Thursday, June 02, 2005

Profiles in Horrible Leadership Skills, with Miguel de Icaza 

Shawn pointed out to me last night that I really have it in for Miguel de Icaza. I just can't help it. He's just such a horrible leader, and I figured I'd write about it, since, you know, it's not like any Gnome people are gonna write about it and let the people who couldn't watch the speeches know what was being said and decided about the future of Gnome.

I watched his keynote GUADEC speech last night, and it was unbelievable. His speaking style is best described as "unrehearsed." Most of the speech seemed like a long ramble based on a few slides, which he made up as he went along. There was no vision for the future of Gnome presented, aside from some vague statements about the need for better usability testing.

Other great moments in leadership:

For contrast, look at Jeff Waugh's "10x10" talk, which at least judging by Planet Gnome, has gotten developers excited. Waugh was the person at the conference who articulated a vision for Gnome. His vision was to have Gnome achieve 10% global desktop marketshare by 2010. He probably knows as well as everyone else that that won't happen.

Still, his talk was a creative, well-spoken list of suggestions for how to get there: why aren't we taking full advantage of our "friends in high places," like Google, Novell, Red Hat, HP, Sun, and the trade press? Why aren't we engaging hardware vendors to see what it would take for them to ship Gnome pre-installed on their computers (in certain configuratios), so that we can work towards that? Why aren't we engaging independent software vendors to make them feel more like their applications are a part of Gnome even if they don't ship in the Gnome distribution? Why aren't we usability-testing our APIs instead of just our user interfaces? Why do we discourage software from Gnome being ported and distributed on Windows, which helps users move to our platform incrementally, and gets them excited about it? Why do we insist on shipping software that we think is abstractly elegant, but our end-users overwhelmingly tell us they hate? Why don't we see keeping our language bindings up to date as part of our job?

All of those suggestions would greatly improve the Gnome project, whether or not they meet their 10x10 goals. As Thoreau said, "In the long run, men hit only what they aim at. Therefore, they had better aim at something high." That is the role of a leader, telling them what to aim at, and how to hit it.

The OSS world seems to be good at (innovatively) implementing known specs. Write a Unix kernel? No problem, give us 5 years and it'll be great. Write command line utilities? Give us some more time, and we'll make them really efficient too. Write developer tools? Heck, we'll improve upon them, even!

But when it comes to "Write a desktop application" they need industry big time. Gimp is an imperfect subsitute for Photoshop (though, it's in a different niche now anyway). And OpenOffice had tons of existing support. To actually think these guys could create their own Unix desktop is silly. Maybe now they can just copy Apple, because they certainly failed at making Unix feel "just like Windows."

The reality is that you need people working full time on projects that might reach completion, but still never get used. NeXT and Apple can do that. But to coordinate a bunch of people who've never met, and to meet the same goal requires either: (1) a great leader, with a trusted vision that can work; or (2) a very well-known goal (i.e., just copy what someone else already did).

But Mozilla seemed like a well-know goal too, so why did that take so damn long?
Oh man, that video is classic. The first five minutes is him trying to get the screen to display, and in the end he can't do it the way he wants. That about sums up the open source world for you, right there. So aware of Linux's strengths, so blind of its many failures.
You've made some interesting comments, but I don't agree with them completely. I think when you say "open source," what you really mean is "non-commercial," or something like that, and I'd agree with your point if you were talking about the majority of unfunded, volunteer-based OSS projects.

Gnome is open source, but it's largely hacked on by people paid to hack on it, from various companies. Other projects, like Linux, Apache, and GCC are also driven by a constellation of companies paying programmers to write the code. In fact, I'd say that OSS quality and corporate sponsorship are highly correlated.

Looking at what some of the Gnome companies are doing, they ARE doing usability testing, and they ARE hiring interface designers. Looking at the recent Gnome 2.10 release, I think it has actually surpassed Windows in usability, for the most part. It just doesn't have any software for it, or an OS that widely supports hardware in a hassle-free way.

The Gnome guys can't control that, and they know this, and they're not trying to be that euphoric solution. They're trying to find people who can work within those constraints, like certain companies and agencies who might be a good fit for Gnome and Linux. ILM, Pixar, and Dreamworks are pretty much run entirely on Linux and Gnome, for example.

It's not exciting, but it's the way to get things done. Find the potential users who are most likely to buy/use your product, get them to buy/use it (through improvements and/or salesmanship), repeat. It's slow, but that's better than aimless wandering through mailing lists.

I've actually posted before about whether or not open source can be innovative: http://mindpoison.blogspot.com/2005/02/some-thoughts-on-linux.html

As for your last paragraph, I actually think you have to look at Mozilla's development process a bit closer. The Mozilla that actually shipped in the end was not actually what they'd been working on since the code was open-sourced. They actually rewrote the code from scratch, which was a terrible idea, and many others (Joel on Software) have talked about this mistake in great detail. Ditching all the code wasted the first few years of the Mozilla project, so what actually shipped was written in a few years less time than most people think, they just don't know what happened in-between. Of course, I'm not saying this to excuse the Mozilla leadership at all, I'm simply saying that bad software management probably accounts for most of the big, notorious open source failures, not something intrinsic in the nature of open source.

Finally, Macneil: I agree with you. Linux has got to get more pragmatic and just accept that they're going to have to freaking have freaking closed-source drivers, because manufacturers just won't allow anything else. I don't think it's a good idea for open source to waste effort trying to produce their own hardware, so they just get over it and realize how easily hardware is swapped out anyways.

you really have a very nice blog here! I'm definitely going to bookmark you!

I have a site dealing with holland seo. You are heartly invited to take some really good tips from it.

Come and check it out if you get time :-)

Very informative blog. Please check out my distribution linux blog.
Post a Comment

This page is powered by Blogger. Isn't yours?