On May 3, 2006, at 9:29 AM, Ralf Stephan wrote:
Finally, as far as I can tell, it looks highly feasible on a technical and temporal standpoint. My only fear (and my main question here), is that I may face an OOo-type situation where the GUI and the engine are so bonded together that using another GUI involves going deep into the general code. It does not seem that this is the case as Inkscape is built on libraries and previous projects, but you never know.
Can you be more specific?
Inkscape is a C/gtk+ application (where the C was OOPish style C) in the process of moving to C++/gtkmm. Of course, libraries are used for this.
*and* some of those libraries include gtk+, gtkmm and glib. Gotta remember that detail. :-)
One thing to keep in mind is that the bulk of Inkscape code, as a visual editor, is to visually do things. This usually involves integration with the UI widgets.
However, all details aside, I can't see where being "highly feasible" on either a technical or temporal viewpoint is a solid assertion. Separate the UI, and you're left with a simple SVG renderer. And that is already completed. And not even looking into the question of GUI and engine being tied together, just the scope of work to get a native UI implemented is rather enormous. And even if there were not major technical issues, the amount of time from pure grunt-work alone blows out the 'temporal' assertion.
But the good news, as others pointed out, is that there is not active work getting GTK++ itself to be OS X native. This has been ongoing, and thus is a large reason for Inkscape developers not to reinvent the wheel when waiting a little can get it all done for free.