Follow Up--Fastest Mac OS X Browser
"Actually, you can make OmniWeb remember its window size. Size and position a window where you want it, and choose Save Window Size from the Browser menu. We chose this option because we weren't satisfied with any methods of trying to be smarter than the user, and would rather give them total control. (For example, open IE and then hit Cmd-N 10-20 times to create a bunch of windows. Note how it nicely cascades and resizes them to fit on screen. Now hit Cmd-Opt-W to close them all, and hit Cmd-N to create a new window... it's small and down in the lower right corner of the screen, because that was the setup of the last window closed.)"
"We'll be releasing 4.0cf4 soon (and hopefully finalizing 4.0.0 soon after that) with a few more speed and compatibility improvements. Thanks for your comments in the article."
This is good news because we are a fan of OmniWeb. We also like OmniWeb's approach to window sizing, now that we know it. The only drawback to some might be that each window is opened precisely on top of the previous one. This is a bug that will be fixed according to The Omni Group. We consider this to be a feature since the window does not have to be moved to be seen.
From Andrew Abernathy of The Omni Group:
"I have a note about window scrolling by holding down the scroll button: Apple's Cocoa frameworks (and presumably also Carbon) manage the base scroll speed (which holding down a scroll button). While you might see _slower_ scrolls if the rendering can't keep up, you're not going to see scroll speeds any faster than this threshold (unless the developer explicitly overrides it, which they almost never should). This threshold might be different between Carbon and Cocoa, and I can't speak with any authority for the Carbon support, but Cocoa apps are all going to scroll at the same maximum speed. (Take TextEdit, for instance: it scrolls at the same smooth speed whether it has a lot of text in it or just a bunch of blank lines. You should see similar results across Cocoa apps.)"
[Also] "...a view in a Cocoa app can specify how _far_ to scroll when the scroll button in pressed. So if one app set it's scrolling distance higher than another, it would seem to scroll faster (but would be potentially less smooth, and users might find that clicking the down arrow took them further than they wanted - it depends on the kind of data you're looking at)."
The second most frequently raised issue was a request for compatible performance data on Internet Explorer 5.0 (Classic). Here are the results:
| Action | IE 5.0 |
|---|---|
| Launch/Load Macs Only! |
|
| Scroll Macs Only! |
|
| Load SiteLink.net |
|
| Back to Macs Only! |
|
IE 5.0 (Classic) launched faster than Netscape but was slower to return with the back arrow. The other actions took identical times. One caution is that the actions in IE 5.0 were measured at a substantially different time than the other browsers and may reflect significantly different internet conditions. (BTW, we used a cable internet connection delivered to our PowerBook via AirPort.) Several people also suggested that we run Fizilla, Netscape's Mac OS X project, but it is still too much of a work in progress.
Finally, several people argued that the launch time for a Classic mode browser should include the launch time for Classic mode (Mac OS 9.1) as well as the browser application's launch time. Doing this would, of course, place the Classic mode browsers (Netscape 4.77 and IE 5.0) as the slowest to launch. Our view, however, is that as long as most applications still run in Classic mode most people will boot Mac OS 9.1 along with Mac OS X. In the future (hopefully the near future), this will not be true.
Anyway, we need to learn to use a new paradigm for application launch and use with Mac OS X--launch once, close the window when not in use, use sleep mode and click on the application's icon in the dock to make the application window open instantly. For PowerBooks to best use the new paradigm, however, Apple needs to address high battery drainage issues during sleep and an unacceptably high rate of freezes/kernel panics when waking a PowerBook.
Saw something? Send a tip
The archive ran on reader tips. What did you see, where, and do you want the credit?