Re: [Savannah-hackers] submission of Waves, Clouds, and Sand - savannah.nongnu.org

[Jaime E . Villate]

On Fri, Feb 21, 2003 at 06:13:17PM -0500, address@hidden wrote:
> 
> A package was submitted to savannah.nongnu.org
> This mail was sent to address@hidden, address@hidden
> 
> 
> M.C. Lewis <address@hidden> described the package as follows:
> License: gpl
> Other License: 
> Package: Waves, Clouds, and Sand
> System name: oceanscene
> Type: non-GNU
> Description: OpenGL-based research project for simulating natural
> phenomenae.  First plan is ocean waves, then clouds, then terrain (sand).
> Intend to use latest features of 3D cards such as texture/fragment shaders,
> register combiners, and vertex programs.  Development on GNU/Linux.  >

> There are a few \"free\" issues here.  One is that the two video card
> manufacturers which have extensive OpenGL functionality (ATI and NVidia)
> have proprietary drivers, which load into the x86 Linux kernel, sometimes
> FreeBSD, Windows (of course), and that\'s about it.  Unfortunately, no
> useful Free drivers exist for this hardware.Hopefully, by making advanced
> graphics software that runs primarily on GNU/Linux and other free platforms,
> 3D on Linux will stay current, so that we won\'t just be starting on 3D
> years down the road when Free drivers become available.

Due to our policy, we can only host your project if it only uses features that
exist in the few free drivers available, for example the ATI Radeon driver
distributed by XFree86. ¿Are you willing to keep your software in Savannah
usable with such drivers?

Without that commitment from your part, we won't be able to host your project.
I sympathize with your plan to stay current and start working on features
which are not yet freely available, but you should keep those newest features
away from Savannah because we do not want to encourage Savannah users to start
using the proprietary drivers of ATI and Nvidia.

> Second, I may wish to make use of some software with free-looking, but 
> nonstandard, licensing.  E.g., GLVU:
> http://www.cs.unc.edu/~walk/software/glvu/
...
> Also, OpenEXR:
> http://www.openexr.net/
...
> Since I will be using the GNU GPL for my software, I am not sure if I can
> \"swallow\" some of the software, meaning release it under the GNU GPL if
> compatible; or if I can only link against it.

Sure, you can use their code in your own project. In both cases it is free
software with a license that amounts pretty much to public domain. As long as
you give proper credit and respect their licenses you are safe; keep the
original copyright notes and copying permission statements as instructed by
the licenses. It would be also useful to make a summary of all the external
packages you have used and their licenses. If you modify one of those external
programs, you can write something such a "this program is a modified version
of ABC; the original copyright information for ABC is..." followed by the
copyright information for your modified version (i.e. the GPL license info in
your case).

> Third, I have downloaded an extensive set of research papers describing
> graphics techniques, some of which are hard to find; I am not sure how much
> of these I can excerpt for tutorials on a technique.

As long as you do "fair use" of that scientific literature, you can do it
without asking anyone's permission. The rules for what constitutes fair use
are not very explicit, but to get an idea of it imagine that you were going to
publish your tutorials in a scientific journal ¿would they allow you?  ¿would
they accept it as an independent work/review that gives proper credit to other
authors, and is not a copy of something already published by someone else?
The answers depend on each particular scientific field; referees in each field
know what is common background knowledge that anybody can freely use and what
is novel knowledge that can be attributed to a particular author.

> Fourth, Intel has a very fast \"approximate math library\", which seems to
> be under a rather onerous license.  They know how to write a faster sqrt(),
> seeing as they made the chips.  Probably any independent implementation of
> some of these functions using SSE2 would result in identical ASM code.
> http://www.intel.com/design/Pentium4/devtools/ As an example of the
> usefulness of fast math in 3D, POVray doubled in speed just from a better
> sqrt() on the Alpha:
> http://www.uwsg.iu.edu/hypermail/linux/alpha/9601.1/0046.html

Sorry, I do not have the time or the skills to analyze that issue. You should
seek legal advice in some other list more oriented toward legal aspects of
software. You can talk to someone at the FSF, for example. From that dramatic
decrease in processing time, it sounds like an algorithm that is important to
keep safe from claims of ownership.

I will keep you project registration on hold until next Sunday, waiting to
hear whether you accept the commitment to stay away from proprietary drivers
dependencies.

Regards,
Jaime