Phoenix Improves

I’m sure that headline could be used daily. I’ve upgraded to the October 6 nightly of Phoenix and am happy to see several long awaited changes. My bookmark toolbar gained a chevron and drop-down menu (like IE) when it was sized too small to fit all the bookmarks in the window (perhaps fixed by bug 171604). Favicons are back in the bookmark menus for Phoenix. I’d forgotten how nice they were before they were yanked from Mozilla. Finally, after years of waiting, bug 28583 was fixed. Tabbing into text fields or focusing them (except with a click) now selects all the text as it should. This means JavaScript prompt with default text also works. This change affected both Mozilla and Phoenix.

Phoenix 0.2 is out

It’s looking really, really good. The only changes I noticed from the previous build I commented on is that the sidebar now has a close button and the scripts and tabbed browsing preferences are now back. Go download it.

Mike Shaver writes what I’ve been thinking: “I guess it’s a little embarrassing that a handful of hackers can produce better autocomplete, better toolbar management, and much better performance than that found in the much more heavily attended Mozilla CVS tree. But we’ve always known that small, sharp teams are vastly more productive than those diluted with a few dozen mediocre-or-worse additions, so it’s certainly not surprising.”

I’ve switched from Mozilla to Phoenix as my primary browser (mostly to test it out) and miss a few things. Type ahead find which was working in previous Phoenix nightlies is now broken. Image blocking is also missing. Look for both in 0.3. I can’t wait to be able to add back some of the Mozilla extensions. I most miss having chatzilla a click of a bookmark away.

Phoenix is moving quickly

Wow! The latest Phoenix nightly builds have some terrific improvements (I played with 2002-09-29-15 windows):

  • Customize the toolbars by dragging and dropping the buttons directly on the toolbar (no dialog necessary)
  • Ability to add new toolbars (create a separate one for URL, for example)
  • History, bookmarks, and downloads can now be sidebars (and you can add toolbar buttons to toggle them on and off)
  • In-form field autocompletion as with IE (start typing in a field and it gives you suggestions from previous things you’ve typed)
  • An interesting search box to add to the toolbar (I think I’d rather have a Find button.)
  • Drag and drop bookmarks to the bookmarks menu (This may have been there for a while, but I just noticed it)
  • Did I mention that it’s fast?

Phoenix, Minotaur, and Mozilla

I’ve gotten some questions about how Phoenix relates to Mozilla. Here’s how I see things as an interested bystander that doesn’t have any inside knowledge about the project. (David Hyatt’s quiz may also help you understand the differences between Phoenix and Mozilla.)

The Phoenix Project’s goal is to build the best web browser for most people. The Phoenix readme says “the interface will not be ‛geeky’ nor will it have a ‛hacker-focus’. Nor will it be ‛minimal’.” The project was started and is run by some of the core Mozilla developers that were frustrated by the restrictions and pressures placed on the Mozilla browser. They didn’t like the compromises forced on Mozilla due to marketing and other pressures within Netscape/AOL. They also wanted to work more quickly and with fewer check-in restrictions. In some cases they wanted to experiment with optimizations that may also be applied to the Mozilla code.

Will Phoenix replace Mozilla? Well, I suppose that depends upon what Mozilla means to you. Phoenix currently builds on top of Mozilla and shares a bunch of the code. Most of the changes in Phoenix are related to the user interface (UI), which is the part of the browser that you see and interact with (menus, toolbars, buttons, dialogs, etc.) Phoenix is going to be just a browser, not an entire suite of applications (email, address book, news reader, irc client, HTML editor, slicer-dicer, julienne fry maker). Don’t let that disappoint you, though. The Minotaur Project is working on a standalone mail client. Other Mozilla components will likely be available as add-on extensions.

The developers hope that if they focus on a particular application they will be able to build it better and make better decisions. They recognize the importance of being able extend the application and are planning for it.

This is a 0.1 release?

On Monday the Mozilla Phoenix project released its first milestone, version 0.1. Download it and try it out. This is a terrific first milestone. Because of its heavy use of Mozilla, this browser behaves more like a 1.0 release in terms of the quality and capabilities of its page layout and rendering. Yes, there are many parts that still need polish, but this is awesome. Keep up the good work, guys.

Of all the things the Phoenix developers are doing, I believe the plug-in/add-on manager is one of the most important. It is getting more difficult for add-ons to integrate into Mozilla, let alone the Gecko-based browsers that they should also work with. In most cases, the add-ons make assumptions about what menus are available and add overlays. I was thinking the other day, what if Windows 95 had included a generic program installer. Every program does similar things when installing. It would have made more sense to just have a standard package format and script for the installer to run. Of course then InstallShield would be out of business and Microsoft would be charging for installs of everyone’s apps. In any case, it would be nice if Phoenix could set up well-defined hooks into the application that mean that the add-on doesn’t need to know much about the menu structure. Microsoft products, especially Word, have had this kind of extensibility for years.

Off the top of my head, here’s a few application hooks that I believe will be important for add-on creators. I’m thinking new menus, new toolbars, and new buttons for specific toolbars will be heavily used, particularly now that toolbar customization is a reality. Event hooks to support things like mouse gestures and context menu changes also need to be considered. I suspect there are more exotic kinds of hooks that haven’t even been considered because nobody’s invented useful things for them yet. For example, I can imagine automatic spell checking of textareas and inputs or adding user page load filters. I’m hoping that Phoenix will provide an elegant mechanism for adding on to the application and managing these add-ons. If they do this right, I hope it can become a part of all Mozilla-based browsers.

Is it a Phoenix bug?

I’ve been trying out the latest nightly of Phoenix and was about to complain that form scrollbars were broken. Turns out that bug 170184 is undoubtably the cause. Sharing most of the codebase with the Mozilla trunk has many advantages, but running into this kind of problem when you’re trying to release a milestone has got to be frustrating. Since Phoenix contains so much of Mozilla, determining where a bug is may be challenging. At least they’re in the same bug tracking system so it will be easy to move them if they’re incorrectly classified.