Binary distributions vs source code

[Bob Harrington]

In article <20364.9410040845 at bsamaga.nerc-bas.ac.uk> you write:

> James Darrell McCauley asks why people want binary versions, rather than 
> compiling them themselves. My own experience is the answer! I am an experienced 
> programmer, having been in the business for about 20 years. I have developed 
> complex systems. However, attempting to compile a complex system is very tricky. 
> I have attempted to compile GRASS on a DEC Alpha, which is not one of the 
> platforms that are in the documentation. I cannot get a clean compilation, and I 
> do not have the information to find out why - at least, not without spending a 
> LARGE amount of time understanding the code. There are also substantial system 
> dependencies - the exact location of X libraries, the exact version of the 
> window manager - the list goes on. Having failed to get a fully functional 
> compilation, I am about to give up on GRASS, because I cannot afford the time 
> that it would take to make it work from source code. If someone were to offer me 
> a binary version, that I could install fairly simply, I would take it up. Any 
> offers? I am running OSF/1 version 2 on a Decstation 3000 AXP, with Motif ( 
> whatever the current version is).

I'm not too crazy about putting our whole GRASS installation into our anonymous
ftp area, but I think I can point you to something that will allow you to 
successfully compile GRASS on an Alpha running OSF.  ftp to ohlone.ced.berkeley.edu
and get /pub/grass/grass-alpha-axp-0.5.tar.Z  It contains some scripts that
make global substitutions to the GRASS source code that allow a successful
compilation on the DEC Alpha.  

Has anyone had problems with this package?  It seems to be working fine for me, 
but I haven't exercised every GRASS program, so I can't say for sure that the
compilation is problem free.

Bob Harrington
bobh at hwr.arizona.edu
Hydrology and Water Resources
University of Arizona, Tucson, Arizona