[GRASS5] GRASS 5.0.3 released

Thierry Laronde tlaronde at polynum.com
Fri Nov 7 16:29:21 EST 2003


On Fri, Nov 07, 2003 at 11:35:29AM -0500, Helena wrote:
> Bernhard Reiter wrote:
> 
> >>>GRASS has quite a conservative user base.
> >>>That is something that always was a concern for Markus, too.
> >>>Beside that the code base is gigantic.
> >>
> >>It is perhaps too late but it took me some time (and the great help of
> >>Hamish's advices) to understand the organization of code/CVS.
> >>
> >>May I say that, for me, now, the simplest way (separating difficulties)
> >>should be first to stop indeed _all_ additions to 5.x 
> > 
> > 
> > This is already in place for many 5.x lines with x<7.
> > 
> > 5.0.x is only for "critical" bug fixes.
> > 5.1 and 5.2 is discontinued.
> > 5.3.x has some improvement, but is only bugfixing.
> 
> I would say that 5.3. has some substantial improvements in terms
> of projection capabilities, nviz and addition of some new modules.
> We are using it quite intensively and it is very stable for the things
> that we do (of course that is just a small part of GRASS). 

Don't misunderstand what I'm saying. I don't suggest to go back to 5.0.
I just say : 5.3 aka HEAD == TO_BE_RELEASED <- 5.4 and this 5.x serie
should be frozen. 

And to prepare a major change (one of the core engine) we
_must_ first discard all the unneeded code and rethink the organization
just in order to not try to put new wine in old skins : putting a new
engine in old weird bodywork won't give a racing car. This will just
lead to a crash!

The fact that I'm new here has the advantage of giving the sight of an
"outsider". I was first pleased to see the power of the portion of GRASS
I used (the perls). And I gave a look in the code... And I thought :
"This is unmaintainable!". There are perls, yes, but there is  mud too.

Just for debugging purposes you can't let two distincts variables
change randomly: the internals and the organization. First change the
organization dramatically reducing the size of the code (and preparing
it to receive a new engine), and once this is achieved try to put a new
engine.

And indeed, beside the new vector engine, cleaning and rethinking the
organization of the code will lead to major changes too that will both
absolutely legitimate an increasing of the major number.

For GRASS it should be the Hercules year: one labour per month. But this
must start by the Augean Stables...

-- 
Thierry Laronde (Alceste) <tlaronde at polynum.org>
Key fingerprint = 0FF7 E906 FBAF FE95 FD89  250D 52B1 AE95 6006 F40C




More information about the grass-dev mailing list