I think a wider view of Open Source GIS related packages might be
relevant.
You can look at Open Source data management, & something like PostGIS for
vector data is pretty effective. Compatability with shapefiles is good.
Raster data can be well managed with GRASS, also GEOtiffs & GMT grd files.
The gmtmath command is very like Matlab in it's approach to working with
scalars & arrays, and it has a number of grd file operators for some very
effective & customiseable raster math.
All here know of the power of GRASS for working with gridded spatial
data 
Commonly used libraries like Proj & GDAL are supported by Postgis & GRASS.
A script to make a world file for a GMT map, then use gdal to make a
GEOtiff creates high quality raster images of the data on which to overly
vectors is straightforward.
Alternative GUI front ends like QGIS give very effective data browsing, &
use the above libraries & tools to happily display Postgis, GEOtiff &
shapefiles, etc.
JUMP gives similar functionality to QGIS as a spatial data viewing/querry
tool, but also allows you to visually edit features (both shapefiles &
PostGIS tables) as well as implementing useful topological facilities.
I suggest you look at Postgis as a vector data repository, add the GDAL,
GEOS, PROJ, libraries, etc, then you can use GRASS, QGIS, JUMP, etc as
tools to work with the data as appropriate. Some scripting will let you
export data to GMT format for publication quality cartography as well.
The same data can also be accessed by Mapserver to web enable things with
minimal work, if required at any time.
The QGIS installation guide includes a description of the install of most
of the underlying packages, so makes a good starting point.
The synergies available by using these interoperable OS packages are
pretty powerful, and give more choice & flexibility than any commercial
suite I know of.
Cheers,
Brent Wood
