[GRASS5] The status of 5.0

GRASS GRASS developer
1

Helena wrote:

all of the enhancements that you have mentioned are extremely useful -

do

you have
at least a brief man pages for the new ones? (especially the MODFLOW
related?)

No man pages yet. In fact, some need a little more tweaking before they
are generally useful. I wrote these in a spasm of productivity last
fall and they allowed me to turn a large project around so fast that it
made my collaborator's head spin. My whining may make it seem like I'm
unhappy with GRASS 5, but I'm not. GRASS is a *very* big hammer.

I realized this morning while looking into the problems with r.out.ascii
that I also inserted a -M option in r.out.ascii to write MODFLOW ascii
arrays.

Markus, would it be possible to include these new and enhanced

modules?

Now the question is where they should go? GRASS5.1? Are they ready for

the

current GRASS5.0?

With documentation they would be ready for 5.0, but if we want to make a
stable 5.0 release in the near future then it probably isn't a good time
to be adding new modules.

As for the "big GRASS 4 installations" are there any? All those that I

have

known here in the US
had to switch to ESRI products years ago (CERL, NRCS, USGS, etc.),
as it was manadated as a standard for government agencies.

Probably a dark day for GIS development. What ever happened to MOSS?

The only large GRASS 4 installation I know of is at Blacklands. I think
the full name is the Blacklands Agricultural Research Station, Texas A&M
University. It's in Temple, Texas. They used (are using?) GRASS and
SWAT for watershed simulation of major parts of Texas to assess the
possible water-supply impact from brush control projects. I came across
them when we needed to review their model of the Canadian River basin.
They sent us several hundred megabytes of GRASS4 and SWAT files.

I've always assumed that Baylor has a functional GRASS4 installation.

I recall from a conversation on this list more than a year ago that
there might be some installations in France and possibly others
elsewhere in Europe.

This is on a barely-related tangent. A friend of mine once expressed an
interest in using GRASS in his research into the Martian
paleoenvironment. I thought that was an interesting application.

Roger Miller

2

On 25 Mar 02, at 9:54, rgrmill@rt66.com wrote:

Helena wrote:

> As for the "big GRASS 4 installations" are there any? All those that I
have
> known here in the US
> had to switch to ESRI products years ago (CERL, NRCS, USGS, etc.),
> as it was manadated as a standard for government agencies.

Probably a dark day for GIS development. What ever happened to MOSS?

The only large GRASS 4 installation I know of is at Blacklands. I think
the full name is the Blacklands Agricultural Research Station, Texas A&M
University. It's in Temple, Texas. They used (are using?) GRASS and
SWAT for watershed simulation of major parts of Texas to assess the
possible water-supply impact from brush control projects. I came across
them when we needed to review their model of the Canadian River basin.
They sent us several hundred megabytes of GRASS4 and SWAT files.

I've always assumed that Baylor has a functional GRASS4 installation.

I recall from a conversation on this list more than a year ago that
there might be some installations in France and possibly others
elsewhere in Europe.

This is on a barely-related tangent. A friend of mine once expressed an
interest in using GRASS in his research into the Martian
paleoenvironment. I thought that was an interesting application.

If no one objects, I will launch a survey on the user's list, so that we can try
to get an idea. I will have people send the answers to me personally, and will
then report back with a summary of the results.

Here are some (quick) ideas for questions to ask. I don't know if it is better to
leave responses open (more varied answers for us) or to give multiple choice
of answers (quicker to answer for users) Please comment and add:

- What do you use Grass for ?

- How many people use Grass in your environment ?

- Which version of grass ?

- If not 5.x, why haven't you switched ?

- Do you use Grass "as is" or have you adapted it for your usage (i.e.
modified/added modules and/or scripts) ?

- Would you prefer to wait longer and have the developers test Grass5 more
thoroughly before its release, or would you prefer a rapid launch with
subsequent bug-fixing releases ?

Moritz

3

On Mon, Mar 25, 2002 at 09:24:09PM +0000, M Lennert wrote:

On 25 Mar 02, at 9:54, rgrmill@rt66.com wrote:

If no one objects, I will launch a survey on the user's list, so
that we can try to get an idea. I will have people send the
answers to me personally, and will then report back with a summary
of the results.

