[NeXus-tech] Schema [SEC=UNCLASSIFIED]

[Akeroyd, FA (Freddie)]

Hi Nick,

Thank you for posting your DTD converted schema files. How flexible is
the conversion process? When I started on the DTD -> schema process I
also wanted to add a new hierarchy of derived types to simplify the
layout and future class/definition creation by instrument editors. In
their current form the DTD converted files would need quite a bit of
editing to get them into this new style, but if it was possible to
modify the conversion process that could indeed reduce the workload
significantly.

Regards,

Freddie


From: nexus-tech-bounces at nexusformat.org
[mailto:nexus-tech-bounces at nexusformat.org] On Behalf Of HAUSER, Nick
Sent: 17 July 2008 07:01
To: nexus-tech at nexusformat.org
Subject: [NeXus-tech] Schema [SEC=UNCLASSIFIED]

My apologies for missing the VC. It is not related to the "temporary"
requirement for me to focus on Bragg Institute operations.

Please note:

Bragg is committed to writing NeXus files in hdf5 and xml format.

Our production systems are writing NeXus files for 6 instruments. We are
as rigorous as possible in our adherence to the standard as "we
interpret it". Bragg fully supports the move from meta-DTD to schema, as
this will remove much of the ambiguity and allow more rigorous
validation. This support currently doesn't extend to funding. This
year's Australian Government budget for ANSTO relates to a 10% decrease
in staffing which has an immediate effect on my team and my role.

Schema generation
How is Freddie travelling with his schema writing? I had a look at
http://definition.nexusformat.org/schema/1.0/. I've attached a zip file
of a conversion done by Ferdi Franceschini to relaxng schema
http://www.relaxng.org/ which was done generatively from the base
classes in Jan 07. This was done with a group of xslt scripts to refine
the metaDTD first to a valid XML file, then to a schema. rng is the
normative form, from which xsd can be generated. I've also attached the
generated NXdetector.xsd. Ferdi will provide more info on this technique
if required. I think it may reduce the workload significantly.

Regarding Darren and his working on metaDTD to schema translation;
Darren left Bragg in September 07. My management no longer provided
funding to the model server and NeXusBeans projects. Lack of progress &
deliverables was the reason for freezing these projects. Darren is a
skillful architect, and a good developer. His weakness is that he won't
compromise architecture to deliver product. Bragg are using several
patterns and architectures that Darren recommended to us during his time
here, but we are not using any of his source code.

If the NIAC were to use Darren services, I suggest a fixed price
contract of say $20k (200 hours) rather than an hourly rate. With a
fixed price contract, NIAC would run the risk of only partial
completion.

Regards,

Nick.