Nice, but it woun't completely replace a Win system. Some base features will be completely missing and some apps wount be able to run. It looks an alternative to a virtualization software like Parallels but it isn't. Depends on what you need.
CrossOver Office had been released for Linux for quite some time. It's not been given much interest in the Mac crowd until now because 1) it needs to run on the same processor architecture as Windows apps, and Macs only went Intel this year, and 2) the product primarily aims to make it possible for Microsoft Office and probably Photoshop on Linux, both of which are already available on Macs as OS X apps. It's basically a product based on the famed WINE project, and if the quirks get ironed out it would probably be the most ideal way of running Windows apps on the Mac - i.e. without Windows. I read it would even support DirectX / Direct3D, making it possible to run games like Half Life 2 or something.
I need true windows for most of my stuff but it would be great for me to be able to run MS Outlook (which I hate but need for blackberry integration) without having to run it in Parallels which I try to reserve and keep clean for development.
If CrossOver / WINE can achieve the level of seamless application running that OS/2 did with Windows apps (Win16 only, though) on OS X or Linux, then at that point, there really wouldn't be a need for 'true Windows' at all. Sadly, I don't think this will happen for at least a couple of years. Another interesting approach to the problem is the ReactOS project - making a Windows-compatible platform on the OS level, but open-sourced. Basically, if it'll do what it's supposed to do, we'll have a free-as-in-beer OS that can run Windows apps. This, too, shows promise, but it's seriously buggy right now.
Wine = Wine Is Not an Emulator. Wine is an abstraction layer. It allows an application to run natively by mapping windows system calls to Linux or, in this case, Mac OS X system calls. It's like Rosetta for Windows apps without the emulation.
Indeed, Darwine for PPC would involve PPC-x86 emulation layer, but Darwine for Intel wouldn't need such emulation and therefore would involve straight call remapping. It's actually better than virtualization in that you don't need a separate OS that needs to be maintaining itself, if it fully works as intended.
I'm currently testing the 2nd alpha version. It does support as many programs as the linux version does, but the programs that do work run quite well. The only other downside is that the programs run under X11 instead of cocoa. The final release should be quite good, it's already one of the most stable alpha releases i've seen.
also, no games are working. directx support is way behind for crossover mac. this may not be supported for a while.
It's not a bit like Rosetta, given that the key component of Rosetta IS in fact the conversion of PowerPC opcodes into Intel opcodes. WINE is a reimplementation of the Windows API, i.e, the functionality provided by the system DLLs that make up the Windows "OS". I bought the commercial version of WINE (crossover) for my Linux box but I was never really happy with it, there was always something that didn't quite work. D
I read somewhere that they have World of Warcraft running well in Crossover Office. I don't know if WoW uses DirectX or OpenGL (probably a beta version). I assume they are mapping directx to OpenGL on Linux. If that's the case, it shouldn't be a problem getting the same functionality on MacOS X.
There are two components to Rosetta. One is the PPC to x86 opcode translation. What I think is a more significant part of Rosetta is that it maps system/library calls to the intel native version. I tried an earlier version of crossover office. The fonts sucked and the window borders looked like windows 3.1.
I don't agree at all. That mapping process is pretty trivial given that the API is identical under both environments - and there's hardly anything to do since the underlying libraries are already there.. In the WINE environment, they actually have to IMPLEMENT the Windows API, and without any access to the Windows source code - that's a much harder thing to do. The ability of Rosetta to take Power PC opcodes and convert those to Intel opcodes and do it in such a way that even some real-time programs (such as digital audio workstations, software synthesizers and so forth) still work is a far more impressive achievement. David