Contents of new GRASS core package - Was: Re: Packaging GRASS

Dear GRASS Community,
it seems to be necessary to split the GRASS package. Up to now
we have more than 340 modules in GRASS 4.2.1.

I would like to invite YOU to make suggestions on a

- GRASS basic package and one (or more)
- GRASS professional package (which could/should be splitted again)

For a discussion about the modules to be included you can make use
of the GRASS modules lists in:

- http://www.laum.uni-hannover.de/iln/grass/grass42/capabilities.html
- http://www.laum.uni-hannover.de/iln/grass/grass42/functional_check.html
- http://www.laum.uni-hannover.de/iln/grass/grass42/Whats-new-in-421.html

Basic contents should be the following modules:
- import/export
- internal format conversion
- statistics/reports
- display
- data query (mouse and through files)
- basic map calculations
- TclTKGRASS graphical user interface
          (this needs improvement: support to define NEW locations!)

The professional edition could contain
- simulation modules (erosion, floodplain, water etc.)
- expert systems (binfer etc.)
- image processing
- database interfaces (Postgres-support is important!)

If you agree I could collect the answers and put a summary onto
the 4.2.1-WWWserver.

   Markus Neteler

---------------------------------------------------------------------
Geographisches Institut | Institute of Geography
   Physische Geographie Physical Geography
& Landschaftsoekologie & Landscape Ecology
Universitaet Hannover University of Hannover, Germany
                     neteler@geog.uni-hannover.de
             http://www.geog.uni-hannover.de/users/neteler/
----------------------------------------------------------------------

Hi...

On Tue, Jun 23, 1998 at 02:23:20PM +0100, Markus Neteler wrote:

Dear GRASS Community,
it seems to be necessary to split the GRASS package. Up to now
we have more than 340 modules in GRASS 4.2.1.

I would like to invite YOU to make suggestions on a

- GRASS basic package and one (or more)
- GRASS professional package (which could/should be splitted again)

I've been trying to compile GRASS 4.2.1v17 on FreeBSD (81 patches so far
just to get it to compile :frowning: ), and my first impression is that the biggest
space savings would come from:

1. Removing all of the Linux binary libs (Tcl/Tk, MPEGEncoder, LessTif,
curses, etc.) that you've included into a seperate file for Linux users to
download if they need it (although Tcl/Tk and all of the other bits are
freely available all over the place). Saves >8MB on both the .tar.gz and
extracted (since they are compressed).

2. Removing ./src.related into a seperate archives. mapgen_6.25 is huge (68MB
extracted), and most of that is in mapfin_docs (36MB). Move the little
programs (those under 1MB) to src.alpha.

3. Removing ./untested to seperate archives. If people want grasslinks, etc.
then they can download them.

4. Cleaning up. There are hundreds of old and unused files lying around,
which should be deleted. These are only confusing, since it's difficult to
tell why they were changed. Also there are sometimes two versions of one
module in the source, and there are duplicated headers and files.

For a project the size of GRASS I'm astonished that the source isn't kept in
CVS. That would make tracking it much simpiler (through anonCVS/CVSup) and
also you'd have logs to show why changes were made (rather than a manually
maintained history file). But then most of the source seems to come from the
dark ages of gets()... If I had the time I would go for a complete
reorganisation and a new build process. If I really had lots of time then
I'd go for a complete rewrite :slight_smile:

-Jeremy

--
.sig.gz