On Sun, Dec 14, 2008 at 2:34 PM, Akarsh Simha <akarshsimha@...> wrote:
> Dear Friends from the Stellarium Team,
> We are the KStars Team [http://edu.kde.org/kstars], and we write this
> to you with a vision of some sort of collaboration between Stellarium
> and KStars as initiated by Stellarium's Fabien Chéreau in a mail to
> one of our colleagues.
I am glad to get a reply :)
> We think a collaboration would be mutually beneficial and could be
> carried out at several levels. The first collaboration that we have
> been able to picturize immediately is in terms of sharing of
> data. This would have several advantages. Our users would not have to
> install twice a similar dataset if they use both softwares. We would
> take advantage of any improvement in the dataset without having to do
> the work twice.
> External contributors would be able to contribute new data to both
> softwares simultaneously. If you agree in such a common dataset we can
> then discuss what are the needs of stellarium and kstars.
I totally agree with this, in fact one of the reasons I sent this
proposal email some months ago is the following: we (the free
astronomy software community) are challenged (or even outperformed on
some aspects) more and more by WWT or GoogleSky & co. It is difficult
for us to catch up with them because to provide great features such as
multi-resolution background images collections, we need resources for:
- gettin data sets, like e.g. the DSS sky survey, and processing them
for our needs
- getting authorization to use and distribute them
- buy and maintain servers for serving the data
All these steps require a lot of work and also potentially money (for
buying servers/bandwith). Hence my idea of trying to cooperate to do
these kind of operations together.
On our side, we finally got the formal authorization from several
observatories to host our own DSS copy on servers. This took about 2
months to get this agreement, and now we could try to extend it to
other open source softwares, not only Stellarium.
Now on the interoperability side there is also definitely room for
cooperation, in Stellarium we are currently thinking of redesigning
important aspects of managing extension data sets (a GUI for adding
new catalogs, landscapes, online images collection etc..). The work is
not very advanced yet. I was thinking to generally use JSON-based
formats for describing the data sets.
> We could also envision some code sharing. Indeed, right now our code
> bases are largely different. Maybe we can think of an astrolib - a
> common library using C++ / Qt + eigen2 / quaternions to implement
> much of the math that KStars / Stellarium uses - in the near future.
I just had a look at the eigen library, that I didn't know, it is
indeed very interesting! Thanks for the info :) We'll need to make
performance tests, but if it is possible, using this library in
Stellarium could make some piece of our code only dependent on
Qt+Eigen. If you do the same on your side we could quickly share
The performance-critical code in Stellarium (for projections etc..) is
already quite optimized and the design is now pretty clean, so I don't
think there is much use for us to change that.
We however have nice, very fast and accurate planets computation code
(from Johannes Gadjosik) that you may want to use. This code is
separated from the rest of Stellarium and has no dependencies (see
On our side we would be quite interested e.g in using your code for
managing Telescope control, we already have such feature but it will
need a big refactoring, so maybe we could just use yours :)
> We have already, so to speak, started collaborating after Jérôme
> Sonrier brought in Stellarium's Sky Culture data into KStars. We
> would like to hear more suggestions regarding ways to collaborate from
> your team. Maybe we should have a common mailing list to discuss such
I just read Matthew' email about that, so here we are :) I will
continue the email there later.
> With Regards,
> The KStars Team
>  Eigen2 is a simple
> (http://eigen.tuxfamily.org/index.php?title=API_Showcase) math
> template library performing automatic vectorisation whenever possible.
> Celestial mechanics can tremdendously benefit from vectorisation and
> the library seem to be efficient:
> http://eigen.tuxfamily.org/index.php?title=Benchmark . The eigen lead
> developer told he could help us using it if necessary.