Re: WineX and the AFPL

On 30 May 2002, Alexandre Julliard wrote:

> That would be true if Transgaming was the only company out there. But
> since it's not, we will actually win more by having as much code under
> the LGPL as possible, since this will maximize the amount of code we
> receive from all other present and future companies, as opposed to
> getting just a single patch today. In the long run the LGPL will do a
> lot more good to Wine than Transgaming's code; but this will only be
> true if we can reach a critical mass of code under the LGPL. And any
> piece of code that is relicensed under X11 reduces the chance of
> reaching that critical mass.

I don't think it's as clearcut as that.  On the one hand, the LGPL demands 
that companies who use and enhance the code must release their changes back 
to the community.  On the surface, this sounds great.  However, there's a 
difficult aspect to quantify -- how much development work never happens at 
all due to the license?  The GPL or LGPL can demand all it wants, but 100% 
of nothing is still nothing.  Giving back only matters if there's something 
to be given back.

The X11 license doesn't demand that companies give back, but gives them an 
incentive to use and improve the code.  The LGPL license demands more from 
the companies, but reduces their incentive in the first place.  In the end, 
which approach results in more commercial contributions to public projects?
I don't know, but I don't think either side is an obvious winner here.

Look at the X11 project itself.  It started out as a free project, and it 
was quickly balkanized by many vendors, each plying their proprietary X11 
implementation, based heavily on the free code.  On the surface, this looks 
rather bad -- a free project led to much development work that was kept 
proprietary, locked away where the community didn't benefit.  Was this a 
failure of the X11 licensing approach?

There's a strong argument to be made that X11 is so dominant now precisely 
BECAUSE of this ability to proprietize the free software codebase and add 
proprietary enhancements to compete in the marketplace.  Where would the 
X11 project be if it used the GPL or LGPL, and thereby failed to become an 
industry standard?  (After all, most of those vendors wouldn't have adopted 
X11 if they couldn't sell their proprietary version starting from a free 
project where 90% of their work was already done!)

Ultimately, the XFree86 project became the standard-bearer for the free X11 
project, and many vendors seem to have abandoned their proprietary X11 code 
for XFree86, because (of course) the community-driven, highly collaborative 
XFree86 project had left most proprietary X11 implementations in the dust, 
despite continuing under the X11 license.  Sure, they could have changed to 
using the LGPL, but it seems that it wasn't actually necessary.

Ultimately, was the X11 license so bad after all?  Even if some companies 
are greedy and don't give back, many of the companies who can benefit from 
X11 licensing might not be in the business under another license.  Which is 
better?  Lots of development with some return, or a little with 100% back?

> But if Transgaming was willing to commit to making future versions
> free then they could just as well adopt the LGPL code without
> requiring any further negotiation. So the fact that they want these
> negotiations shows that they have already rejected the option that
> could be the basis for what I would consider a fair trade.

I was suggesting that they might be willing to commit to making PART of 
their code open in future releases.  The LGPL would force them to make ALL 
of their code available, which they're evidently unwilling or unable to do.

Obviously contributing all of their code is a fair option -- does that make 
a partial contribution necessarily unfair?

> I'm not refusing at all, I have repeatedly said that I'm willing to do
> all necessary efforts to make things work for Transgaming if they are
> willing to change to a more open source model. But I have no interest
> in negotiating on the patch-by-patch basis that Gav has proposed,
> because it does absolutely nothing to address the very serious
> problems I see with Transgaming's model.

I don't understand why you'd worry about this.  What does it matter if his 
business model is flawed?  Transgaming will ultimately live or die based on 
their business model, good or bad.  If you think Transgaming's business 
model is bad, then by all means, don't invest in their company!  If it's 
not your money on the line, why worry about it?  Even if you think an open 
source business model would be a better choice, those who have their money 
on the line should make the decision of whether or not to switch models.

Whether they're right or wrong, they should be allowed to succeed or fail 
on the merits of the business model they choose to use.  Suppose (for the 
sake of argument) that the open-source business model you recommend to them 
won't work for them (for whatever reason), and would cause their business 
to fail.  Would you volunteer to take the blame for the failure?

Whether their business model is good or bad, it's probably a good thing 
that there are people out there trying different business models -- some 
models that sound good on paper don't necessarily work out in practice, 
while others that sound dubious may turn out to be successful.  The only 
way to know is for someone to try them -- and it may as well be tested by 
the true believers, not the doubters.  If they doubt the open-source model, 
they might fail at it -- and if you doubt their model, you might want to 
avoid it too.  To each his own.  Diversity isn't such a bad thing...

