Re: layers of abstraction, and how Gnome can win



On Sun, Dec 10, 2000 at 09:51:42AM -0800, Brian Behlendorf wrote:
> 
> b) Is it Gnumeric's mission to be the best free spreadsheet out there,
> whether you're using KDE or Gnome or even just a bare windowmanager?  If
> someone contributed a series of patches to get Gnumeric to compile cleanly
> and look nice under Win32 native widgets, would they be accepted?

Gnumeric's goal is to become the best spreadsheet (free or
otherwise).  This is obviously a large target.  In some places we
have a long way to go, in others we're already beginning to pull ahead.
One such area is probably maintainability.  Our code base is fairly
modular and quite nimble at this point.  As such I'd have to divide
potential win32 patches into two categories.

    1) Core changes to work better with different OS/C libraries.
    These types of things have been accepted in the past to ease
    support for *BSD vs Linux.  They would be accepted in the future
    for win32 if the changes are kept neat and tidy.

    2) Support for a native win32 gui.  I would not accept this as a
    'patch' to the current Gnumeric tree.  Such a beast would be too
    far reaching and too prone to #ifdef forests.  However, there is
    now a mostly complete MVC split.  This was intended to allow
    different frontends (no-gui, text, etc).  If someone wanted to
    write and maintain a win32 gui they are welcome to.  That seems
    to be an XP approach similar to abiWord.  I do not think this is
    a good idea.  The marginal gain of exact win32 look and feel
    does not seem worth the duplication.  IMHO it would be far
    better to improve Gtk+/Gnome-libs/GAL under win32 to match win32
    behaviours, than to 'fix' each application separately.

At this point Gnumeric is window manager agnostic and there seems no
pressing reason to change that.  It runs nicely under GNOME and
'bare' windowmangers.  I have not personally tried it with KDE, but
we get enough bug reports and wishlsts from KDE users that that it
seems likely to work there too.




[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]