Can someone tell me what this means?

I see. Thanks. I understand that 'Open' in this context does not mean open source, just an open specification. Apple, Microsoft and presumably the Linux volunteers had to spend a lot of time writing their own versions. I also understand that Microsoft's version is getting left behind and may be axed at any time. Maybe I am out of date on this though as we don't currently have a windows versions of our 3D products, only the 2D ones. Yes I should have known that, sorry. Speed is relative. If you have a million lines of code to keep up to date and sales revenues only allow you to employ yourself and one other programmer, then the speed of development is not sufficient to keep up with API, OS and processor changes AND add the features, which you need to stay competitive, to your programs. One of our programs, which was first issued soon after the first Macintosh shipped, has been rewritten from Pascal and assembler into C/C++, its basic drawing engine has been changed from using QuickDraw to using Quartz, the user interface has been partially rewritten to use Quartz/ Carbon and look like Aqua, but a lot more needs to be done. The printing side has been updated to use use the OS-X printing model rather than the OS1 to OS9 model. The clipboard handling has been changed many times and needs to be changed again to support PDF input rather than PICT vector format. Its been moved from 68K to PowerPC and now to Intel. Its been moved from development on Lisa to MPW on Mac, to CodeWarrior and now onto X-Code. Debugging has moved from 2 machine debugger to MacsBug to CodeWarrior's debugger to GDB under X- Code, and not always for the better. GDB under X-Code is still not as good as CodeWarrior, it fails to show some variables, cannot display dynamic arrays except as a hex memory dump and cannot single shot over some C++ constructors. Another program has been updated from using QuickDraw3D to using Quesa, from using PowerPC to Intel (with lots of endian problems), Its been partially re-coded from C++ to Objective C. We've only re- issued two thirds of the program, dropping the less useful features until (if ever) we get the time to do the more exotic parts. Its been changed from the classic event handling system to Carbon events. Its been changed from using a single processor to using multiple processors in critical parts. From using resources and Resourcerer to using NIB files with Interface Builder (despite that, in my opinion, Interface Builder is still incomplete and that MacOS is still not as well documented as OS9). I could go on, and that's just two programs of all the ones we ship. It been moving TO GCC that is one of the problems. It does not allow temporary variables which need to be constructed, to then be passed as procedure parameters. Debugging of anything using templates is very confusing (presumably because the compiler does not output the correct source code references), displaying source code which has nothing at all to what its actually executing so that you have to look at the object code, which is Intel code which is another learning curve I need to master to add to my repertoire of a dozen or so machine codes. Any recommendations of a good book on the latest Intel processor instruction set? GCC does not even generate very efficient object code when optimisations are all turned on. Rather difficult when the libraries are in PowerPC code and you're running on Intel, and when the libraries make OS calls which are no longer available.