Mitchell Baker has a good post clarifying the decision to look for a new home for Thunderbird. Emotions are running high on this issue, so I should make it clear that I was in no way disagreeing with the need to preserve focus in my previous post on this topic. In fact, what I took issue with was the conflation of Firefox and the Mozilla platform into a single organization. To my mind, this is at the heart of the confusion that seems to arise whenever the discussion turns to the topic of how to handle platform issues and non-Firefox products that are based on Mozilla.
I’ve thought about this a lot in the past 24 hours, and I still think that separating Firefox into a totally distinct commercial organization would be the best solution. I guess this is unlikely to happen, for historical and ideological reasons, but it’s an alternative that is at least worthy of consideration.
There’s been a fair amount of controversy regarding Mitchell Baker’s recent announcement that Mozilla is seeking a new home for Thunderbird. From my perspective, the issue stems in large part from a growing disconnect between Mozilla’s stated goals, as laid forth in the Manifesto, and its increasingly exclusive focus on Firefox. This isn’t to say that the latter is necessarily a bad thing, but the decision to saw off Thunderbird doesn’t seem to sit well with the open-ended objective to “build and deliver great consumer products that support the Manifesto’s principles.” If an open source email client based on the Mozilla platform with 5-10 million users (including me) doesn’t make the cut, what does?
Another concern that I have, from my perspective as a probably ill-informed outsider, is that Firefox itself is the product of a grassroots effort from inside the Mozilla community, not a top-down dictate. It would seem apt for Mozilla to retain a structure that is conducive to fostering the Next Big Thing, rather than merely nurturing the current hit product.
The final consideration is the ongoing platform debate. It’s no secret that I and some others think that Mozilla may be missing a truly forward-looking and visionary direction by not investing more into the platform as a competitor in the RIA space.
Mozilla has taken on an incredibly difficult and ambitious task in trying to innovate so deeply not just in the technological but also in the organizational sphere. This is exciting thing to watch, and I would hope that the Thunderbird debate will in the end help to clarify the overarching vision of what Mozilla, as an organization, wants to be.
I haven’t seen any comprehensive explanation of how the decision was made to create the Mozilla Corporation as a wholly owned subsidiary of the Foundation (though I may well have missed this). I’d be interested to know whether an alternative was considered where the Foundation would remain “pure” and focus on development of the platform as a general-purpose framework for client-based internet applications. After all, it doesn’t seem right to say that the vast majority of Mozilla staff is working on the browser product. At least as far as developers are concerned, everyone I interact with is actually working on the platform.
An independent Firefox Corporation could focus on development and marketing of the browser product and contribute from its revenues to the operation of the Foundation. Other successful companies using the Mozilla platform would be expect to contribute as well (perhaps as part of some hybrid licensing scheme?). This would suggest a neat solution to the Thunderbird conundrum. The Thunderbird developers should put together a business plan and raise funding to enable them to continue operations under the auspices of a new Thunderbird Corp. Obviously this would require some promise of revenues and profits in the future. If they can’t raise money then I would submit that they probably aren’t meeting a pressing and unfulfilled need.
Another thing I like about this idea is that it promises increased incentivization for outstanding individuals in the Mozilla universe. Firefox Corp could eventually be acquired or go public, so people who could potentially be major contributors but aren’t motivated by idealism alone would have a more concrete reason for getting (and staying) involved. This might seem anathema to the whole open source ethos, but it’s suggestive that some of the key drivers behind Firefox have long since left the project to set up venture-funded companies.
I just attended a session here at OSCON squaring off a couple of new Rich Internet Application (RIA) technologies: Sun’s JavaFX and Adobe Flex. Nandini Ramani of Sun and James Ward of Adobe were on hand to demo their respective approaches.
As I understand it, JavaFX is Java but better integrated with other web technologies. The original pitch for Java was as a way to run multiplatform code on a webpage. As it transpires, it had its clock cleaned by JavaScript and Flash, and has found far more success as a server-side technology. Nandini plausibly suggested that one of the biggest barriers to adoption has been the inadequacies of Java graphics frameworks such as Swing. Now Sun is giving it another go, using HTML as the UI language with lots of whiz bang script goodness. The result looks a lot like, well, Flash.
This is certainly a big improvement. The only problem is that Adobe has moved on, and the Flex demos were nothing short of jaw-dropping. James kept muttering mea culpas to the effect that “I’m not sure why anyone would actually want to do this,” but who cares when flipping the pages of a virtual book that overlay as transparencies or run live video looks so damn cool? He made a big deal about Adobe’s Tamarin scripting engine, which has been open sourced and donated to the Mozilla project. Thanks to just-in-time compilation, it apparently runs tens or even hundreds of times as fast as old school JavaScript.
Bottom line: If JavaFX is Sun’s last gasp in the web client space then they have their work cut out for them. When the session ended, James was mobbed by adoring open source geeks as Nandini looked on forlornly.
