Why Wine is like playing the guitar

[Michael Robertson]

Ok, I've had some time to digest the two packed days of wineconf. It was 
great being around so many smart and passionate people. I've long since given 
up coding, so much of it went over my head but the various biz presentations 
hit home and I had a chance to talk to many of you to crystallize some 
thoughts. 

We need 20 million avid Linux desktoppers. Look at what Apple's 
vocal minority gets them. With 4% marketshare they command sections in 
stores. They get hardware with Mac/Win drivers. They get documentation in the 
box. Even stores dedicated just to them. That's what  we need to get Linux 
really going. If we can get the user inertia and the business inertia around 
Linux, then that will take it to new heights. I believe Wine will play a role 
in this. 

But I think a slight re-focusing may be necessary. Have you ever tried to 
learn how to play the guitar? I think most people have at one time. First you 
start learning chords and you say to yourself "I'm going to learn all the 
chords so I can play all the songs". After about struggling through 4 of the 
700 chords, you realize that it will take until 2037 to learn all the chords. 
To make guitar rewarding, you have to start with a list of popular songs and 
learn the chords just for those. This way you get that feeling of success and 
it motivates you to stick with it, even if it's just learning Happy Birthday. 
I believe a similar strategy could really energize Wine. 

We need a "top 10 tree". 

What WINE needs is a specific value promise to the consumer. In many ways, 
the same challenge confronts WINE that confronts Linux. Great things will 
happen *if* we get a bunch more users. But to get more users, there needs to 
be some specific value that people get from Wine. And it can't be fleeting 
value. It needs to allow them to do something specific and when they get the 
new version, it needs to no take away that value, but add more. 

The challenge we have now is that the goal for WINE is an architectual one 
(learn every chord). While nobody can deny the value of a solid underpinning, 
users don't care about the architecture. They want to hear music! And 
with Wine this means "what programs can I run?"

I would also suggest that focusing around specific applications focuses 
energies of developers on solving specific user problems. The more problems 
you solve, the more users will get excited about Wine. 

One question arose at wineconf which never got addressed. "How do we get more 
people excited about Wine?" I've been thinking about that. I believe a 
specific part of the answer is making Wine actually work for a subset of 
programs. Take it from a theoretical white paper stage to a stage where 
people actually use it. I'm not suggesting its at a white paper stage, but 
rather if the world can't use it with any regularity for even a narrow set of 
applications, the perception is it's nothing more than theory. 

What I believe needs to happen is that we have a 'top 10' tree. A version of 
Wine for which the primary goal is to do a good job of running a set version 
of win32 programs. This serves both parties. It gets developers all 
laser-focused on the same goal. Because we've narrowed the universe we can do 
specific testing. And perhaps MOST important, consumers know what to expect. 
If you tell them "we have a Linux OS which will run Win32 programs, they'll 
bring out a geneology program from 1992 which won't run and then they'll 
think it sucks. If however the public commitment is "this software is 
designed to run the following:
	MS Word
	Lotus Notes
	Quicken"
and it actually works, you have a believer. You have someone who will get 
excited and will offer their help. You give momentum to Wine. You get program 
sponsors (those willing to oversee ongoing supervision/testing of programs. 
You get more people offering to be on documentation teams. And you get more 
developers. 

One final thought. I think the only tenable way to implement a Top 10 type 
approach is to get widespread support. Lindows.com alone is too small to do 
this in a timely fashion. So is Codeweavers, transgaming, etc. The only 
approach with any chance of success is one which leverages the entire 
community. 

I can assure you that Lindows.com would put our full weight behind such an 
endeavor. While we only have a couple of coders, we've got a super QA 
department. We've got solid organizational skills. We have an enthusiastic 
user base we can tap for testing and support. 

I'll shutup now before I get banned from wine-devel. 

-- MR
Ok, those are my thoughts.