community.
>>
>> Instead of worrying I think we will adopt a solution which does not depend on the app defs until we are sure they are being used by multiple sites.
>>
>> Regards
>>
>> Andy
>>
>> On 13/02/2015 11:43, Carlos Pascual wrote:
>>> Hi Andy,
>>>
>>> Here is some info of usage of app defs in the Tango world (you may be
>>> interested):
>>>
>>> -We at Alba implement sardana recorders for 2 application definitions:
>>> NXscan and NXxas (the first is the default and widely used, and the
>>> second is a proof of concept that is not in production).
>>>
>>> -We also convert propietary  (t)xrm files from Xradia into NXtomo-
>>> compliant files for interoperability with data analysis programs
>>>
>>> -Also, AFAIK the people at Desy are using sardana macros that talk to a
>>> custom Tango Device which can use the application definition files to
>>> store data conforming to "arbitrary" app defs.
>>>
>>> Cheers,
>>>
>>> Carlos
>>>
>>> On Thu 12 February 2015 18:47:36 Andrew GOETZ wrote:
>>>> Dear Tobias + Pete,
>>>>
>>>> thank you both for your quick replies. They clarify the situation a
>>>> lot.
>>>>
>>>> I understand it is difficult to know who is using the application
>>>> definitions but maybe a survey could be done on the mailing list to
>>>> find out how many use which application definition and which software
>>>> implements them.
>>>>
>>>> Regards
>>>>
>>>> Andy
>>>>
>>>> On 12/02/2015 18:07, Tobias Richter wrote:
>>>>> Dear Andy,
>>>>>
>>>>> To answer your first second question first (before I forget): I have
>>>>> not come across anything in NeXus that requires you to duplicate
>>>>> data. Linking is always allowed. There are some opposite cases like
>>>>> the one you quote below, where linking is mandated in a particular
>>>>> way. In my view that’s more of a legacy item. As long as data can
>>>>> be retrieved from the specified location using the underlying
>>>>> contain format, I don’t care how it gets there.
>>>>>
>>>>> Otherwise Pete has written something on subentry that may help. If
>>>>> it
>>>>> doesn’t speak up!
>>>>>
>>>>> The majority of the application definitions where generated at some
>>>>> point far back in time with the best intentions. You have to start
>>>>> somewhere, right? Most facilities are only now slowly taking up the
>>>>> task to implement them (fully). When problems are found we try and
>>>>> resolve them. Unfortunately there is not yet an army of volunteers
>>>>> out there that implements and tests everything for us.
>>>>>
>>>>> Regards,
>>>>>
>>>>> Tobias
>>>>>
>>>>>> On 12 Feb 2015, at 17:39, Andrew GOETZ <andy.gotz at ESRF.FR
>>>>>> <mailto:andy.gotz at ESRF.FR>> wrote:
>>>>>>
>>>>>> Hi,
>>>>>>
>>>>>> at the ESRF we were looking at the application definitions and saw
>>>>>> by
>>>>>> chance that the fluorescence one:
>>>>>>
>>>>>> http://download.nexusformat.org/doc/html/classes/applications/NXflu
>>>>>> o.html
>>>>>>
>>>>>> has a typo in it which makes it unusable as it is :
>>>>>>
>>>>>> *data*: /NXdata/
>>>>>> <http://download.nexusformat.org/doc/html/classes/base_classes/NXda
>>>>>> ta.html#nxdata>>>
>>>>>>      *energy* –> /entry/instrument/fluorecence/energy
>>>>>>           *data* –> /entry/instrument/fluorecence/data
>>>>>>
>>>>>> This begs the question : how many sites are using the Application
>>>>>> definitions and how do you get around the bugs?
>>>>>>
>>>>>> A second question we had is that the way the application
>>>>>> definitions
>>>>>> are defined today implies they duplicate data in the file instead
>>>>>> of
>>>>>> linking to it. This is because they insist on creating NXsubentry.
>>>>>> Please correct me if we have misunderstood the use of NXsubentry.
>>>>>>
>>>>>> Our second question is : does this bother anyone else than us?
>>>>>>
>>>>>> Kind regards
>>>>>>
>>>>>> Andy + Christophe
>>>>>>
>>>>>>
>>>>>> _______________________________________________
>>>>>> NeXus mailing list
>>>>>> NeXus at nexusformat.org <mailto:NeXus at nexusformat.org>
>>>>>> http://lists.nexusformat.org/mailman/listinfo/nexus
>> _______________________________________________
>> NeXus mailing list
>> NeXus at nexusformat.org
>> http://lists.nexusformat.org/mailman/listinfo/nexus