Seeing these sexy demos steeled my resolve to soup up the AllPeers user interface with some sort of RIA technology. As a Mozilla shop, the natural inclination would be to explore the use of more open technologies like SVG or simply HTML and JavaScript. (Am I still allowed to call it DHTML?) But Flex has the indisputable attraction of building on tried-and-true Flash and has a slick Eclipse-based IDE. I asked some Mozilla folks at their booth what they thought, and they said “try both”. Hopefully I’ll be able to free up some of our UI people’s cycles to do exactly that, in which case I’ll be reporting back here with our conclusions.
There have been a couple of great articles recently about our upcoming BitTorrent support on Ars Technica and Wired. So you might be excused for asking: where the heck is it?
We’ve already got a great release candidate of v0.70, which is the new AllPeers version with BitTorrent. The public release should be out in the next couple of weeks. In the meantime, if you have any questions about our BitTorrent support, feel free to ask here in the comments.
Update: The outage is now over.
We are performing some necessary upgrades to our servers. The AllPeers network will not be accessible during this period.
Expected downtime should start at:
24/07/07 6pm PDT
24/07/07 9pm EDT
25/07/07 2am GMT
25/07/07 3am CEST
25/07/07 11am EST
and last for 5 hours.
We apologies for the inconvenience while we are upgrading our back-end infrastructure.
A complete transcript of the last Mozpad meeting is available on the wiki.
We seem to have nailed down the initial action items for Mozpad, and I’ve updated the wiki accordingly to change the wording from “proposals” to “projects”. There are seven, just like the Deadly Sins (coincidence?). Now I know that some people will be saying that we’re biting off more than we can chew or that some of these projects are worthless. (”IDE? We don’t need no sticking IDE!”) As far as I’m concerned, all these projects are fantastic, and if people are motivated to work on them I can’t imagine why we wouldn’t want to do just that. If it becomes apparent at some point that one or more projects has been abandoned, we’ll act accordingly.
Anyway, the bulk of Wednesday’s meeting was spent running down the status of the various projects. I’m championing the API project. I recently published a shell script that extracts the names of all Mozilla APIs used by a project, and the results when the script is run on AllPeers. Next steps will include improving the script’s coverage, getting data from other projects and reorganizing the presentation of the interfaces on the website.
My original intent was to use this data to decide why interfaces to include in future versions of the Mozilla SDK. Benjamin Smedberg suggested that we also gather information about use of XBL, XUL templates and the like. He further noted that, in his view, all APIs should be included (using separate directories to indicate which are subject to change and therefore riskier to use). He also explained that Mozilla 2.0 may have a totally different approach to expressing APIs, calling into question the value of data on existing APIs in the longer term. Nonetheless, I’m convinced that this work has a lot of value. If anyone can help to qualify that value by suggesting uses for API usage statistics, please let me know.
The champion of the documentation project (Mukunda Modell) wasn’t around, but we did get some useful suggestions. Brian King wants someone to translate any XULFr documents that maybe of interest to the broader Mozilla community. If you can translate French to English and want to help, please consider tackling this. Mark Finkle proposed that we restrict the scope to “that which is XULRunner but not Firefox”. The next deliverable should be an outside of the future platform section on MDC which makes the vision and scope of this work clear.
There is apparently some controversy as to the utility of the platform viability project (Andreas Wuest, Ben Longoria). Some feel that it will produce a vague wishlist but no resources to actually work on stuff. We didn’t have time to finish this debate, so Andreas promised to post a thread to the newsgroup. Personally I am pushing for a Pligg-based voting system which I think will have great PR value and probably provide some very interest data from the community as well.
None of the IDE project champions were present (ominous?) so we didn’t make much progress there. Clearly a lot of people are still skeptical about the idea of a Mozilla IDE. I think it rocks but I don’t have any time to devote to it. If we end up with a good requirements analysis and then functional specification, I’m optimistic what we can find a way to push the implementation forward. We’ll see.
A couple of weeks ago, Jason Kottke published a couple of brilliants posts comparing Facebook today to AOL in the nascent days of the web. Like Jason, I’d like to stress that I have nothing against Facebook as a company. In fact, unlike every social networking app I’ve tried in the past, I find it entirely too addictive and check it at least once a day, more if I’m engaged in a cut-throat game of Roshambull.
I was blown away when Facebook unveiled their apps strategy. I’ve felt for a long time that a highly leveraged online identity is key to the next generation of web applications. Everything from matchmaking to job hunting to discussion forums to shared calendaring requires a strong identity and detailed user profile, with much of the same data (often including a list of people I know) needed for each app. Web apps also need to be much easier to install with a more consistent user experience. Facebook apps do all of this, installing with a single click, piggybacking on your existing profile and integrating cleanly into a unified dashboard with a sidebar, minifeed, etc.
