[GRASSLIST:1170] Re: creating a desktop GIS application using GRASS

[Glynn Clements]

Greg Ford wrote:

> > > > The problem is that because
> > > > of GPL, it is impossible to create proprietary applications for GRASS.
> 
> I imagine that many GRASS programmers have contributed their time
> for free over the years - with the confidence that the GPL would protect
> their work from being exploited by someone making a proprietary
> product from their code.  Perhaps GRASS could be converted to LGPL,
> but you would need the agreement of every programmer who has
> contributed code (or documentation?) to the project.

Yep. Which makes Radim's objections to the GPL largely irrelevant.

> But you do have a point... (I have not looked at the GRASS source so
> someone correct me if I'm wrong) the architecture of GRASS does not
> appear to expose a library interface - so any application needs to
> call external applications/commands or scripts which is not usually
> as efficient or slick.

However, there is one big advantage: robustness. A bug in an
individual command can crash that command. With a single monolithic
application, any bug could crash the entire application. Furthermore,
a command's memory allocation is automatically released when that
command terminates. A monolithic application would be much more prone
to memory leaks.

> Consequently the GRASS GUI is a little clunky,
> but would be quite difficult to replace.

Not necessarily. There aren't any significant technical obstacles to
producing something better than tcltkgrass. There are many minor
obstacles, in the form of programs which perform largely gratuitous
user interaction.

The only major obstacles are:

1. Finding someone to actually write it.
2. Agreeing upon a GUI toolkit to use.
3. A small number of programs which are inherently interactive
(v.digit, i.points).

-- 
Glynn Clements <glynn.clements at virgin.net>