A few weeks ago Opera released their version latest version (9.5.0 back then, now at 9.5.1) of their desktop browser. As a huge fan of OperaMini on my Nokia phone I downloaded and tested it. And it's really a good browser, which some huge advantages, like the speed dial, the responsiveness of the GUI the gadgets (I have fishies!), the dock icon makes me run the same same instance for weeks, the standards compliance (my browser here gets 83, not 100 but still much more then all other browsers), the idea that you can search a piece of selected text… and more. I mean much more… like that crappy small package is packaged with RSS reader, a Mail client, an IRC client… and I probably missed a feature or 5… and it looks great…. this is really a visually nice GUI. I changed the theme to a Tango based theme, but still the default GUI is nice!
However, after using it for several weeks I have to revert back to using Firefox. It's not because of web standards, but it's because they don't mean a lot in real life, sadly. It seems Joel was right (1), and the most important thing is how they are implemented – which means once more you need to test your site on all browsers available, and testing the syntax against the W3C test application is not enough.
For example, on a site I visit I see some elements outside of a div, which not hapenning on IE*, Konqueror, Firefox… on another site which uses CSS to layout the main page the left columns is pushed down. And then we get to site designed to work on Firerox…
One nice example is GMail which is set to work on specific browsers. Then we get to Flash (youtube does not work on Opera/Mandriva 2008.1, and again some other random flashes do not work …). Also the richtext editor of WordPress does not work on Opera. Even the plain text editor of Opera is really unpredictable on BiDi texts (how the hell did they mess up QTextEdit3…?).
(1) I hope this will not be true in future. The acid test are cool, but we need a large, I mean 10,000-500,000 tests which browsers need pass before they are called "standard browsers" (2).
(2) We have this for pages, the W3C validator, but we need this also for web browsers.