I discovered Mitchell Baker's article about Application versus the Platform too late on friday to be able to comment it before, but her article ran so much in my head until now I can't sleep and need to comment.

First, it is good to see such an article. The Platform is immensely important to a rather special group of Mozilla users and contributors : companies building and selling services upon that Platform. I started advocating loudly for XULRunner when it became clear that Disruptive Innovations' customers were more or less all asking for or targeting XULRunner-based solutions. Then Laurent wrote about it ; the Paul ; then Matthew. And others. Beyond Disruptive Innovations, Joost, Songbird, Wengo, there are dozens of corporations or organizations bulding upon XULRunner. And I feel myself well positioned in that world to say they're not happy with the way XULRunner is managed at Mozilla, despite of the fact Mitchell sees "an enormous amount (invested) in the platform technologies; probably more than in any other area".

  • XULRunner has no marketing support at all. Where is XULRunner on Mozilla.org's home page ? Where is XULRunner on Mozilla.com's home page ? What's the real and stable name of XULRunner ? What's the logo of XULRunner ? Where are the marketing material like "Built with XULRunner" logos ? XULRunner is not even listed in the Mozilla projects linked from mozilla.org's home page ! Where is XULRunner in the press ?
  • the XULRunner roadmap does not really show the investment mentioned by Mitchell in her article
  • the XULRunner documentation is  far away enough from what embedders expect it's a severe break for the spread of XULRunner. There is a lot of magic in XULRunner for embedders, too much magic. The organization of that documentation is too randomish, and figuring out which document contains a given set of information is almost hopeless.
  • where is the official XULRunner 1.8.1.3 matching Firefox 2.0?
  • XUL needs new features. We need new stuff like editable trees, real dependant windows (moving the main window moves dependant windows with it), more window features. A lot of embedders work with flash and plugins, need software update, extensions update and skins. We desperately need simple and easy profile portability à la PortableFirefox.
  • ...

We need more investment on the platform, sorry to say. When Mitchell says "should Mozilla focus mostly on the platform as a general purpose platform?", I feel she's asking a wrong question here. We all know that Mozilla is nothing without Firefox and Thunderbird, that the buzz, the fame and the money come from the applications. Nobody is asking Mozilla to focus more on the platform than on the apps, and nobody ever asked for that.

What we desperately need is the following :

  1. a real name for XULRunner ; we keep hearing "xulrunner" is only a codename.
  2. a technical writer dedicated to XULRunner building a stable and well-organized set of docs and tutorials embedders can rely on. Of course, this can't be done w/o interviewing the developers and embedders, and the technical writer needs some good knowledge of moz.
  3. marketing.... In the world of Mozilla, XULRunner is just nowhere at all. The Platform does not need the kind of marketing backing Firefox or Thunderbird, but embedders need to feel their technological choice is supported by Mozilla or they will choose another platform from another software provider ; the fact we're good on browsers is pointless here, they can embed a browser component in a simpler/better platform.
  4. packaging... XULRunner does not even have an installer ! Even making a flash plugin work with a XULRunner zip downloaded from mozilla.org is pure magic to most developers (according to the number of people asking us!).
  5. the XULRunner team is... Benjamin Smedberg ! We all love bsmedberg, we all rely on him, but that's not enough and working on XULRunner itself requires knowledge that embedders rarely have. We need more of his time on XULRunner (if that's even possible), and we need at least one other person here so we're not lost in void when bsmedberg is away.

When I read Tristan Nitot's answer to my friend and colleague Laurent Jouanneau's article, the first word that comes to my mind is "bollocks". The second one is "blah-blah". The third one is "sigh". Like it or not, XULRunner is a reality and is a small part of the Mozilla big image. Like it or not, there are now projects and products with a big visibility counting on it, and the PR would be terrible  for Mozilla if these projects and products drop Mozilla for instance in favor of Apollo. Playing the ostrich, head in the sand, is a dangerous game here because we're almost at crossroads. Without changes in the way XULRunner is supported by Mozilla, Adobe and Microsoft are going to kill it and kill it rather easily.

Mozilla, please, turn XULRunner into a Mozilla product.