Galeon

» see galeon.sourceforge.net as it lived, on the wayback machine

~ the obituary ~

I missed Galeon. The browsers on the machines I sat at through those years were Internet Explorer and then Firefox 1.0, and Galeon never left the GNOME desktop. It is a grave I can only read about, and the reason it is worth reading about has almost nothing to do with browsing the web.

Marco Pesenti Gritti released the first public version in June 2000, numbered 0.6. It used Mozilla's Gecko engine for rendering and GTK for everything else, on the argument that Mozilla's own suite was slow and enormous while its layout engine was good. Galeon's pitch was the engine without the suite: a browser that started fast and did nothing except the web.

For about two years it was the obvious choice on Linux. It had a configurable tab bar before tabs were standard equipment, along with mouse gestures and smart bookmarks that dropped a search field straight into the toolbar. It also had an enormous preferences dialog, because the people writing it wanted every behavior exposed, and that dialog is what eventually broke the project.

~ the case file ~

~ the fork ~

In 2002 GNOME adopted its Human Interface Guidelines. The core instruction was that an application should make a sensible decision on the user's behalf rather than offer a checkbox, and that a preference nobody could explain should be deleted rather than kept. Applied to Galeon this meant removing most of what its contributors considered the point of it.

Gritti agreed with the guidelines. His own contributors did not, and the disagreement was not about any particular option but about who a program is written for. In November 2002 he stopped working on the browser he had created and forked it into Epiphany, which shipped with the simplified interface. The remaining Galeon developers carried on with the 1.2 line and later a 2.x branch that put the options back.

~ the argument about settings ~

This fight has been had in every software project since, usually badly. The case for stripping options is that each one is a decision the developer failed to make, and that a program with four hundred preferences has no design at all, only a record of past arguments. The case against is that defaults encode assumptions about who is using the thing, and the people most likely to be poorly served are the ones who care most.

Galeon is useful because both halves were shipped and both were maintained for years, by people who had been on the same team the week before. The simplified branch became GNOME Web and is still in the distribution today. The branch that kept its settings ran for six years and then ran out of people. That is not a clean verdict on the argument, but it is a result.

~ where it went ~

The thing that actually killed Galeon was neither side of the fork. Firefox arrived, shipped as the default browser in every major distribution, and made a Gecko-based GNOME browser into something you had to go and install for reasons you had to explain. A small project without its founder has no answer to that.

Development stopped in September 2008. Release 2.0.7 went out on the 27th and the maintainers announced that the project was not sustainable and that their effort was moving to writing extensions for Epiphany instead. Debian and a few other distributions carried the package into 2011 out of inertia. The features survived, ported one at a time into the browser that had forked away from it.

@}-->--

              )
             (_)
            .-'-.
            |   |
            |   |
            |   |
            |   |
          __|   |__   .-.
       .-'  |   |  `-:   :
      :     `---'     :-'
       `-._       _.-'
           '""""""

« previous  ·  random  ·  next »