Slashdot Mirror
Mozilla Jumps on 'Lean Browser' Bandwagon
fader writes "Following in the footsteps of fast (and often fantastic) wrappers around Gecko (the Mozilla rendering engine), Mozilla has just released their own lightweight browser, Phoenix. Only Phoenix will still use XUL, the cross-platform markup language used for the current Mozilla interface. Will it still be fast enough to overcome the final gripe about Mozilla, namely that it's just too slow?"
is for my gtk theme to take over the Mozilla theme. Widgets and whatnot, not just color. I don't mind having buttons and layout set by moz, but I'd like an integrated feel, like it's part of the system... esp since it's the app I use most. I won't use galeon, mainly because it doesn't have some of the bells and whistles that mozilla does(that I do use).
Is this just YAGBB (Yet Another Gecko Based Browser) or will this be the start of a modularization of the Mozilla browser???
I am a happy user of Mozilla, but i dislike the monolithic approach of integrating browser, mailreader, newsreader, composer and you name it into one executable. What happened to the old and proven Unix approach of "Do only one thing, but do it well!"?
I hope Mozilla in the future will be split into a suite of components, that work well together and with a consistent interface.
there are a few simple things that would make it feel so much faster....
1) Cache a picture of a blank page instead of mucking about drawing everything from first principles every time. Show this (or whichever part the user has chosen to start up with) FIRST before doing anything else. It doesn't matter if the thing isn't clickable yet, there is plenty of time to get to that stage while the user is moving the mouse. Buffen any clicks the user manages to make before you are ready and they will never notice.
2) Accept and buffer keyboard input while pages are drawing. I get so annoyed that I can't fetch one page and then get a new browser window to open - even Netscape 4 let me do this!
3) Cache the way the mail window looks and restore to that when it's opened (see point 1)
Things like this would give an impression of improved speed with practically no change in the actual code. Hell, you could even take the startup pic away earlier in the loading process and it would make the thing feel faster!
A pizza of radius z and thickness a has a volume of pi z z a
This all started with Apple's QT 4 player, which completely broke the highly regarded Apple Human Interface Guidelines and was put onto the Interface Hall of Shame just for that. Then Winamp came out, creating one of the first in-app skinnable applications, which is cool, but led everyone to release skinnable apps, such as Windows Media Player, and a lot of similar ones on the *NIX side. Sure, it's a media player, you don't interact with it like a word processor or the like, but there's something to be said about interface consistancy when teaching computers to newbies. That's why it's odd that Apple broke that mold with QT4, as they lived and died by the HIG in their efforts to promote the Mac system.
Now with MOz's interface scheme, as with a lot of other cross-platform libraries like Java, QT, etc, it doesn't tie into the OS control toolkit and instead relies on drawing it's own widgets. To do the former would have to break cross-platform ability (I've yet to see a fully cross-platform system that uses the system's native toolkit, mostly due to lack of certain features in some kits compared with others. Even those that try to do this typically have to hard code certain settings that the user would normally be able to change -- I have a friend (hi paul!) that typically likes light text on black, and it's amazing how many Windows-native programs alone don't use the system colors, or use them inconsistantly as to make programs unusable.) It's understandable that WORA is a lofty goal, but there should be more push to try to provide some system native level that can be easily built without too much problem. For example, Nethack is a good example where out of the entire source tree, only a few special files are needed for supporting a different interface, including text and graphic variations; someone even pasted a Diablo-like orthorhomic few on top of the Nethack code, by only adding the appropriate hooks for that GUI. I'd rather see more effort here with Moz and other programs to provide this, though with much effort, than to keep on reinventing customization wheels that are inconsistant with the OS's customization.
"Pinky, you've left the lens cap of your mind on again." - P&TB
"I can see my house from here!" - ST:
What is the point of developing another nerdy XUL based Mozilla browser? Have the lessons of the Mozilla project not yet been understood? Some of the biggest weaknesses of the Mozilla browser can be attributed back to XUL. XUL enables cross platform applications to be quickly built, but for this developer convenience the biggest trade off for your end users is that your application will never fully conform to the native user interface of the operating system it is run on. A secondary concern is the memory and processor cost of the XUL layer - no one wants a fat and slow browser, caused by having to compile and run a Java Script based user interface at runtime.
Why doesn't the Mozilla project develop fully native user interfaces around the Gecko HTML rendering engine instead of wasting precious time and development resources on another dead-end XUL based browser. A number of separate teams have already started such projects independently (Chimera, K-Meleon & Galeon). The Mozilla team need to refocus their efforts from developing half-caste XUL based browsers toward building native front-ends for each operating system that can complete head-on with the more popular commercial browsers. An XUL based application will just never cut it for the masses.
I'm having a heck of a time finding a lean browser to run on this thing. I haven't even attempted Mozilla. Galeon is too big, sending my poor machine deep into swap. I tried downloading Opera, but it kept complaining about not finding the right version of libXm.so, even with the statically-linked version.
I see lots of talk about how fast this Phoenix is, but I've yet to see *any* mention about its memory footprint. Is it really lean, or is it simply lean as compared to Mozilla?
I now have dillo running, and it looks promising. Any other suggestions?
(No, buying a new computer is not an option. I remember running browsers on my old 486, so this shouldn't be impossible!)
I'm a leaf on the wind. Watch how I soar.