[GRASS5] GRASS development roadmap proposal

[Markus Neteler]

On Thu, Jul 17, 2003 at 06:51:52PM +0200, Bernhard Reiter wrote:
> On Thu, Jul 17, 2003 at 05:48:33PM +0200, Markus Neteler wrote:
> > On Thu, Jul 17, 2003 at 05:29:07PM +0200, Bernhard Reiter wrote:
> > > On Thu, Jul 17, 2003 at 05:08:31PM +0200, Markus Neteler wrote:
> > > > On Thu, Jul 17, 2003 at 04:36:01PM +0200, Bernhard Reiter wrote:
> > [...]
> > > > > > Two problems have to be addresses:
> > > > > > - in a 5.0.x version we should not introduce
> > > > > >   datum transformation as the results will be
> > > > > >   different
> 
> That seems to be a point which should be easy to agree upon.

Perfect.

> The reason is that the proj support would change to much
> in the code.

Yes, you got it :-)
 
> 
> > Sorry, should read:
> > In general it's the idea that 5.0.x support hopefully can be discontinued
> > in favor of the new developments.
> 
> As long as there is no established stable version it certainly cannot
> be discontinued and even after this, conservative users might continue it
> by themselfs.

Well, 5.0.x should be discontinued as soon as possible in terms of
introducing new features. Otherwise we never arrive at a better GRASS
implementation which is really needed and currently only almost
on Radim's shoulders.

> > > Thinking more about it, renaming in a form close
> > > to your proposal, might be a possible solution.
> > 
> > Fine, so let's continue the discussion. More comments welcome.
> > 
> > My goal is to get out a new release for bugfixes (5.0.3) and
> > a new release containing the datum transformation and PROJ4
> > improvements (5.4.0) and probably a first release of the new
> > vector developments if Radim agrees.
> 
> One drawback of the renaming would be the higher communication costs
> and the further delay of the 5.1.x experimental release.

It always depends on how many developers participate (hi!).
The higher communication costs seems to be my problem as usual
(updating the web site, changing docs etc). With a clear roadmap
published in a good place (first page?) the users can quickly
understand what's going on.

> Also the release with the proj improvements would be more experimental
> than the 5.0.3 release, so we'd probably should to release it as 5.y
> where y is an odd number.

As I proposed: have another look at the initial mail
http://grass.itc.it/pipermail/grass5/2003-July/005913.html
"
stable                    experimental
----------------------+--------------------------
5.0.3                     5.3.0 (was: CVS 5.0 HEAD)
- bugfixes for 5.0.2      - datum transformation
- G3D activated           - improved NVIZ
                          - r.terraflow (C++)
"

> Renaming this to 5.1.x or 5.2 would bluff people expecting 5.2.x to the be
> stable version with vectors. Thus we had to rename it to 5.3.x.
> 
> Subsequently current 5.1.x can only be renamed to 5.7.x.
> That was just writing down a couple of reasons 
> or thinking loud.

Also as I proposed :-)
"                          5.7.0 (was: 5.1.0 renamed)
                          - new 2D/3D vector
                          - DBMS support
                          - keep sites format [1]
                          - new Makefile system
                          - initial cleanup libs
                          - initial code inline docs
                            (doxygen)
"

> I now believe that Markus proposal is fine and will support it,

Thanks.

> given the following remarks:
> 
> 	It all only makes sense if the 5.3.x cycle is 
> 	compartively very short.

Well, we'll see, So far I haven't seen any short cycle. It depends
as usual on our "human resources".

>       In other project we'd wait
> 	until it is integrated with 5.1.x or have it as an add on
> 	with a big warning sign.
> 
> 	Renaming 5.1/5.2 -> 5.7/5.8 has to happen early, 
> 			when communicating the whole plan
> 
> 	It should be clarified that 5.9.x does not necessarily
> 	lead to 6.0.0, it can also lead to 5.10.x. :)

Sure, why not. We had 22 version of GRASS 4.2.1 and 11 betas of 5.0.x,
so we can keep the tradition :-)

Cheers

 Markus