Re: Proposal for a geodetic deformation model format

Hi Martin

Thanks for your comments.

Who is the intended audience?  Short answer is anyone interested!

From our (LINZ) point of view our primary motive is to get something going,
and having it working in PROJ/GDAL would be a huge step forward.  The GeoTIFF
format provides us the opportunity to do that, as without that there is not a suitable
grid format available to build the general deformation model with.

As you rightly say there is also a lot value in a standardisation process with OGC,
though of course that is a (much much) longer process.  Certainly that is something
that elements of this could feed in to.  It could end up with something similar to
this, in which case it would be great for us, or it may end up with something different, in
which case it wouldn't work with the current LINZ deformation model in any case.
Either way it is useful for us to have this capability now.

So initially the audience is anyone else who thinks they might use it, and particularly
anyone who thinks they could use it, but it would need to be just a bit different for
it to work for them.  Primarily I imagine that would be from geodetic agencies who
are using or developing deformation models.  I did forwarded the proposal to the IAG
working group on time dependent datums.

On HDF5 vs GeoTIFF.. As you rightly say apart from some details this proposal
is largely format agnostic.  I know the ESRI GGXF initiative is now proposing using
HDF5, and certainly that could serve as well.  Curiously I did propose HDF5 as an
option for deformation models a while ago, but at that time there wasn't much
appetite for using it as (IIRC) it wasn't so well suited to deploying with online web services,
From a PROJ point of view GeoTIFF will already be there.  Using HDF5 would add another
huge dependency.

The choice of multiple/single file implementation was a surprise for me, as I've
been wanting to come up with a single file implementation for a long time.  But I
found as I was developing this proposal that there seemed to be more advantages
to a multiple file implementation.  Because GeoTIFF provides the nested grid capability
the number of files is not excessive (unlike the LINZ publication format as individual
grids).  One advantage is that different versions of the deformation model will only differ in a few
of the component files.  Most would be common to the versions, so using a multiple
file implementation allows these to be shared.

Cheers
Chris

________________________________________
From: PROJ [[email protected]] on behalf of Martin Desruisseaux [[email protected]]
Sent: Thursday, 5 December 2019 10:52 p.m.
To: [email protected]
Subject: Re: [PROJ] Proposal for a geodetic deformation model format

Hello Chris

Thanks for the proposal, I had a quick look on it. The explanation about
rational, required metadata, etc. are very nice. My first question would
be who is the intended audience. Is it a proposal for the PROJ/GDAL
community, or e.g. a starting point for standardization at OGC? The
reason why I ask is because in my understanding, the choice of GeoTIFF
format for PROJ/GDAL is based on pragmatic reasons specific to that
community (close connection between GDAL and GeoTIFF since they are
maintained by the same peoples; more familiarity with GeoTIFF than with
HDF5 and CF-Conventions). They are valid reasons for PROJ/GDAL, but if
the target audience is wider then I would suggest to reconsider since,
if it was not for PROJ/GDAL convenience, I think HDF5 + CF-Conventions
(i.e. NetCDF 4) would be more appropriate for those kind of data.

However a "GeoTIFF versus HDF5" discussion could be left for a separated
thread. The vast majority of the proposal document would still apply in
both formats, except the discussion about single versus multiple files
which would not be needed anymore. Some attributes like "extent_east",
"extent_north" would be revisited in a HDF5 format because they overlap
(but are not identical) with standard CF-Conventions attributes, but the
ideas would still the same.

     Regards,

         Martin


_______________________________________________
PROJ mailing list
[email protected]
https://lists.osgeo.org/mailman/listinfo/proj

________________________________

This message contains information, which may be in confidence and may be subject to legal privilege. If you are not the intended recipient, you must not peruse, use, disseminate, distribute or copy this message. If you have received this message in error, please notify us immediately (Phone 0800 665 463 or [email protected]) and destroy the original message. LINZ accepts no responsibility for changes to this email, or for any attachments, after its transmission from LINZ. Thank You.
_______________________________________________
PROJ mailing list
[email protected]
https://lists.osgeo.org/mailman/listinfo/proj