What would most aid WINE development?

Mike Hearn mike at plan99.net
Fri Nov 18 06:11:30 CST 2005


Susheel Daswani wrote:
> In terms of what a court could order, I think remedying the
> documentation and scope problems wouldn't be overly difficult.  

Actually I think it'd be very hard. It's not like Microsoft is just 
ULTRA-EVIL here, they often face the same problems we do with 
application developers relying on odd quirks of the way Windows works. 
Every release of Windows involves a massive Wine-style reverse 
engineering effort, on the apps themselves. Then Windows is sometimes 
changed to make apps that were broken by changes work again.

Now I'm not saying all is rosies and posies. Clearly, Microsoft go out 
of their way to screw us over at every possible opportunity. Dmitrys 
example of MFC is one, though I doubt it was intentional (Windows 
includes MFC for ages now because apps that ship with it use it). 
Another example is the way they class Internet Explorer as an operating 
system component, ditto for media player, etc etc.

Media player is a real problem also as many of the new online music 
stores rely on the Windows Media DRM platform, which would be *very* 
hard to implement in Wine for all kinds of technical and legal reasons.

> proven monopolist would have a hard time arguing against an order for
> complete documentation, even if a few trade secrets were revealed in
> the process.

Most of the time, the code itself is the best documentation, and it'd 
take a government with steel balls to force Microsoft to open source 
Windows. That'd almost guarantee its bankruptcy.

> Right
> now, it seems that WINE lacks support because WINE is chasing a moving
> target (and no one except MS knows what the target looks like)!

It lacks support because it's a REALLY REALLY HARD thing! Look at it 
like this.

In its history Wine has had about 700-800 people work on it in its 10+ 
year history. I forget the exact number. Yet, according to the last 
accurate figures we have, about 7000 people work on Windows full time!

In other words for every person who has ever worked on Wine, 10 people 
work on Windows every day. Now, some of them aren't writing new APIs. 
Some do design, others do usability, some do documentation etc .... but 
a lot are writing and implementing APIs. The numbers just don't work: 
how can we match that sheer force of manpower? A DLL like shlwapi that 
Microsoft implemented entirely within one team and one or two releases 
takes us years to reimplement.

The only reason Wine works at all is that Microsoft generally churn out 
new APIs faster than application developers can learn how to use them.

thanks -mike




More information about the wine-devel mailing list