Re: XML scientific data formats

• To: Russell Steicke <r.steicke@xxxxxxxxxx>
• Subject: Re: XML scientific data formats
• From: Bill Hibbard <hibbard@xxxxxxxxxxxxxxxxx>
• Date: Fri, 14 Dec 2001 08:58:39 -0600

Hi Russell,

> > I'm about to start some work on an XML Adapter and have considered two
> > formats:
>
> I've written an XML Adapter for a project at the Australian Bureau of
> Meteorology.  It was initially fairly specific to that project, but we
> believe it's now general enough for other uses.  It will load and save
> one or more Fields or FlatFields in XML format.  For a while we have
> been talking about releasing the code for wider use, but have yet to
> receive executive approval for that.

I encourage BOM management to release your XML Adapter. It
would be a real contribution to the science community. It is
too generic to have much commercial value in itself, but
releasing it demonstrates BOM's expertise to those shopping
for commercial systems built on VisAD.

> Also, I've done some preliminary work on saving arbitrary VisAD data in
> XML format, ie having all of the VisAD MathType and Data hierarchies
> expressable in XML.  The direction I was heading was specific to VisAD,
> (eg using VisAD names for the XML element names), not a general format
> like XSIL.

I looked over the XML extensions XSIL, XDF and ESML
mentioned yesterday as possible bases for representationing
VisAD data objects. None of them capture the generality of
the VisAD data model. XSIL is explicitly extensible, and XDF
and ESML can be extended implicitly by designing custom
encodings of data in their tables and arrays (similar to the
various conventions for encoding curved and irregular grid
topologies in netCDF files). This explicit and implicit
extensibility can be used to encode VisAD data objects, but
other sofTware reading these XSIL, XDF or ESML files wlll
not be programmed to decode those encodings. While XSIL, XDF,
ESML or even generic XML software will provide some abilities
to access and manipulate encoded VisAD objects, the benefits
will be limited.

The point of this is to not attach too much value to using
XSIL, XDF or ESML. Thus if your XML Adapter does not, its no
big deal. Of course, I will be open minded to anyone who tries
to convince me otherwise.

> If you're interested, I could describe the format we're currently using.

I hope BOM management will release your XML Adapter. I'd
be delighted to make a link from the VisAD web page to a
description of your work, and/or to include your adapter
as part of VisAD.

Cheers,
Bill
----------------------------------------------------------
Bill Hibbard, SSEC, 1225 W. Dayton St., Madison, WI  53706
hibbard@xxxxxxxxxxxxxxxxx  608-263-4427  fax: 608-263-6738
http://www.ssec.wisc.edu/~billh/vis.html

• References:
  • XML scientific data formats
    • From: Ugo Taddei
  • Re: XML scientific data formats
    • From: Russell Steicke