[NeXus-tech] NXDL definition of "link"

[Mark Koennecke]

Pete,

Pete Jemian wrote:
>
> These are the NXDL definitions that will need revision due to this 
> change:
>        applications/NXlauetof.nxdl.xml
>        applications/NXmonopd.nxdl.xml
>        applications/NXrefscan.nxdl.xml
>        applications/NXreftof.nxdl.xml
>        applications/NXsas.nxdl.xml
>        applications/NXsastof.nxdl.xml
>        applications/NXscan.nxdl.xml
>        applications/NXtas.nxdl.xml
>        applications/NXtofnpd.nxdl.xml
>        applications/NXtofraw.nxdl.xml
>        applications/NXtofsingle.nxdl.xml
>        applications/NXtomo.nxdl.xml
>        applications/NXtomophase.nxdl.xml
>        applications/NXxas.nxdl.xml
>        applications/NXxbase.nxdl.xml
>        applications/NXxeuler.nxdl.xml
>        applications/NXxkappa.nxdl.xml
>        applications/NXxnb.nxdl.xml
>        applications/NXxrot.nxdl.xml
>        contributed_definitions/NXsnsevent.nxdl.xml
>        contributed_definitions/NXsnshisto.nxdl.xml
>
> If groups can have links, that can and should be written into the 
> Schema.  Is this up for discussion?
>
this is really  a lot of editing. Is the change necessary? Links to 
whole groups are entirely possible.
I think that was also the rationale to make those links a separate 
entity. The important bit is that it
works with the validation tool. So it is Freddies call really.


Mark


> Pete
>
>
> On 10/19/2010 01:27 PM, Peterson, Peter F. wrote:
>> Thank you for noticing (and fixing) this. You are right about links 
>> being an attribute of fields rather than fields themselves. Freddie 
>> will have to speak up on if (and how) this affects validation.
>>
>> A practical issue that you can probably ignore is that we can have 
>> links to groups themselves. I don't know if we need to deal with this 
>> at all.
>>
>> P^2
>>
>> On Oct 14, 2010, at 3:48 PM, Pete R. Jemian wrote:
>>
>>>
>>> While developing documentation for NXDL, I see that the definition of a
>>> link (in nxdl.xsd) is not what I was expecting.
>>> It's wrong.  I want to fix it.
>>>
>>> See http://trac.nexusformat.org/definitions/ticket/129
>>>
>>> Presently, the NXDL schema defines a link (in<xs:group
>>> name="groupGroup">) as a child element of a<group/>. This means an NXDL
>>> specification could have a<link/>  element at the same level as a
>>> <field/>  element. That's not how we use links in NeXus. Instead, links
>>> are used within the target attribute of a<field/>, such as
>>>
>>>    <polar_angle target="/NXentry/NXinstrument/NXdetector/polar_angle">
>>>
>>> My proposition is to change nxdl.xsd:
>>>
>>>     1. remove the existing definition of "link" from
>>>        <xs:group name="groupGroup">
>>>     2. remove the definition of<xs:complexType name="linkType">
>>>     3. add an optional attribute target to
>>>        <xs:complexType name="fieldType">  defined similarly to
>>>        the existing linkType definition.
>>>
>>>
>>> Pending remarks to the contrary, I should like to commit this change on
>>> October 21.
>>>
>>> Pete
>>>
>>> -- 
>>> ----------------------------------------------------------
>>>   Pete R. Jemian, Ph.D.<jemian at anl.gov>
>>>   Beam line Controls and Data Acquisition, Group Leader
>>>   Advanced Photon Source,   Argonne National Laboratory
>>>   Argonne, IL  60439                   630 - 252 - 3189
>>> -----------------------------------------------------------
>>>      Education is the one thing for which people
>>>         are willing to pay yet not receive.
>>> -----------------------------------------------------------
>>> <nxdl_xsd_Complex_Type_nx_groupType.jpeg>_______________________________________________ 
>>>
>>> NeXus-tech mailing list
>>> NeXus-tech at nexusformat.org
>>> http://lists.nexusformat.org/mailman/listinfo/nexus-tech
>>
>