This is fairly accurate, and in the case of GNUStep, it's an achilles heel, because Apple is actively pushing those developer's to rework the User interface in Cocoa, but they aren't saying rewrite the logic. At this point much of the Windows code you are talking about is C++, some built around MFC or ATL, but most of the logic is pure C++. It will require some work to make cross platform if it isn't alreay, but that's not a big deal. With ObjectiveC++ it's trivial to consume that logic with a new UI, and as such the UI works fairly well. That's the rub. Objective C++ isn't well support on the non-Apple toolchain yet, so efforts like WebCore will continue to languish on GNUStep.
Making matters worse is that GNUStep is installable on Windows, but ProjectCenter won't build on that install. GORM will, but what use is that without PRojectCenter ?. These issues continue to plague GNUStep. None of them are unresolvable, given enough resource, but therein lies the problem. The trick is to get some of those Windows developer's along with the Linux developers and the Mac developer's to embrace GNUStep as a workable toolkit, which right now as a Windows and Mac dev who has a linux box as a server, it's not. I keep working with it on a casual level, hoping that the situation will improve enough where I feel that I can contribute, but it's simply not there yet. Andy -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Markus Hitter Sent: Friday, November 04, 2005 12:47 AM To: Adrian Robert Cc: [email protected]; Lars Sonchocky-Helldorf Subject: Re: OpenOffice.org on OS X and GNUstep FWIW, I can pretty much support this opinion. If you listen to Apple mailing lists, there are many developers out there coding primarily on Windows and wanting a similar API on Mac OS X as well. They obviously feel more comfortable with Carbon as Carbon shares roots with the Windows API. Some of them even report they have no influence on the Windows code they have to port to the Mac; the quality of the Mac API is counted by the number of lines they need to get it running. Apple, with the advent of Tiger, marked all the QuickDraw functions (the classical Carbon graphics model) as deprecated. But if they ditch Carbon any time soon, they could easily loose half of all the applications available. Their answer to the situation seems to be CoreFoundation. A low level framework with mostly Cocoa in mind, but accessible through a plain C API. What does this mean for GNUstep? Well the hack-until-it-builds-and- debug-until-it-seems-to-work type of application development appears to be wide spread. A lot of developers firmly hold grip on their existing, huge code bases. They prefer low level hacks over revising or even rewriting logic. IMHO, GNUstep can't do much about this, but to continue making the development of well thought code design even easier and quicker. Some steps towards fewer diffs to Cocoa development surely wouldn't hurt. Again, only $0.02, Markus - - - - - - - - - - - - - - - - - - - Dipl. Ing. Markus Hitter http://www.jump-ing.de/ _______________________________________________ Discuss-gnustep mailing list [email protected] http://lists.gnu.org/mailman/listinfo/discuss-gnustep
smime.p7s
Description: S/MIME cryptographic signature
_______________________________________________ Discuss-gnustep mailing list [email protected] http://lists.gnu.org/mailman/listinfo/discuss-gnustep