It’s hardly surprising to see innovation like this happening in the walled garden of a high-powered startup rather than in some smoky back room at the W3C. Nonetheless, Jason is absolutely right: the things that make Facebook apps cool will be so much cooler when they are available on the open web. I want all the zero-registration instant installability and harmonized user interface of Facebook apps, but with the ability to deploy my application on my own domain, in the programming language of my choice and without the permission and oversight of another company. I want to be able to contribute at the platform level. And I want much more inter-application integration (implying a common data model of some sort, something Facebook sorely lacks).
Nick Gonzalez spots the trend and hypothesizes that Plaxo’s “social graph” could be the basis of an “open Facebook”. Perhaps, but I feel strongly that the right place to mediate between web apps and web services is in the browser. The obvious candidate to bring us strong identity and a social graph that can be leveraged across applications is Mozilla, whose stated aim is to promote choice on the web. Any walled garden is the natural enemy of this mission. It’s a heady time of change in the web world right now, and I’ve been inclined lately to contend that Mozilla’s primary competition is AIR and Silverlight, not IE and Safari. But maybe, just maybe it’s actually Facebook.
Our shell wizard Jakub wrote me a script to analyze our source code and list all nsI* interfaces that we use, along with the occurrence count for each one. Here are the results. The script is a quick-and-dirty first version that needs to be extended to eliminate false positives (”nsI” in the middle of a word, “nsI” followed by a lowercase letter, etc.) and to add other interface prefixes (e.g. mozI*). Concrete suggestions would be most appreciated.
Once the script is a bit more mature I’ll start bugging other companies and projects to run it and give us the results. Over time I’d like to get a page per interface onto the Mozpad wiki, linking to the corresponding page on MDC and adding info about who is using it, how much and for what purpose(s). This information should be invaluable in deciding what should and shouldn’t go into the SDK.
I don’t know if you’re serious, Asa, but this is a fantastic idea. I’m not sure what kind of prize you can give to a whole country (a new gas pipeline?) or who exactly we would give it to, but a contest for the first European country to cross the 50% mark could potentially get us a ton of publicity here in Europe. I can totally see the Czech press writing about this and vaunting our not entirely ridiculous 36.6%. I don’t think that most people realize how exceptionally successful Firefox has been in Europe, and this could help greatly to get the word out.
Perhaps the best kind of prize would be virtual: a congratulatory banner on mozilla.com for a week or something. I’m not really in a position to be more than a cheerleader for this (though we could help with Czech press contacts) but I hope someone who is will pick up the ball and run with this.
I got a lot of great comments to my most recent Mozilla IDE post. The message that came through loud and clear is: don’t try to develop your own development environment from scratch. My attitude is more Not Inventing Here than Not Invented Here, to the extent that the success of any development project often seems to be inversely proportional to how much it is trying to invent. So I’m very receptive to the idea of using Eclipse or NetBeans as a starting point. I’m also interested in knowing more about how well Komodo IDE covers our requirements, although the fact that it isn’t free software is obviously a major drawback. I plan to take a closer look when I have time.
Once the Mozpad IDE team has come up with a list of requirements, I would submit that choosing the appropriate starting point will be one of the most important next steps.
I totally forgot to mention pipe-based IPC in my XULRunner wishlist. We definitely need this, so chalk up another vote. I’d have to think of a plausible use case, but I believe that profile sharing (multiple XULRunner instances all using the same profile directory) is sorely needed as well.
John Battelle posted a short “wow Facebook is hot right now” piece which is basically an extended introduction to an intriguing question: why is Facebook so damn hot right now?
I think the answer is related to the well-known technology hype cycle. (Mark Andreesen had a funny take on this recently, and I’ve talked about it in the past.)
Everyone sees the social networking space as a winner-takes-all-proposition. So what will determine the winner? Well, usually it’s the player who managed to generate the most hype and excitement. Facebook has done a great job attracting attention, and has now crossed a notional threshold where people see it as the most likely eventual winner. At this point a virtuous cycle takes over. The more hyped the site is, the more obvious it is that it’s going to be the last man standing. This results in even more hype and an ever-ballooning potential valuation. Rinse, soak, repeat.
I also gave a demo of AllPeers to Scoble. The wifi in the hotel was superflakey so I wasn’t connected to the internet, causing me to do a double-take a couple of times when AllPeers didn’t do stuff that it can only do when it has connectivity. But think I recovered reasonably well (grace under pressure is my middle name, people). Disappointingly, Scoble’s hand did not make an appearance.
3 weeks ago, we met Robert Scoble for a little chat about AllPeers and what’s coming next. So if you want to see Matt, me, Scoble’s hand and a couple of Nokia N95 click play below.
I’d like to have a Mozpad meeting next Wednesday (same time and place as usual). Two things I’d like to discuss are status of the various action items and a more general look at the IDE project (which I’ll be posting more about in the interim).
If people can’t make it or don’t think it’s worth meeting in this timeframe, please let me know.
