[GRASS5] why GPL

[strobe anarkhos]

>
>Your missing the argument.  Most of the people developing for GRASS
>are end users (including myself).  The primary motivation is to have
>a free and freely available GIS system.  By having many people 
>contributing freely to GRASS under a free license we all reap the
>benefits of each others work.  If we re-license code to allow third
>parties to use our work to make non-free software the motivation to
>work on GRASS is lost.  If I wanted to pay for the software, I'd just
>buy one of the many commercial GIS's already on the market.  Basically
>you end up with a one-way street.  Developers of proprietary add-ons
>reap all the benefits, where develepors of the free parts have to pay
>for the non-free parts.  I don't see any point to it.  Might as well
>use ESRI or Integraph products and write to their API's (I know they'll
>happily use code developed by third parties if it is robust and general
>enough).

You're the one missing the whole point (which I'll summarize at the end of this email).

Any changes to GRASS itself would have to be LGPLed. GRASS is a system, not just one tool. If somebody wants to extend GRASS they would have to release those changes under the LGPL. The motivation for extending GRASS does NOT come from GPL, it comes from whether the GRASS library is useful or not. The realization of this becomes even more clear when you deal with more flexible models like the one I proposed. When more people use GRASS, THEN more people have an interest in extending the API.

The advantage of LGPL (or a license like it) is that people can release software which uses GRASS but doesn't have to use the friggin GPL! This software doesn't extend GRASS in the least (and can't in it's present monolithic state), it just uses GRASS.

I wanted to release MY code as LGPL so a) people can't sell my code and b) other people can write useful software for it without releasing it under the GPL. I do not have this freedom even though I am using my OWN model. I want MORE people to use my software, not less.

> > Nobody can take your code and sell it, it's available to everybody.
> > The problem with the GPL is it isn't available to everybody and it
> > will prevent it from becoming widely used. Instead it will be stuck as
> > a nice stand-alone app instead of a general tool. I don't have any
> > plans to sell GRASStep or any project branching from that work, but I
> > would like to LGPL license it in case somebody wants to write a new
> > tool for it under their own license. 
>
>GRASS already is a pretty specialized application framework.  Even
>though GIS's have seen tremendous growth, they will always fill a very
>specialized niche.  

A niche which is substantially smaller than GIS software because it can't be used as a general GIS toolbox.

GRASS is more than a specialized application 'framework', it's a very inflexible one. I scoff at your "tremendous growth".

>Anyway, anybody can sell the GPL code already (as long as the source is
>made available under the GPL license).   That's not the issue.  The
>issue is one of reciprocity.

Look at other projects I have used which are LGPL. For example GNUStep and Quesa.

Anything within the GNUStep or Quesa API which doesn't work I have to fix myself, and if that code is released it has to be under the LGPL. If you define a model for GRASS like I started to do in another thread on this list any functions pertaining to a particular model will have to become PART OF GRASS!

For example one of the model classes is a vector set. Let's say I want to implement a method which calculated the length of that vector set and returns it. I would place that method within the vector set class where *everybody* can use it for free.

> > The LGPL is not 'weak', I think you should try to be more open minded.
>
>Well, RMS himself has argued against using the LGPL, and he wrote it!
>Anyway, it's not about being open minded, its about being selfish.  I'm
>selfishly withholding my code from those that would use it to make a
>deritative work and then turn around and charge me for that derivative
>work and never offer me unencumbered access to their source code.  I
>see no benefits from the LGPL.

Ask me if I care what RMS says. I only care about the license verbatim. 

The problem with that logic is nobody is forcing you to pay for squat. Nobody can patent an API and nobody can extend GRASS's API without releasing the code under the same license.

The advantage of the LGPL comes if we implement a GRASS framework whereby it can become a general tool more people have an interest in contributing to it. Some of these people may have commercial interests in mind, others like myself do not.

OpenStep is one example of this. Even though it's closed source Objective-C allows people to add methods to existing closed-source classes. These were in general use and many were merged with the larger body.

IF you see NO UTILITY WHATSOEVER in general wide-spread use of GRASS where the GRASS framework remains free and grows as more people add functionality to it then your point is taken. But if that is your position why bother open sourcing it in the first place. It's a matter of unrealized potential, something I may even accept as "tremendous growth".

As for ESRI, if GRASS becomes the general framework used for dealing with GIS data, that pressures companies to adopt the GRASS model. You make it sound like this would be a gold mine for ESRI, it would be a poison pill. Instead of using their APIs, people would use the GRASS framework and only tools using the same general framework could interface with their apps. Thus for ESRI to make any sales they would have to make components for the GRASS framework. And how many of these components will provide unique functionality? They could only sell those which provided unique functionality. The value of their monolithic application would drop considerably.

However if the GRASS framework were GPL this could never happen and the uses for and growth of GRASS will remain highly limited. I mean there is virtually no point in making GRASS a more useful framework if the primary advantage of this model can't be exploited. 


---------------------------------------- 
If you want to unsubscribe from GRASS Development Team mailing list write to:
minordomo at geog.uni-hannover.de with
subject 'unsubscribe grass5'