It is always good to known more about the users.

Here are some (quick) ideas for questions to ask. I don't know if
it is better to leave responses open (more varied answers for us)
or to give multiple choice of answers (quicker to answer for
users) Please comment and add:

Doing short surveys it usually better to leave responses open,
but have good explanations or examples.

- Would you prefer to wait longer and have the developers test Grass5 more
thoroughly before its release, or would you prefer a rapid launch with
subsequent bug-fixing releases ?

The answer of this question probably will not have a big influence.
It leaves out too many details which are important.
Very important is that we have motivated developers and we can
construct a good process.

It is never a good idea to release really bad untested software,
but all software will have bugs. The truth is somewhere in between.

Because of the unique situation GRASS is in,
I do not think we have much of a choice as trying to make
a transition in the way GRASS and its releases are handled.

4

On Mon, Mar 25, 2002 at 09:24:09PM +0000, M Lennert wrote:

On 25 Mar 02, at 9:54, rgrmill@rt66.com wrote:

>
> Helena wrote:
>

>
> > As for the "big GRASS 4 installations" are there any? All those that I
> have
> > known here in the US
> > had to switch to ESRI products years ago (CERL, NRCS, USGS, etc.),
> > as it was manadated as a standard for government agencies.
>
> Probably a dark day for GIS development. What ever happened to MOSS?
>
> The only large GRASS 4 installation I know of is at Blacklands. I think
> the full name is the Blacklands Agricultural Research Station, Texas A&M
> University. It's in Temple, Texas. They used (are using?) GRASS and
> SWAT for watershed simulation of major parts of Texas to assess the
> possible water-supply impact from brush control projects. I came across
> them when we needed to review their model of the Canadian River basin.
> They sent us several hundred megabytes of GRASS4 and SWAT files.
>
> I've always assumed that Baylor has a functional GRASS4 installation.
>
> I recall from a conversation on this list more than a year ago that
> there might be some installations in France and possibly others
> elsewhere in Europe.
>
> This is on a barely-related tangent. A friend of mine once expressed an
> interest in using GRASS in his research into the Martian
> paleoenvironment. I thought that was an interesting application.
>

If no one objects, I will launch a survey on the user's list, so that we can try
to get an idea. I will have people send the answers to me personally, and will
then report back with a summary of the results.

Here are some (quick) ideas for questions to ask. I don't know if it is better to
leave responses open (more varied answers for us) or to give multiple choice
of answers (quicker to answer for users) Please comment and add:

- What do you use Grass for ?

- How many people use Grass in your environment ?

- Which version of grass ?

- If not 5.x, why haven't you switched ?

- Do you use Grass "as is" or have you adapted it for your usage (i.e.
modified/added modules and/or scripts) ?

- Would you prefer to wait longer and have the developers test Grass5 more
thoroughly before its release, or would you prefer a rapid launch with
subsequent bug-fixing releases ?

You may have a look at:

http://grass.itc.it/survey.html
(this was disabled, no links to this page for some time). The results
should be available "grass@baylor.edu". Maybe you can ask
for the results of the last two years and continue (eventually
modified).

Markus

5

On Tue, Mar 26, 2002 at 11:29:08AM +0100, Bernhard Reiter wrote:
[...]

It is never a good idea to release really bad untested software,
but all software will have bugs. The truth is somewhere in between.

Luckily GRASS 5 is quite tested - at least what a few people
can test within 3 month. Now it is time to get more users on
that.

Markus

6

On Tue, Mar 26, 2002 at 12:02:40PM +0100, Markus Neteler wrote:

On Tue, Mar 26, 2002 at 11:29:08AM +0100, Bernhard Reiter wrote:
[...]
> It is never a good idea to release really bad untested software,
> but all software will have bugs. The truth is somewhere in between.

Luckily GRASS 5 is quite tested - at least what a few people
can test within 3 month. Now it is time to get more users on
that.

