[GRASS5] New raster formats

I've noted that there has been talk of a new raster format for future versions of GRASS. What kind of changes are being considered? One thing I miss from SPANS was the quadtree format which allowed for very quick processing of data without loss of spatial resolution. I realize that implementing this in GRASS would mean rewrites to much of the existing code to deal with the complexities of the format but I thought it would be worth throwing out the idea for consideration, as it seems that we constantly face bigger and bigger data sets as computing power grows. Since PCI's takeover of SPANS, they've dropped this and there is no commercial vendor willing to put in the effort to gain the benefits from the use of this format (speed and disk space). I'm curious if this has been considered.

T
--
Trevor Wiens
twiens@interbaun.com

The significant problems that we face cannot be solved at the same
level of thinking we were at when we created them.
(Albert Einstein)

Trevor Wiens wrote:

I've noted that there has been talk of a new raster format for future
versions of GRASS. What kind of changes are being considered?

Tiled storage (as an alternative to rows), more flexible support for
maps with a small number of categories, fixed-point maps, built-in
caching (eliminating the need for the segment and rowio libraries).

One
thing I miss from SPANS was the quadtree format which allowed for very
quick processing of data without loss of spatial resolution. I realize
that implementing this in GRASS would mean rewrites to much of the
existing code to deal with the complexities of the format but I
thought it would be worth throwing out the idea for consideration, as
it seems that we constantly face bigger and bigger data sets as
computing power grows. Since PCI's takeover of SPANS, they've dropped
this and there is no commercial vendor willing to put in the effort to
gain the benefits from the use of this format (speed and disk space).
I'm curious if this has been considered.

It isn't amongst the features which I've considered for a new raster
format. Unfortunately, I suspect that it would probably make writing
modules too complex.

--
Glynn Clements <glynn@gclements.plus.com>