Martin Vermeer a écrit :
I belive that qt3 will stay of a (long) while yet.
That's it... what I don't like is opposing an agreed good idea with the
argument "don't you have anything better to do".
Well, it is surely a good idea for the user but, if you look for fun
coding, you will enjoy Qt4 coding better than Qt2/3 ;-)
BTW on a side note: did anybody else notice how good it is that we have
GUI front-end independence, even if one would hold that having multiple
front-ends is undesirable: LyX contains a number of external elements
that evolve in time, like C++, boost, and the GUI toolkit(s). Of these,
C++ is standardised and unchanging (even if the compiler is not), boost
a bit less so. But the most volatile external element is the GUI. It is
actively developed and changes rapidly.
If you talk about the gtk and qt4 frontends I agree but when I look at
the qt2 frontend, I don't see much development in cvs history in the
last two years. Whereas I saw a flurry of new code in the core parts of LyX.
And thanks to GUI-I, we could just drop in Abdel's qt4 without in any
way destabilising the rest of LyX. Great isn't it?
This is great yes and I am very happy about that :-)
But, as a side note, I would like to point out that the controller
framework seems overly complicated to me. I remember that while
developing the Qt4 frontend I found myself spending more time trying to
fit into the controller framework than doing the actual porting to Qt4.
Well, it's done now so need to simplify it. But I would have very much
prefered a clean lyx core API that the frontend would use instead of
delegating the intelligence to the controller.
That's just IMHO and it's Friday isn't it :-)
Abdel.
Another BTW: when will be have the 'personal' or 'sandbox' branch
directory? We have SVN now but psychologically we're still in CVS.
- Martin
