Mapserver/PostGIS general questions about how best to do things

[Stephen Woodbridge]

Hi all,

I posted the bulk of this on the PostGIS list, but got not comments
back. Since this is also mapserver oriented and the questions are
general in nature, I'm hoping I might get some suggestions here.

First, I have to say getting PostGIS up and running on freeBSD was very
easy so Kudos to the port maintainer and freeBSD. Also get data loaded
and get mapserver running with PostGIS was not nearly as hard as I had
thought it might be and I'm really liking what I see, even if I'm not
sure what to do with it all yet. Also getting mapserver running was just
as easy.

What I have done (and I will try to get you a link soon to see it) is
load up the Census zipcode polygons and some of the Census summary File
1 and 3 data and I can generate some thematic maps.

Questions:

1) What are your thoughts on loading all Tiger into PostGIS versus
keeping most of it in shapefiles and only using some of the data out of
PostGIS? Performance? Are there other benefits to put it all in PostGIS?
If I had it all in PostGIS, are there other cool features that I could
have that I don't now with shapefiles?

2) What is the best way to scale or other map an attribute column into a
known range for use with mapserver CLASS EXPRESSIONS. The Census data
has literally thousands of columns any of which you might want to select
as the basis for a thematic map, or even some expression of multiple
columns. Conversely, a mapfile would have some fixed number of classes
that would say render you polygons as say shades of color based on the
attributes.

Has anyone done this? anyone have any ideas or suggestions on how to do
this? I realize that 2) is slightly off topic, but any help would be
appreciated.

Thanks,
   -Steve