[GRASS-user] [Fwd: [OSGeo-Discuss] Can I do the same GIS tasks with OS (as with ESRI)?]

[John C. Tull]

On Apr 25, 2008, at 1:17 AM, Hamish wrote:

> Wolf:
>>> Arc is perhaps a bit stronger on the cartography side of things, but
>>> with a bit of patience you can produce nice maps with GRASS.
>
> John wrote:
>> Where GRASS fails relative to ESRI products is in cartographic map
>> output. Others may differ with my view, but creating high-quality map
>> outputs in GRASS is non-trivial.
>
> Yes, you are right, it is non-trivial -- but is possible! :) We had  
> hoped
> to fund a student to write a wxPython GUI frontend for ps.map as  
> part of
> this year's Google Summer of Code, but were only given 3 slots to work
> with (and very grateful we are to have those 3). Competition for those
> slots was very tight and we had to say no to some very nice proposals.
> It's an important component that I hope is ready for GRASS 6.4.  For
> GRASS 7 we will be less constrained by the old code to improve the
> PostScript, PNG, and Cairo drivers for a more integrated hardcopy  
> output
> approach.
>
> http://grass.osgeo.org/wiki/GRASS_SoC_Ideas#Cartography
> http://grass.osgeo.org/wiki/WxPython- 
> based_GUI_for_GRASS#GUI_for_ps.map
>
> I find that by sitting down with a hand calculator (1"=72pt) and  
> keeping
> the ps.map reference guide handy you can get publish-quality results.
> Maybe that still takes longer than it should, and the jumping  
> between %
> and inches is a bit awkward, but the results make the trouble worth it
> IMO. If you have many maps to produce from a single template or in an
> automated way, then ps.map is very hard to beat.
>
>
>> I have started using Mapserver to create my cartographic maps.
>
> some ideas here; additional hints are always welcome:
>  http://grass.osgeo.org/wiki/GRASS_and_MapServer
>
>
>> Generic Mapping Tools is another option, but one that I have found to
>> be more challenging than using a Mapserver mapfile.
>
> ditto:
>  http://grass.osgeo.org/wiki/GRASS_and_GMT
>
> GMT will be similar to ps.map in that it is based on scripts and text
> file configurations. Perhaps it is just my exposure level, but  
> personally
> I find ps.map to be less cryptic. (shrug)

These are great points, and good links for reference. Thanks for  
taking the time to provide such a detailed response.

I was trying to use ps.map last fall for some production work, but the  
lack of transparency options for vector layers in my ps.map output was  
a show-stopper for me. I created some nice maps, but have found  
mapserver, with the shp2img command, to be more efficient for my  
needs. Getting a high-quality legend was also a tad difficult. (This  
is not to say that I have been without problems in MapServer, e.g.,  
tiny legends in large, high-resolution maps.)

I agree that ps.map was much easier to get to a usable state than GMT.  
The scripting language was more intuitive for me as well. Similarly,  
mapserver mapfiles are easier for me to work with than GMT.

Cheers,
John