> Not because of disrespect of devels efforts, no.

Good. :)

> Python is not suited for a desktop application: it's too slow, uses
> too much resources, the legibility of the code is worse than c++,  it
> has no good library for reading dive computers, it's hard to package
> it on osx Windows and Linux regarding libraries, bindings to another
> languages are a pain to be made. You don't see c or c++ apps going to
> python, but the opposite : once the python app stops scalling, it
> must go to either c or c++.

I would respectfully disagree on several of these points: the GUI is
not appreciably slower than in C (granted, I only compared gtk and
pygtk, and have no experience with qt), and any speed loss is
certainly offset by convenience; the legibility ultimately depends on
the programmer (and the programmer's affinity to document code) in both
languages; and C APIs can be wrapped more-or-less easily. Packaging
might indeed be an issue for windows and osx, I don't know enough to
comment because I am ignorantly exclusive to linux. I would never
suggest C or C++ number crunching code to be ported to python, but I
would advocate for porting the (G)UI into python. Mixing C and python
(not sure about C++, never tried it) is really not that much of a pain
once you get used to it. I am an astrophysicist by day and I can tell
you that most of our field is rapidly migrating all of the UI stuff to
python. Our own code (phoebe) runs C for number-crunching under the
hood, but all UI is done in python, and we never regretted the move.
Your mileage may vary, of course, I just figured I'd ask because, if
subsurface UI was in python, I'd be much better positioned to
contribute to the code -- for whatever that's worth.

Cheers,
Andrej
_______________________________________________
subsurface mailing list
[email protected]
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Reply via email to