metadata: where should we be going?
I am working on a bunch of perl modules which are nominally GIS
related. I'm further than before, but I am not ready to say it
will be done soon.
The latest hurdle, is Akima interpolation. Yes, Math::GSL has one
particular Akima interpolant. With options, I have 8. And
currently Math::GSL doesn't work on i86 where the hardware
supports 64 bit, and the user is running a 32 bit OS. Some
strange stack problem, which effects a bunch of programs.
A long time ago, I wrote a natural cubic spline in FORTRAN-77.
I've seen cubic splines implemented in 6 lines of APL. Akima
splines are unusual, in that they allow "corners" in a data set.
And they are better at avoiding wiggles. But, to introduce a
corner into a dataset takes some deliberate effort.
Many materials properties have "corners", but to satisfy Akima
interpolatants, artificial points would probably still have to be
added.
But, if someone goes to the effort to properly introduce a corner
into a dataset, such that Akima interpolation will recognize the
corner (if it is capable), I suspect this interpolant will be used
for a long time. And it should have metadata accompaning it.
The place where I've seen the biggest recognition, need and use of
metadata is GIS. So, I am calling you experts. :-)
Is there an accepted metadata standard which works with GIS? Do
people have ideas on what is required?
Gord
Reply to: