On Wed, Aug 22, 2007 at 09:22:51AM +0200, Abdelrazak Younes wrote: > But you talk like if it was an easy task to begin with. Believe me, it > was and is still not. And it was very time consuming too. If I may > sumarize the pros about multiple frontends: > > - system integration: this is good I guess for commercial application > (even if I have my doubt) but this is open source.
Qt pretty well integrates into a variety of systems. This includes _native_ Windows, Mac, and even the _Gnome_ desktop. I have still to see significant integration problems that would be solved by using either a Qt3 or a Gtk frontend. > The moment the development effort costs too much with regard to the > number of users, I am sorry to say that this is not worth it. Those > users should either switch to another OS or make their OS (or LyX) > good enough for their use (or pay someone to do it). Right. > - force the kernel/frontend separation: I really think this was a good > thing at the beginning of the GUII work. But at one point, the multiple > frontends approach hampered the development because any change in the > separation layer needed to be fixed *four* times. By concentrating on > getting _one_ frontend clean, I am 100% sure that the kernel/frontend > separation state is much much better now than when we had multiple > frontend. I'd even dare to say that it should be much easier to add a > new frontend now than when we had multiple frontend. Exactly. > >Yes, they are better spent in adding new build systems. I stand in my > >opinion that the Qt3 and gtk frontends were murdered. > > The difference is that some people _are_ _willing_ to maintain the build > systems. You cannot force anybody to work on anything. Their time is > their time. Or somebody pays. But then, even payed LyX development (we had that already...) almost failed as the development happened against an older version and syncing was too tedious. Andre'
