Wandering electrons... but not too far

Version 9.5 of our beloved Opera browser has arrived, and with it a pretty good number of changes.

You can of course just "upgrade" your browser. What I mean by that is you go to the download page and install the new version over the old version. But with a relatively major upgrade such as 9.5, it might be (understand "is") worth it to do a clean re-install. When you've been using Opera for a long time, all the customizations, plus the upgrades, plus this and that, can start to have a negative impact on performance (Like Windows does as time passes). Sometimes a good thorough scrubbing is a really good thing. And here, with a major upgrade, certain unanticipated problems may appear that can be avoided by starting with a clean slate.

So!

Read more...

Opera caches webpages that you have already visited to improve the speed of the browser.

Nevertheless, starting with the 9.x version, Opera does, by default, send a request to the website to ask if anything has changed, and if so, it loads the new items. This of course is not a bad thing as it assures that you have the most up-to-date webpage to look at. But sometimes this can be "too much of a good thing". In particular when you use the back button. Indeed, if you're using the back button, that means that you were just there a few minutes ago, and the chance of something new AND important (i.e. not ads) being there is virtually nil.

For example, go to the New York Times and click on any article, and then on your back button. Did you see how there was a slight pause during which certain elements were reloaded? That's because the NY Times site "told" Opera that there was new stuff, so Opera loaded the new things in. But I'll bet a million bucks (uuh, that's just a figure of speech) that the only thing new was the ads.

What we're going to do is tell Opera to simply reload what it had previously cached, without checking for new items from the server. The result is that the back button takes you instantly and exactly to where you were, without the lag provoked by the renewing of content. I find that this really enhances my web-surfing experience.

Read more...

Ryan at CyberNet News published the results of his Windows browsers comparison, which includes both beta and stable versions of the four biggies: Internet Explorer (versions 7 and 8 Beta), Firefox (versions 2.0.0.12 and 3 Beta 4), Opera (versions 9.26 and 9.5.9841 beta ("Kestrel")) and Safari (version 3.1).

He tested Javascript performance, page loading speed, and memory usage. The results found good to excellent performance both for the stable Opera 9.26 version and especially for the 9.5 beta, but the competition is fierce from Safari and the Firefox beta.

For javascript, the Opera 9.5 beta came in second after Safari 3.1 and 9.24 was fourth behind the Firefox 3 beta. The last to cross the finish line--and long after the others were home in bed--were the Internet Explorer versions.

For the first page-loading speed test, done on the official Google blog, Opera 9.5 was the winner, beating out Safari 3.1 in second place and again 9.26 was fourth behind the Firefox beta. In the second speed test done on the Yahoo! Search blog, Opera 9.5 and 9.26 were respectively second and third behind the winner Safari. IE7 managed to finish 7th in front of Firefox 2 in the first speed test, but IE8 was last in both.

As for memory use, the surprise winner was the Firefox beta, using globally between 21.3MB and 124.7MB according to how many pages were open and for how long. The Opera versions were "middle of the pack", let's say. The Stable version of Opera was indeed better than the stable version of Firefox, but the beta version of Opera was behind the Firefox beta in all but startup memory use. Looking more closely at startup memory, the IE versions were, of course, the winners there, and the loser was Safari; surprising considering its excellent results elsewhere. The IE versions were however the worst for in-use memory usage (with multiple sites open).

The results of Ryan's tests seem to suggest that there is indeed a new player in town, and his name is Safari (despite a rather rough start). The difference in Javascript performance and page loading speed are pretty insignificant between Safari 3.1 and Kestrel (not to mention the significant improvements for the Firefox beta), but the Opera development teams are definitely going to need to start thinking of Safari as serious competition in the Windows environment (I'm sure they already are).

Yes, to summarize, the Windows browser performance wars are now a three-way battle between Safari, Opera and Firefox. Internet Explorer seems to be KO for the moment, its only advantage being its ubiquitousness.

It would be interesting to revisit these results once the various betas are all officially released. Indeed, testing beta performance is a bit like tasting a cake while it's still in the oven. In particular, I think one should not jump to conclusions concerning IE8's poor performance in Ryan's tests, as it is the least mature of the betas. But what is sure is that the folks at Microsoft need a lot of Gatorade if they expect to catch up with the others concerning performance.

For those of you out there who have been looking for a clarification on this whole "version targeting" thing that Internet explorer 8 is going to introduce, and how this ties in to other browsers, such as, ahem, Opera, I recommend a tour on the "A list Apart" website, where you will find three articles that address this question:

“Forgiving” Browsers Considered Harmful by J. David Eisenberg

Version Targeting: Threat or Menace? by Jeffrey Zeldman

and

They Shoot Browsers, Don't They? by Jeremy Keith

You're in for about, oh, perhaps 20 to 30 minutes of reading, but when you're finished you'll have a much clearer picture of Microsoft's logic behind the decision to create still another line of code for website developers.

As to whether it's good logic or bad logic, the discussion is open. Let me know what you think!

This is a tips and trick for Opera versions 9.2x. In version 9.5 re-opening closed tabs is done with Ctrl+Shift+T (unless you've activated 9.2 style keyboard shortcuts.)

A forum post entitled "2 old bugs" attracted my attention this morning. One of the "bugs" in question was indeed something that's been bothering me for a while as well.

The problem:

You probably know already that the keyboard shortcut "Ctrl+Z" will re-open a tab that you (usually accidentally) closed. However, this may not work when there is only a single tab (web page) open. This is the case for me: If I accidentally close a single page, my Speeddial appears and Ctrl+Z will no longer open the previous page.

Bpm took the time to look at this problem closely and found the problem.

What's happening is this: When you close that last webpage, Opera sends you to a "blank" page (or an "empty tab" if you prefer). According to your setup, this may bring up the Speeddial, but the Speeddial is a blank page from Opera's point of view. Because the page is blank, Opera activates the address field (you'll see the cursor blinking there). You can think of the address field as a type of text box. When you're in a text box, Opera presumes that you will want your keyboard shortcuts to affect the text your typing and not the browser as a whole; it can't do both at once. For example, as I'm typing this blog entry, Ctrl+Z undoes what I just typed, but if I click outside of the text box, Ctrl+Z opens the last page that I closed.

So two things are now together: 1) you're on a blank page, so it's impossible to "activate" it, and 2) Opera has moved you to the address field, but you haven't yet typed anything. Result: Ctrl+Z does nothing.

(For that matter, this can also happen when multiple tabs are open. If you close a tab, Opera sends you, by default, to the previously active tab. If by chance that tab has its address bar or some other text box active, you'll have the same problem: Ctrl+Z will affect the text field and not re-open the just-closed tab.)

The solution:

As bpm pointed out: the fact that Ctrl+Z can re-open closed tabs in certain circumstances is just an extra that comes from it's nature: "I, Ctrl+Z, undo that last thing that was done, according to where I am." For the situation described above, Ctrl+Z finds itself in a conundrum: "here where I am, there is nothing to undo."

So how do we resolve this?

Simple.

Use the keyboard shortcut that was meant all along to re-open closed tabs: Ctrl+Alt+Z

Yep, it's been there forever and I had never noticed. Click your trash can and you'll see this shortcut next to the last-closed webpage that's listed there. This shortcut will re-open the last-closed webpage (and then the next-to-last-closed page, etc.) whatever the circumstances may be.

And then finally, think of Ctrl+Z as a 'text-oriented' keyboard shortcut, like in Word, just to avoid confusion.

A big thanks to bpm!