Gav

(With apologies to members of the choir who've heard this sermon before)

If you use the Internet a lot, you need to check out Firefox 3 when a flavour appropriate to your personal pre-release-software-comfort-zone is released (Beta 5 is out already, Release Candidate 1 is due soon, Final release is due in June). Here's why:

The primary methods that people use to return to sites they've visited before are the location bar autocomplete and Google. Which is crazy, because browsers have features to help with that - history and bookmarks (IE calls them "favorites"). Unfortunately, bookmarks are a pain to use, especially if you have lots of them and can't be bothered to organise them. I have at least hundreds, and only really do a cursory sort when I get pissed off at the mess they've become. I gather that makes me far more organised than most people. It's better than nothing, but a pain to use.

(Personal history diversion: One of the main reasons I stuck with Netscape 4.x almost to the very end¹, and swapped to the Mozilla Suite (now known as SeaMonkey) when I did leave, was bookmark handling. Back when NN3.x/4.x was my browser, search engines sucked (this was in the dark days before Google) so using bookmarks quickly became a habit if I ever wanted to find stuff again. In essence, the competition's bookmark handling sucked. Netscape/Mozilla's sucked as well, but was less sucky than everybody else's.)

I am somewhat boggled that this area has been neglected by browser manufacturers for so long. The Mozilla people reached the same decision some time ago, and decided to do something about it.

In summary, they've made bookmarks much more usable for both normal people and power users, and added loads of cross-linking between the address bar and history/bookmarks so that the autocomplete is scary-good at finding stuff you've visited in the past.

So, it turns out I was wrong about Microsoft's IE8 standards-mode rendering switch - they're proposing that web authors use either a meta element or an HTTP header as an opt-in.

My initial gut reaction: wow, that sucks.

So: flip the default around, and this is a good idea. Otherwise, it sucks.

Finally, a word from the lemurs.

There's more good news about IE8 - it won't have "hasLayout" either. The message revealing this is extremely light on details, but if this has the consequences implied, then a whole category of IE-related layout strangeness disappears into thin air with version 8.

After a long silence, Microsoft have finally started talking about IE8 features, and it's actually good news - IE8 internal builds are now passing the Acid 2 CSS test¹.

While many will hurl (not entirely undeserved) general derision in MS's direction for being the last of the major browser vendors to get there², the important point is they got there. The Acid 2 Test is eeeevil. This means IE8 has substantial and wide-ranging fixes to their CSS support (including, but not limited to, position: fixed, float/clear, margins, generated content, ignoring bad declarations, display: table and associated gubbins) and also fixes to more obscure bits of HTML like the object element.

IE7 had quite a few fixes to CSS support, and was welcomed for that reason, but still lagged a bit behind the competition. If IE8 is passing acid 2, that's a huge leap forward. Once this version hits lots of users (probably a good 2-3 years after they release it, at least) this will finally open up to general use several areas of CSS2 that are currently off-limits. Headlines: position: fixed (menus that stay fixed in the window while the page scrolls, without frames), display: table (table-style layouts without table markup), generated content (tricky to describe, but it allows all sorts of cunning stuff). It will also significantly reduce the pain of making float/clear work cross-browser.

¹Unfortunately, someone managed to actually break the main Acid 2 test site recently, but the powers that be are on the case, and it should start working again sometime soon.

²Out of the big four, Safari were first (internal build 27 Apr 2005, general release (v2.02) 31 Oct 2005) then Opera (public experimental build 10 Mar 2006, general (v9.0) 20 Jun 2006) then Firefox (semi-public dev build 12 Apr 2006, general release (v3.0) early 2008)