Re: [Pkg-grass-general] debian-gis stable

[Hamish <hamish_nospam@yahoo.com>]

> I work in two GIS labs and both are interested in
> using debian GIS for some production workstations and
> servers.
> 
> I thought it would be good to discuss:
> 
> What are our plans,if any, for creating a stable
> release of our packages?
> 
> do we have a timetable for when we could
> release/promote a stable release? or at least promote
> our repository as being ready for production use?
> 
> Besides alioth going down, I have found debian gis
> repository to be fairly stable. what steps are left
> before we internally would consider our packages
> release infrastructure ready for others to rely on for
> production use?




with respect to GRASS,



current stable is 5.4.0.
current testing is 6.0.0beta1
current development (5.7) is frozen, new development will be 6.1(?)

6.0 release schedule:
  http://grass.itc.it/pipermail/grass5/2005-January/016924.html

GRASS version roadmap + comparison:
  http://grass.ibiblio.org/devel/roadmap.php



with respect to GRASS in Debian,

hopefully the rest of the /tmp/ files can be removed before 6beta2 and
we can replace 5.0.3 in Sarge.

- scripts/     : done.
- init scripts : need to rewrite g.tempfile to not need GRASS fns.
                  (maybe today)
- C code       : use GRASS G_tempfile() fn as needed
                  (help is welcome)

Steve: is that the only thing holding it back?


with respect to "stable" and "production use",

These are both loaded terms which mean different things to different
people so are rather hard to define.

Plenty of people use 6.0 for production work, but as it is still in
testing there are bugs to be worked out & it really helps if you have
someone on site with a bit of programming experince who can isolate the
problems and communicate with the developers' mailing lists for fixes.
This can really put some people off. Others enjoy the speed with which
bugs get fixed and are happy to contribute. That's up to you...
6.0 is certainly "getting there" though. The Mac OSX version for example
is quite polished and non-unix programmer friendly (due to packaging &
not base code I think! which is a good sign).

5.4 is "stable" but that mostly means it is not getting new features
which could introduce or expose new bugs, or break end-users' scripts.
Compared to 5.7/6.0 it can be harder to use and much less powerful, but
it is much better tested - most bugs are known and work-arounds
documented on the mailing list. In fact it may be more buggy than 6.0,
the difference is that they don't change.

5.0.3 is outdated but known.



hope that helps some,
Hamish