I've spent 50 minutes watching Chris Messina's Thoughts on Mozilla video. There is a lot in there. Some of it is sensible thinking, some of it is totally off-base, and the rest is in-between. For additional thinking, please read Anne Zelenka's response.
I don't have a Webcam, plus I'm more of a writing person anyway, so I'll respond in writing.
I know Chris for quite a long time. I think we first met in Brussels, at Fosdem, maybe 3 years ago, and we recently had dinner in London. I've been reading his blog on and off for quite a while. I've followed his adventures at Flock, and I'm a BarCamp fan. To be clear, I like the man and what he has achieved so far!
I wanted to send Chris an email, but I'll do it publicly, in the spirit of his message. Please note that this message is posted on my personal blog, as an individual, and not as someone involved in Mozilla. To make things very clear, please note that this is not an official Mozilla response in any way. So please, if you want to quote me, please write "Tristan Nitot, Mozilla contributor".
Let's get started with what I agree on with Chris. I'll make this part short, as it's pretty obvious that Chris and I share many values.
What Chris and I agree on:
- The Web and the Internet are wonderful things
- Choice and innovation is important on the Web
- Open Standards are what make the Web and the Internet so interesting, universal and powerful.
For the sake of keeping this post short, I won't list the topics on which Chris is uninformed in my opinion. (Plus it could very well be because Mozilla does not communicate enough on it). But for the record, on being Open, on having events/unconferences (Tokyo and Paris in June, baby!), and Campus Reps program, I'm confident that we're doing reasonably well.
I would like to focus on a couple of key topics where discussion may become interesting and fruitful.
In my very own opinion, Mobile is a difficult market for Mozilla to play well, as its driving forces, mobile operators, are not willing to open it. Because they want to sell Services that they can charge a premium, they don't want to sell plain bandwidth, as it leads to a war on price, which is not good for their shareholders. They control the mobile phone manufacturers, decide which applications are going to be on phones. In such an environment, Mozilla is not well equipped and cannot do very well. On the contrary, Opera is doing wonders, as they have pre-sales teams and professional services teams to taylor their proprietary codebase to the needs of phone vendors, influenced by operators. Let's note that the market is evolving, so things may be very different in a few months from now: Opera Mini is a very interesting technology, doing a lot to open up the real Web to mobile phones. In the meantime, Mozilla is experimenting with Project Joey, which approach I find quite innovative.
Mozilla as a platform
"Mozilla should become a platform", says Chris. I would argue that it's already a platform. Twice! It's a platform for Web applications. But it's also a platform for extension developers. Being a platform vendor is nice, but this means a lot of commitment in terms of API freezes, as otherwise you piss off all your users and application vendors. And committing to API freezes slows development down, which Mozilla cannot afford. I'll explain why below.
IE7 has caught up with Firefox, and browsers are dead...
... at least that's what Chris is saying. I do not agree, but I do recognize that IE7 has implemented things that may make it look "good enough" for many users, namely pop-up blocking, better security, tabbed browsing and search integration.
So what's left for Mozilla? Give up on Firefox? Give up on innovation? Leave the flawed bookmarking system as is? Leave Microformats on their own? Leave OpenId in the dust? I don't think it's a good idea. There is a lot to be done in the browser, in my not so humble opinion. But it's not easy to do. Tons of ideas sound cool but end up becoming feature-creep. There are a few very interesting things, and they need to be implemented very carefully. Easy to discover, to understand and to use, without implying security issues. Figuring this out is incredibly hard. Look at Flock. Tons of money, smart people, a proven, extensible and reliable codebase (Firefox), and still no product. Last time I checked they had to ship 1.0 last December. 6 months later, nothing. I'm not saying that the Flock team is bad, not at all. I see this as a proof that building innovative and useful features in a browser is very very difficult. Mozilla is very successful in launching a new major version of Firefox every 12 months or so. But let's face it: it is hard. Really hard. Do we want to slow down our development process in order to become a platform on a third level (on top of the Web and Extensions)? One of the biggest advantages that Mozilla has over Microsoft, it's that we're moving faster. Giving up on our ability to release often (which is something very important, considering our Open Source / Free Software nature) in order to become a platform is a very tough choice...
"Targeting Joe Sixpack"
While I'm not a big fan of using this expression, I stand by the fact that market share is super important for the Mozilla project. The reason why is that it's the only way to keep the Web open. I've been directly and professionally involved with browsers since May 1997 I've been involved with Web standards since mid-2001. I've spent time finding Web sites that did not render correctly in Gecko starting in 2001. At some point, we had a team in 4 countries testing the top 5,000 Web sites in France, Germany, UK, US and Japan. Broken Web sites were approached and given a fix. You just can't imagine how much effort and energy has been put by our team to "fix the Web", because so many Web sites where "
optimised for limited to Internet Explorer". And too few Web sites where willing to apply the patches we sent them for free. Then Firefox came, it gained critical mass, and now there are way less websites that appear to be broken in Gecko. Conclusion: Significant market share of different yet standards-compliant user-agents is the only way to ensure the health of the Web. Firefox has roughly 15% market share worldwide. It's not enough, as Vista is eventually going to take over the desktop market. We need to make sure that Firefox is still innovative to entice users to download Firefox.
The future of the Web, RIAs, Silverlight and Flex
This is the 1 million dollar question: who is going to win the battle of the next-generation Web? Is it Flash-based technologies pushed by Adobe, and the promise that Adobe will open up the source code? Is it Silverlight from Microsoft? Is it going to be Ajax-based, with added sweetness from HTML 5? I'd rather see the latter happen, as HTML 5 is documented, open, can be used without getting into ugly patent issues, and HTML 5 is an evolution of the existing Web. This means this would be a broader continuum encompassing static documents, Ajax applications (2007-era) and better, HTML 5-enabled applications. The other possibility is that we throw XUL in the mix, but what's the point? I'm thinking aloud here, so forgive me, but if Adobe open-sources Flex, we'll have an open-source, cross-OS platform to build Rich Internet Applications, which is still better than Silverlight on paper, as Microsoft's solution will not run on Linux or other platforms.
This is a very interesting subject, and I have to admit that I have not yet fully wrapped my mind around it (I'm working on it). It will be the subject of several posts in the near future, I think. In the meantime, I suggest that those of my reader who are interested in this subject read recent posts from AllPeer's blog, Peer Pressure, including this one and that one. Oh, and don't forget to read Mike Shaver's thoughts on the subject: The high cost of some free tools.