I want to run URL Snooper and that requires WinPcap. On Windows, WinPcap requires a kernel driver. So I would need a guide for installing a kernel driver into Crossover. I also need to run a program which directly accesses a USB device, but again there is no device driver support. I don't mind installing device drivers manually if that were possible... but I don't think it is, and thats a serious limitation of the product. Crossover is a noble effort, but nothing I have tried it with runs properly even when the application does not require access to device drivers. I'm willing to do the work to make things run but there is no clear and concise troubleshooting guide.

Right now this system is not very practical. According to the web site, CodeWeavers' mission is to "transform Mac OS X and Linux into Windows-compatible operating systems" ...but there is still a long way to go. I think it's time for CodeWeavers to reassess its long term development strategy and its priorities. And just as Microsoft is making some effort to support Linux*, it would be right and proper if Apple contributed something to the development of Wine and Crossover too. There are all kinds of USB gadgets like engine analyzers, television sets and blood pressure monitors that only come with Windows software. These things will probably never be supported on Mac or Linux by the manufacturer until they know how to design and package Windows applications that can be made to run in Wine or Crossover. It seems like TransGaming is setting an example here that CodeWeavers might want to emulate. Aside from the lack of device driver support, I think there is a serious lack of good documentation for both the developers and the end users. The result is that very few applications can be made to run on Crossover even when it is theoretically possible. That's my opinion, anyhow.

*Microsoft supports Linux

http://blogs.technet.com/mattmcspirit/archive/2009/07/20/microsoft-releases-linux-integration-components-linuxic-under-the-gplv2.aspx

Thanks for the post.

We'll be the first people to say that Crossover/Wine have "a long way to go". Wine is extremely hard technology to work with and develop - it's very slow going (moreso than with most any type of programming), and there's a pretty small pool of developers who have the knowledge and desire necessary to work with it. Of that pool, there's a much smaller segment who want to do it full-time - and this is where the Codeweavers team comes from.

Device drivers have long been and will likely continue to be a major hurdle for wine/Crossover. Kernel drivers of course hook directly into the kernel, rather than calling upon functions from several layers "up" in the OS structure. From an application's standpoint, Crossover/Wine looks just like Windows, but at the kernel level it is completely unrecognizable as windows. Thus, for cases in which a linux- or mac- native driver does not exist (which are many in the case of proprietary drivers), a lot of work needs to be done to get the driver working in Crossover, so much so that it is most often not cost-effective for Codeweavers to pursue it unless the company which produces the application makes an investment in the process.

Since you're spent some time looking through our documentation, I presume you've seen this article -

http://www.codeweavers.com/support/wiki/faq/serialoverusb

The success rate with USB devices in Crossover is admittedly quite low, but it has been known to happen with some serial devices.

(For more info, please see this page about device drivers in wine from the wine wiki - http://wiki.jswindle.com/index.php/Drivers)

Lastly, your point about the documentation is well-put. We're currently in something of a half-way point between setting up a comprehensive online wiki and phasing out our FAQ's/documentation section. We do add new troubleshooting articles to our wiki quite regularly for specific issues as they are diagnosed, but our "general troubleshooting" guides have grown a bit stale and could, I think, use a lot more detail.

Thanks for the thoughtful responses, and the reference links. If nothing else, that gives the public a clearer picture of the state of this project and clarifies what is necessary for people to achieve the kind of results they are looking for. I can see now that this is a labor of love, and you guys have not been slacking off. However, the slick user interface and the presentation of the product on the web site gave me the impression that I could expect a much higher degree of compatibility. I guess it was hard to contain my disappointment when nothing I wanted to run worked properly using the automated installation routine. In one case, the bug appeared to be nothing more than the failure to detect a mounted CD-ROM and it seems that sort of thing would have been addressed a long time ago. Now that I know how to generate a log file, I will take a second look. I'm sure you can appreciate how difficult it is for the end user who is trying to determine if Crossover lacks support for a specific function, or if the problem could be solved by some kind of manual configuration.