> And despite the fact that I have been saying this to him for a long time 
> now, Gav has shown no intention at all to try to address this issue.

If you can't convince him of the error of his ways, then it's probably best 
to agree to disagree.  You continue to endorse your way, he can stick with 
his, and what happens, happens.

Just because he steadfastly refuses to come around to your way of thinking, 
is that any reason why cooperation should be impossible?

> That's why I prefer to spend my time writing code, instead of wasting 
> even more time trying to make Gav give us something that I know he 
> doesn't want to give, no matter how much we could offer in exchange.

It doesn't sound like he's willing to give up on his business model any 
more than you're willing to give up on the LGPL.  At least trading patches 
would give you both an opportunity to benefit without turning your backs on 
your goals and strategy...

> My worst fears are already coming to pass, which is the whole point
> about the change. DirectX development has been dead for more than a
> year now, and all the enthusiasm there was for running games with Wine
> a couple of years ago has evaporated. That's a very serious harm, and
> while it's too late to do anything about DirectX I don't want that to
> happen to other parts of the code.

To me, it seems like there's been an OVER-emphasis on games in the last few 
years, to the detriment of the project as a whole.  There's been such an 
emphasis on games that it seems like little else gets attention.  How come 
basic Windows IPC mechanisms are still unsupported (or so I've been told) 
while effort is being spent on gaming?  It's possible that the dearth of 
DirectX development could be PARTLY a natural correction as well.

Transgaming's purpose in life seems to be to serve the interests of the 
portion of the gaming community that wants to run Windows games on Linux.  
This is a fine goal, but it seems to have absorbed so much attention over 
the last few years that the goal of having a general-purpose replacement 
for Windows has been sidelined somewhat.  It might not be an entirely bad 
thing if Transgaming siphons off some of the gaming enthusiasts from Wine 
development, if it leads to a better focus on the original goal.

Is the real goal of Wine to replace Windows, or JUST to run games?  Think 
about it -- if Wine gets to a point where everything BUT games works, and 
works flawlessly, then the attention would fall on gaming again to finish 
the job anyhow.  And if everything ELSE was flawless, we might be seeing a 
lot more desktops converted to Linux in the first place, because everything 
would work just as well as in Windows, except maybe games.

Let's be realistic here.  Gamers are an obsessive breed.  They'll use the 
platform that meets their needs, convenient or not.  They'll dual-boot with 
Windows if they have to.  They'll use bleeding-edge drivers.  They'll buy 
cutting-edge hardware.  And yes, they'll even pay Transgaming for their 
little proprietary toys, if it improves their gaming experience.

So leave the gamers out of the picture.  No matter how much effort they 
might put into improving Wine's DirectX support without Transgaming mucking 
up the works, such support is primarily of value to gamers anyhow.  Apart 
from gamers, what does the rest of the world want?  They just want their 
applications to run flawlessly and be rock solid.

Right now, the only real option for most users is to have a real Windows 
partition, since Wine promises nothing at all about what will run.  It's a
"developers-only" release, as it has been for many years now.  The light at 
the end of the tunnel has seemed to be just a year or two away almost all 
along, or maybe it's been a mirage this whole time?

If you really want to see Wine and Linux take off, forget about the gamers 
and DirectX support and worry about everything ELSE.  That's what really 
matters.  The gamers will take care of their own.  Worry about Joe Sixpack.
When Wine works so well that Joe Sixpack doesn't see any point in having a 
copy of Windows around, then you'll have won.

And guess what?  When Joe Sixpack starts using Linux and Wine instead of 
Windows, and the desktop marketplace moves to Linux because it works better 
and crashes less often than Windows, it comes full circle.  At that point, 
the Linux market would become more compelling than the Windows market, and 
the great new games would start being targeted for Linux in addition to (or 
instead of) Windows, and at that point it won't really matter how good the 
game support in Wine is anyhow.  (Most gamers tend to focus on recent games 
anyhow, so even supporting old games will be less than critical.)

Looking at the big picture, I don't see harm to the Wine PROJECT.  Maybe 
DirectX development has fizzled (for now) due to Transgaming.  So what?  
There's plenty of other important work to be done, isn't there?  Focus on 
that and let Transgaming do their thing.  And trade when it's advantageous.

> Go read Dimitrie's analogy about children. Would you let someone hurt
> your children even if they gave you a lot of money for that?

Of course not.  But I think you're jumping to conclusions to say that it's 
harmful when that hardly seems a foregone conclusion at all.  Obviously, if 
it's true, there's no amount of money that can justify it.  But is it true?

Deven