[
Date Prev][
Date Next][
Thread Prev][
Thread Next][
Date Index][
Thread Index]
[
List Home]
Re: [science-iwg] Eclipse as a platform for scientific applications
|
Hi Torkild,
my 2 cents:
In addition to the reasons given by Philip (which I agree with) I also see:
(o) main reason for me is that Eclipse offers the possibility of
building a commercial quality UI with open source which is
multi-platform. Many of the features of a good UI in commercial tools
are available for free in Eclipse (e.g. help, search, cheatsheets,
workbench, perspectives, etc.)
(o) the modular approach of the Eclipse plugin system is excellent and
is well suited to writing complex UI tools because it scales well
Some disadvantages I have encountered are:
(o) a minority of scientists are sceptical of any tool based on Java -
probably because they do not know and/or like Java
(o) Java is not a common choice in the scientific world. Python has
better acceptance but most Python UI's are not professional nor do they
scale
(o) workbenches need to interact with different languages especially in
the scientific world. The Java to Python/Fortran/Matlab/IDL bridges are
not always that good
That being said I think Eclipse/RCP offers the best platform for UI
development for the kind of tools we need for multi-disciplinary
science. At the end of the day the majority (fortunately) of scientist
do not care about the language in which their tool is written (e.g.
Matlab UI is written in Java and it does not seem to bother most
scientists).
Scientific visualistion used to be an issue in SWT but this has since
been solved with the various plotting libraries out there. Take a look
at Dawn as example (http://www.dawnsci.org).
Best regards
Andy
by
On 12/07/2012 09:09 PM, Torkild U. Resheim wrote:
Hi Philip
You have some good arguments there, even a few I did not think very important. But in retrospect it makes perfect sense. I guess working with the Eclipse platform for so many years made me forget about how hard it is to do extensibility, platform independence, modularity etc. on other platforms. That you also like Java is great. I've mostly seen Fortran in math heavy systems, however integrating these (into Eclipse) is not really an issue, and there are solutions such as JCuda and JOcl/Aparapi that makes Java even interesting even for these types of problems.
Thanks for you help!
Best regards,
Torkild
7. des. 2012 kl. 20:44 skrev Philip Wenig <philip.wenig@xxxxxxx>:
Hi Torkild,
the main reasons why I chose Eclipse as a platform for my scientific application are:
* Its modularity - it enables to make a clear cut between different concerns. Moreover, it enforces to think about a well designed architecture of the software previously.
* Its extensibility - if I need a e.g. a GIT plug-in for my RCP, I just can install from the marketplace and don't have to write it by my own. Great!
* Its platform independence - it's marvelous that it runs under different operating systems.
* Its native look - frankly, who really likes Swing? The first time I saw a RCP/SWT app I didn't believed that it was Java :-) ... and people like to have a native look, me too.
* It's Java - previously I've done a lot of VB6 and I seriously thought about C#, but there are so many fantastic Java libraries for scientific approaches (Apache Math, SWT-XY-Graph, SWT Chart, ...). Hence, it was clear to take Java.
Cheers,
Philip
Am 07.12.2012 20:19, schrieb Torkild Ulvøy Resheim:
Hei all,
I've been asked to do a presentation for SINTEF about Eclipse as a platform for scientific applications. And while I'm quite familiar with the platform itself I'm curious about your reasons for choosing Eclipse for your projects. It would be great if you could help me out by giving me a short summary.
Best regards,
Torkild
_______________________________________________
science-iwg mailing list
science-iwg@xxxxxxxxxxx
http://dev.eclipse.org/mailman/listinfo/science-iwg