[Freegis-list] SPRING GIS software

[Adrian Custer]

Dear Mr Camara,

Like Bernhard Reiter, I thank you for taking the time to explain your
position to this list. I have always had the sentiment that the SPRING
project was close in philosophy to the FreeGIS project and it is nice to
hear your views. This response also follows Mr. Reiter's in suggesting
that you could, in the framework of your current institutional
structure, open up the source code of SPRING without as much difficulty
as you present. Hopefully these comments can help you re-frame the
issues you face.

I signed up for the right to download SPRING about two and a half years
ago but I decided that I could not use the product. I am a scientist
collaborating with researchers in Africa and must guarantee that the
work I do can easily be packaged onto a CD-ROM and redistributed to
groups who have computers but no internet connectivity. For this
purpose, the right to redistribute is paramount.

A separate part of my work has been exploring the data structures
required in a GIS to hold 4-dimensional raster, vector and other (say
mathematical field descriptions--not raster fields but 4-d continuous
functions) in such a way as to be rapidly searchable and analyzable. For
this work, the ability to see other implementations could be a huge
source of knowledge.

It strikes me that neither of my interest in your code conflict with
your work and that both could be quite useful as a source of credit to
the INPE. Both are satisfied simply by releasing a snapshot code under
an open source license. So what obstacles remain? The obstacles you
present are the following:
1) Legal issues
2)  Product responsibilty
3) Committment to the product rather than the development process
4) Code documentation
5) The TERRALIB open-source effort as a substitute

Like Mr. Reiter, your second point seems superfluous since any
responsibility you have, you also bear under the current distribution
system. The legal issues do not seem to be insurmountable and if you can
persuade your lawyer to HELP you do what you intend, you undoubtedly can
figure out a way to acheive this.

Your third point arises, I believe from a misunderstanding of the open
source development process. By opening up the code, INPE bears no
responsibility to become the maintainer of an open source project. The
FreeGIS community is able to use portions of your work (with credit
still going to INPE for those portions) or the FreeGIS community can
fork the SPRING effort to continue from that point on. There is no
requirement that INPE set up a public CVS archive and allow the
community into your development process. Indeed the cost may very well
be prohibitive in time and energy to allow this. Especially if INPE has
decided to move into a new, modern project of this kind, it seems much
better to focus on your new effort.

In this context the fourth issue is unfortunate but not a show stopper.
It would certainly be useful for one of the developers to spend an
afternoon to write a simple white paper explaining the structure of the
code base but this is not a requirement. However, if INPE is moving into
an open source effort, it would be helpful for us to see the previous
effort and give you feedback about the documentation which is really
necessary to make code readable from the outside. I am currently trying
the read the code for the GNOME spreadsheet Gnumeric, a classic
open-source development effort, and it is quite difficult with little
documentation beyond a few white papers. 

Finally the TERRALIB effort. I am really excited to hear you are working
on such a library. As I stated earlier, the presence of the code of your
earlier work could only help the community at large by allowing a
comparison and enabling discussion of the successful part of INPE's
previous work.

I encourage you to pick a stable snapshot of SPRING for release, to
explain to the lawyers that they must help you pick a well understood
open-source license to use for your release and to post the snapshot on
the web with an explanation that you do not have the resources to move
the project to an open source development effort but do want to offer
your work to the public. At least then I could install openSPRING and
see your great work! :-)

Thanks again for your time in explaining your position. I look forward
to the discussions over TERRALIB.

kind regards,
adrian custer

acuster at nature.berkeley.educational (truncate)

Dept of Entomology
U.C.Berkeley
Berkeley, CA 94720-3112
USA




> 
> A recent thread on FreeGIS-list
> which discusses the availability of
> the SPRING source code has been brought
> to our attention, and I would like to
> clarify some issues.
> 
> First, we are extremely supportive
> of the FreeGIS effort and free software
> in general. That said,
> why don´t we simply put SPRING on the
> web as source code? 
> 
> As a starting point, there are legal
> problems which we haven´t been able
> to sort out. But these are not the only ones.
> 
> Putting software on the web, on whatever
> form, implies an enormous responsibility
> on the part of the producer. It does not
> matter if no formal warranty is given,
> the individual or the institution need
> to act in a responsible fashion. If people
> use your software for anything, you´d better
> provide quality, or else your reputation 
> will be compromised.
> 
> SPRING is a very large software product, with
> hundreds of functions, and 500.000+ lines of
> C++ code, produced over a period of a decade,
> by a large team (more than 100 man-years, at my
> last count). The project did not
> begin as a free software effort, but rather as
> an integrated GIS+IP freeware product for the
> Brazilian and Latin American users. 
> 
> More than 10,000 people worldwide have downloaded
> SPRING, and we have a strong commitment to SPRING 
> as a full GIS product, capable
> of supporting real-life projects, and we
> have a significant number of real-world users.
> 
> Making SPRING available as open source on the web 
> would require a major effort by our team in terms 
> of documentation, and further effort in terms of supporting 
> fellow code developers. These commitments are 
> outside of our current capacity.
> 
> We have given priority to supporting users
> of SPRING, producing enhanced versions, correcting bugs,
> generating user documentation, preparing training courses
> and maintaining the same code running on Windows, Linux 
> and Solaris. These activities keep a lot of us busy all
> the time.
> 
> Therefore, SPRING-product has, for us, 
> a much higher priority that SPRING-open source.
> I hope you appreciate the difference.
> 
> Taking everything into account, we have taken the
> option of using our experience with SPRING to design
> a product that´s meant to be open source. That 
> product is called TERRALIB, whose first version will
> be made available on-line sometime in 2001. 
> Our aims in TERRALIB is to build a library for GIS
> software development, and our next-generation GIS 
> products will be built using it. 
> 
> All of you that develop and maintain 
> open source code know how much effort is involved.
> Therefore, I hope you will appreciate your cautious
> approach, which is aimed at only making commitments
> that we can maintain.
> 
> Regards,
> Gilberto Camara
> Co-ordinator for R&D in GIS
> National Institute for Space Research (INPE)