Read an interesting article about NASA. that basically argues the agency was an order of magnitude more productive during its Apollo-phase than during the current low-earth-orbit Shuttle-phase. The author, who admitedly has an axe to grind, compares NASA's output from 1963-1971 with its output from 1990-now. The Shuttle missions, he claims, are basically done with the argument that "we could use them for something, anything, in the future". The conclusion? When NASA was actually working toward concrete (lofty) goals, a lot more got done. I think its a very general phenomenuum in engineering (and maybe other areas too?). Both in a micro (single person across a day or two) and a macro (hundreds of people across months or years) scale... to achieve great things you need to have great goals that you really buy into.

The article struck a chord with me because it reminded me of GNOME. GNOME exists in a state of perpetual shuttle-missions. It has no goal, no vision. I don't wake up in the morning dying to hack on GNOME (sometimes when one of my projects is doing something especially cool and is almost working I feel that way, but never really about GNOME itself).

On a "micro scale": Ever noticed how a few (or one) motivated people can do really amazing "feats of programming"? I think a lot of times this gets chalked up to either the individual's brilliance or the advantages of small groups of people in terms of communication (think mythical man month material here)... Another important factor, maybe even more important, is that these people are driven by a capital-v Vision. They know what they want to create, they are not intimidated by the steps they need to get there, they dive in and just DO IT. As a developer, how much more productive are you when things are just clicking compared to an average day? For me, the difference is a solid order of magnitude, and it happens when I've got a vision I'm working toward and I've overcome my fear of the hurdles to get there.

On a "macro scale": Were it not for the presence of commercial software to copy I think GNOME (and the free software desktop in general) would be totally stagnant. The acceptance of copying commercial software provides a design goal that people generally agree upon. I (occasionally ;-) hear GNOME people putting forward really solid ideas. They receive no backing until, hey what do you know, a year later Apple or Microsoft comes out with them. Suddenly everybody is in favour of incorporating the idea. In that sense, commercial software provides the closest thing to a vision that we have (of course, this guarantees that we will be perpetually at least a release behind, and because we haven't developed expertise in really understanding the issue we will probably do a poor clone).

I don't know why this happens exactly. For one thing, beauracracy and the like plague any collection of more than a dozen people (sometimes even then)... but that can be overcome. I think its deeper than that. I think many free software developers might feel inadequate: they may not think they can produce things or think of things that are as good as the "big boys".

Maybe we are just really risk averse: unwilling to try anything unless we are sure its a safe bet. We have more room to be betting than companies! If anything we should be less risk averse than they: our users are probably more tolerant of interesting ideas that turn out to be failures (they can work around them and turn them off), we have a smaller user base, we aren't liable to third quarter financials, and GNOME 1.x proved we can suck a lot compared to the "competition" and it'll take a long time before that starts showing in "the numbers".

Maybe we were burned by the long development cycle between 1.4 and 2.0 and aren't willing to try anything "big" again. The problem with 1.4 to 2.0 wasn't the time it took, it was that we didn't have anything to show after that period. That was a perfect example of shuttle-mission mentality. I mean, things were improved, sure (just like space tech keeps creeping forward during 1990-2003 for NASA), but not commensurate with the work and time that went into them. If we had produced something markedly better, really cool, etc people would have quickly forgiven the time it took between releases. If we had focused on an idea we could have done it.

I mean, I'm not saying we should ditch quick releases, I think they're a good thing, but we MUST be planning much farther than the next release if our releases are going to be 6 mo apart. Right now the release team has become the defacto "forward thinking" power in GNOME by virtue of the board being explicitly not-involved with technical issues and the release team being the only other organized body in GNOME. Release team is doing a good job planning 6 mo releases. But we need more than that, and I don't think the release team would do a good job at it. They are focused (by mandate and by personality) toward worrying about more immediate and pressing issues, and I think people in such a position naturally sacrifice "non urgent" issues (that are exceedingly important, but in the nebulous and distant future) for "urgent" issues (where you can see the consequence of not doing something TODAY). Its good to have an RT that's dead-set focused on what needs to be in the release several months from now.

How many "man years" have gone into GNOME? How much better could GNOME be today if all that work had been toward a cool, coherent vision?

We also need a vision for GNOME that's farther out. And it can't just be something that some powerful cabal writes down on paper, it has to be the life blood of day to day GNOME programming. It has to motivate our technical decisions, clarify our priorities, and goals, focus our interface improvements, help us decide what "applications" and technologies to develop next, etc. It has to seep into every pore of the GNOME body and take over the organism. On the other hand, it can't be designed by committee or consenus, because then it would inevitably suck. What we (aka I) want is a vision infused with creative ideas from everyone, bounded (but not too bounded) by the dire warnings of failure pronounced by experts on certain ideas, and drafted by a small cabal of fascists... a vision that is so compelling that we all (or next to all) buy into it. And that's the tricky part. A vision that will make people drop their disputes and differing goals (I do not think diversity of goals makes GNOME stronger, diversity of talents toward the same goal is what works). Does such a thing exist? I think so, I wish so, I hope so. Can we find it, will we even look for it? Probably not.

Oh well.