That is exactly the plan.
Schedule something like:

GRASS 5.0.0 developmemt time: now

  All developers have time to fix and bring in (moderate) additions
  for the grass5.0.0 release now.

Code Freeze and Branching for Release: about 22th of April
  We will create a release branch.
  We will release grass5.0.0pre4 shortly after alpha tests.
  This will be beta tested.
  Only release critical bugs will be fixed on that release branch.

Release: about 22th of Mai

  After another month we will release grass5.0.0 from the branch,
  or release pre5 or merge back without release as a last resort.

We can't get faster than this with a project as huge as GRASS.
This is the tightest reasonable schedule I can imagine.

After this repeat schedule for GRASS 5.0.1.
Note that _major_ enchancement should go into grass5.1
or be publised as extra additional patches or tarball releases
if testing is needed. These extra file should always be based
on the last tarball release (currently pre3).

  Bernhard

7

On Tue, Mar 26, 2002 at 12:20:50PM +0100, Bernhard Reiter wrote:

On Tue, Mar 26, 2002 at 12:02:40PM +0100, Markus Neteler wrote:
> On Tue, Mar 26, 2002 at 11:29:08AM +0100, Bernhard Reiter wrote:
> [...]
> > It is never a good idea to release really bad untested software,
> > but all software will have bugs. The truth is somewhere in between.
>
> Luckily GRASS 5 is quite tested - at least what a few people
> can test within 3 month. Now it is time to get more users on
> that.

That is exactly the plan.
Schedule something like:

GRASS 5.0.0 developmemt time: now

  All developers have time to fix and bring in (moderate) additions
  for the grass5.0.0 release now.

Code Freeze and Branching for Release: about 22th of April
  We will create a release branch.
  We will release grass5.0.0pre4 shortly after alpha tests.
  This will be beta tested.
  Only release critical bugs will be fixed on that release branch.

Release: about 22th of Mai

  After another month we will release grass5.0.0 from the branch,
  or release pre5 or merge back without release as a last resort.

We can't get faster than this with a project as huge as GRASS.
This is the tightest reasonable schedule I can imagine.

As you proposed the plan - what about electing you as the release
manager for the upcoming pre4 release? We may rotate this job,
this is also found in other Free Software projects.

Related to this some document is available now:
GRASS CVS and release strategies
http://grass.itc.it/codemanagement.html
(authors: Bernhard Reiter, Radim Blazek and Markus Neteler)

Markus

8

On Thu, Mar 28, 2002 at 06:29:37PM +0100, Markus Neteler wrote:

On Tue, Mar 26, 2002 at 12:20:50PM +0100, Bernhard Reiter wrote:
> On Tue, Mar 26, 2002 at 12:02:40PM +0100, Markus Neteler wrote:
> > On Tue, Mar 26, 2002 at 11:29:08AM +0100, Bernhard Reiter wrote:

> That is exactly the plan.
> Schedule something like:
>
> GRASS 5.0.0 developmemt time: now
>
> All developers have time to fix and bring in (moderate) additions
> for the grass5.0.0 release now.
>
> Code Freeze and Branching for Release: about 22th of April
> We will create a release branch.
> We will release grass5.0.0pre4 shortly after alpha tests.
> This will be beta tested.
> Only release critical bugs will be fixed on that release branch.
>
> Release: about 22th of Mai
>
> After another month we will release grass5.0.0 from the branch,
> or release pre5 or merge back without release as a last resort.
>
> We can't get faster than this with a project as huge as GRASS.
> This is the tightest reasonable schedule I can imagine.

As you proposed the plan - what about electing you as the release
manager for the upcoming pre4 release? We may rotate this job,
this is also found in other Free Software projects.

The most important job as a release manager is to judge
if certain components are ready.
As I cannot conduct many tests myself I'm not a good choice.

I offer to be the release assisstant for 5.0.0 for you, though.
This means we will get a 5.0.0 release, but if developers and
testers do not take me serious enough it will be pretty ugly.
Nevertheless we need this release.

--
Professional Service for Free Software (intevation.net)
The FreeGIS Project (freegis.org)
Association for a Free Informational Infrastructure (ffii.org)
FSF Europe (fsfeurope.org)