[GRASSLIST:1161] Re: creating a desktop GIS application using GRASS

[Bernhard Reiter]

On Wed, Sep 10, 2003 at 11:31:56AM +0200, Radim Blazek wrote:
> On Tuesday 26 August 2003 12:15, Bernhard Reiter wrote:
> > > The problem is that because
> > > of GPL, it is impossible to create proprietary applications for GRASS.
> >
> > I consider this an advantages, because in GRASS' history
> > each time a vendor made a proprietary application with GRASS code
> > it was locked in and GRASS development stagnated.
> 
> You mean proprietary version of GRASS, I mean applications for GRASS.

I actually mean proprietary applications for GRASS
that extended GRASS and thus would have been a part of it.

> I don't think that GRASS could stagnate more than it is stagnating now,
> we are mostly in the state of bugfixing last few years. 

I don't see that picture of stagnation that you are painting.
Bug-fixing and winning over users is very important
and GRASS has made siginificat progress in the last years.

> Proprietary versions don't seem to be problem for projects like Mozilla
> or Open Office!

That is the other way round, 
those projects originated from proprietary software
and those original propritary versions still exist for some reason
or another.

After having unbalanced (and therefore not GNU GPL compatible licenses)
they are also mainly available under the GNU GPL (openoffice)
and an obscure triple licensing for mozilla (also including the GNU GPL
which does not make much sense).

So they do not really aim at having more proprietary versions 
coming into existing.


> > On the other hand, if you make scripts for GRASS you might be able
> > to make an application which is proprietary,
> > so your statement above is only partly correct.
> 
> ???!!! Application using scripts? That definitely cannot be competitive
> compared with applications for proprietary GIS, tightly integrated into
> the base system.

It depends on the application.
But know you exactely point out the reason why GRASS cannot allow
proprietary applications, they would need to be tightly integrated
thus being a derived work of GRASS' core not just using a fixed interface.

> > > While for commercial GISes,
> >
> > (I assume you are talking about proprietary GIS).
> >
> > > it is possible to distribute both free
> > > and proprietary applications, GPL GIS allows GPL compatible extensions
> > > only.
> >
> > As explained above this is not correct, both is possible,
> > depending if you derive your application from GRASS or build it on top
> > like using GRASS commands as scripts.
> 
> As explained above, usually it is impossible to make good,
> copetitive application using commands and scripts.

I doubt that, but for the sake of arguments let's assume that is true.
This would still not force the conclusion that we should relicense GRASS 
to allow such applications being made proprietary.
It is possible to make commercial Free Software applications
which will allow us to actually build a competitive advantage over
proprietary commercial offerings.

> > > It is illusion to think, that all required extensions can be created
> > > under GPL.
> >
> > Why that?
> > In 1983 Richard Matthews Stallman called to create a complete
> > operating system (which is a lot more than GRASS extensions)
> > and was successful.
> > Many people thought this wouldn't work, but it did.
> 
> We must think about a ratio between the work required
> and the number of users, say 'code size'/'number of installations'.
> This ratio is too high for higly specialized GIS extension
> comparing to operating system. It seems to me that GPL is good
> only for SW with that ratio under certain limit, that means to have 
> enough users and contributors.

It is an interesting argument.
The GNU example shows that you can start with a high ratio and lower it
over time. The copyleft licenses helped to accelerate the process.

GIS itself is a broad subject so GRASS has the potential of a higher
number of installations. We should seek ways to raise that (what we
are already doing, the transition from the first author group to
a join world-wide and more stucture developement community is going fine, 
I'd say).

As for specialised applications with a low number of installations 
there are also example in other Free Software areas that this can be done
without being proprietary. Especially because if the needs are special
the companies have the tendency to put more money into the solution.
The two most prominent examples are the GNATS (GNU ADA compiler) and Cygnus.

> There are also other aspects, working on OS is attractive while 
> most GIS applications is just routine and boring work. 

I cannot see why an operating system (kernel) is inherently more interesting
or attractive. There are parts of GNU/Linux that are very boing work.
Of course you are right, we should make GRASS development and
contribution to it more attractive and fun if we can. 

> HW and SW producers are interested in free OS development,
> but almost nobody having money in free GIS development.

There are several possible reasons for this observations of you.
The potential of money for GNU/Linux developments or Free GIS
development is almost the same. I even think that there is more
potential per GIS user for Free Software GIS.

> GIS is for us something like OS for others, i.e. enviroment, starting
> point. While OS is covered by exception in GPL, GIS is not.
> To argue by OS under GPL is not correct, because of that exception in GPL.

It is quite easy to seperate the OS system from the applications.
For GRASS for technical reaons this is far away.
It is okay to argument with the GNU system, because this project
also includes "userland" applications which are not under that exception
and it still worked. The OS exection is a compromisse there because it 
needed to bootstrap itself 
and Free Software operating system weren't there at the time.

> > You don't know until you try.
> 
> I believe that somebody can realy waste his money and write an application 
> under GPL, why not. 

There are many situations and examples for this 
where this is not a waste of money.
You are clearly speaking the proprietary software vendors way here.

> The problem is that we need massive movement in this 
> area not one exception. More important than to talk about theoretical 
> possibility is to look around at the reality - no such applications. 

You mean no Free Software applications for GRASS?
I believe there are other reasons for this.

There have been many in the past, but GRASS fell in a hole that
we are now crawling out again. Bringing clear freedom protecting
licensing terms to GRASS were an important step in doing so,
otherwise GRASS would have died the canabalised way.
We need to continue that clear way without getting fuzzy again.

> > > Until such applications are available (good quality,
> > > localized, with commercial support) most serious GIS users cannot
> > > choose GRASS as their main system.
> >
> > I believe that GRASS usage could be a lot higher
> > and is already raising in small steps.
> 
> Very small steps, the gap between GRASS and proprietary GIS is 
> bigger and bigger every day.

You need to be more specific in what areas the proprietary GIS
are improving, then Free Software can try to catch up sooner or later.
Also GRASS is bound into the framework of other Free Software 
for geoprocessing. It might be that they will tackle part
of the proprietary tasks. I basically see that the other way round
Free Software is catching up fast on almost all areas.

As to the size of steps, I'd rather make small solid steps
that bug ones that carry a high risk.

> > What you demand is "the permission" to vender-lock users in
> > and in my view that is not bringing more freedom, but less.
> 
> Not at all, what I want is to give users the freedom to use both free
> and proprietary extensions for GRASS. 

In principle that is a goal worth debating,
however I don't see how we can do this with the current GRASS technology
without giving away our "crown jewels".

> GRASS users are not infantile idiots, who need to be protected by you GPL. 

I didn't write the GNU GPL, the FSF did.
As part of the development team I'm strongly in favour for using
its protecting for the work.

> It must be their decision what SW they want to use, not your or our.

The underlying assumption of you sentence is wrong
and often seen in proprietary FUD tactics.
You are not forced to use a software with a license you don't like.
That includes GRASS and the GNU GPL. 
If you use it, then you are bound to the license,
in particular when you make derative works.

The clear Free Software licensing of GRASS has helped it a lot
to regain credit with users and developers in the last years.
We know have GRASS available on some scientific GNU/Linux distributions
which can directly run off the cdrom drive.
New efforts go into documentation and the German Grass User Association 
has made GRASS beginner courses and will now send a flayer to many
institituions in Germany to promote GRASS.
GRASS under GNU GPL has come a long way!

	Bernhard
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : http://lists.osgeo.org/pipermail/grass-user/attachments/20030910/cadb7b5b/attachment.bin