On Sat, 2 Nov 2002, Andreas Mohr wrote: [...]
The only advantage I can think of would be the use of Unix code, but then you could achieve the very same thing by writing the Unix equivalent DLL of a Windows DLL (e.g.: hardware access or so) and using the Unix specific DLL version instead on Wine.
That's one of the advantages and it allows you to do things you would probably not be able to do by just loading a Winelib dll from a non Winelib application. One such thing is mixing Win32 windows with QT ones in a single application. This is not a pipedream, it has already been been done on a large application at least once.
The other advantage of course is that Winelib makes it possible, at least in theory, to port your application to non x86 platforms like the Sparc or PPC. In fact I think the path to removing the x86 monopoly is: 1. run existing Windows applications using Wine 2. Linux grows market share 3. Wine improves enough on non-Linux x86 Unix platforms (FreeBSD, OpenBSD, Solaris) that you can run Windows apps on them as well as on Linux 4. now that Linux has a significant market share on the desktop ISVs start taking Wine into account when designing applications 5. ISVs use Winelib to migrate off the Win32 API to more cross-platform APIs 6. ISVs use Winelib to port their applications to non x86 platforms or port them directly if they already moved off Win32 using Winelib as a transition
Not saying this will happen, there are too many factors involved: Java, .Net, continued survival of non-x86 processors, etc. But Winelib might play a significant role in the future.