Microsoft rolled out its Internet Explorer 9 Platform Preview 3 today, and, yes, it does include the HTML5 canvas element, and yes, that element is hardware accelerated, and yes, Redmond is still saying that performance and compatibility share "job one" status. Platform Preview 3 obsoletes the earlier versions, the first …
I probably slept through the apocalypse
All I can say is "hurray". Even if it took ten years, having the default mom and pop browser on the world's default mom and pop OS finally support the features we're using on established websites as well as experimental "standards" compliant HTML5 websites means today is not a day for pointing at MS and going "took you bloody long enough".
Hurray. plain and simple.
Are firefox and Opera currently hardware accelerated? or is this the only way microsoft can keep up?
Firefox 4 has hardware acceleration support using the Vista/Win7 technologies that IE9 is using (https://bugzilla.mozilla.org/show_bug.cgi?id=527707). This is currently being tested to hake out any remaining issues.
I don't know about Opera or WebKit/Chrome, but I suspect they have a similar thing in the works.
Not accelerated yet
Both Opera and FF are planning to introduce hardware acceleration in future versions. In order to prepare for this Opera has completely rewritten its implementations to take advantage of core OS functions. The IE test suite is pretty useful in indicating what can be done with the new toys (unlike Apple's turd of a demo). Of course, IE comes out top but I was pleasantly surprised to see how well Opera performs in some of the tests - lower frame rates than accelerated IE but still reasonable performance. FF was mainly too slow to be useful and it looks like the FF -> Chrome migration is happening. Worth noting that IE 9 won't run on all those XP machines out there while Opera still runs on Win98!
@Microsoft - kudos for adding support for Canvas. Already a reasonably heavy user of it with browsermob's charts.
Re: Safari only
Safari has, I think, always been hardware accelerated on OS X and on the iPhone. As of Safari 4, Apple are advertising that it is also hardware accelerated in Windows but I've no direct personal experience to tell you if they've made a decent job of that.
Super kudos to Microsoft for implementing canvas, especially after being equivocal at least as recently as D8 (I seem to recall?). I guess this'll leave people on XP who are sticking to IE as the only sizeable chunk without canvas in the next year or two?
Hardware acceleration & multicores
IE9 is certainly a fantastic effort from MS, and I'm not being sarcastic.
But as things stand now, can you imagine how much fast still Opera 10.6 would be if it also used a separate CPU core to run Carakan and used hardware acceleration for Vega?
In fact, Vega was designed to be hardware accelerated, they just disabled it by default for the time being. I'm sure within a couple of years, as the speed race hots up, they'll enable it and make a fanfare of it.
I remember IE6 being a fantastic browser when it came out (Mozilla Suite/NS6 was a bloated old dog), but it soon aged and was left languishing at the back. I do wonder how long IE9 will seem so fantastic for. Once you've played your ace cards of using hardware acceleration and multiple cores, all you're left with to keep up with the competition on a technical level is better coding. And MS are no longer at the height of their empire, where technical merits didn't count for anything. If IE doesn't deliver a good experience for the user, they will switch to another browser. It's not 2000 anymore.
I thought Flash Player 10.1 was ALREADY hardware accelerated?
Only specific parts of flash, like video rendering are hardware accelerated, most other areas arn't, so things like games do not benefit from hardware acceleration.
There is an OpenVG version of Flash
So, if you have an OpenVG accelerator, you can have HW accelerated Flash (not just the video). However, this may only be in FlashLite and therefor for mobile devices only (ish). And VG is a compile time switch...
No Decision Should Be Necessary
"And speaking of standards, director of IE product management Roger Capriotti told The Reg that although patent-backed H.264 is Microsoft's video codec of choice for IE9, it will also support Google's recent open source gift to the world, WebM/V8, if a user has that codec on their machine. However, Redmond has yet to make a decision, Capriotti said, about how to handle the open source Ogg Theora codec."
Why? The presumptive purpose of a codec is to provide a library which allows encoding*/decoding of a given format from any application. So why would any multimedia application not simply support whatever codecs are installed on the machine? This is not pointed solely at Microsoft, but at all browser/multimedia application developers -- and codec people as well. In my mind, if your software can't or doesn't simply identify the codec of a particular file (or codec requested, for encoding), pass it to the right decoder (or encoder), and pipe the output to the appropriate location, then you're not doing it right (or the codec author didn't do it right).
This is why I believe HTML5 SHOULDN'T specify codecs for the various media-related tags -- it shouldn't HAVE to.
* Yes, I know they make software that will only do the decoding and call it a codec, but that's technically a lie -- those would properly be called "decs" given the underlying meaning of the term.
Handling OGG Theora
FFDShow + the OGG Splitter filter from GuliverKli: Wouldn't that work?
Assuming IE9 just attaches itself as part of WMP and uses DirectShow filters, most people with FFDShow and the OGG Splitter would already be able to play OGG Theora files as soon as they get IE9 installed.
re: Just curious
Opera is not currently hardware accelerated, however all the work in 10.50 and 10.60 in Vega (their drawing engine) is focused about hardware acceleration being plugged in at a later date.
I wouldn't be surprised to see Opera 11 use hardware acceleration on Windows at least. However right now, they don't really need to introduce it, as it's already the fastest browser in pretty much every area, and makes toast of Chrome6..
I wonder how well **ventilated** it is???? and correct me if wrong, but I hear FF 4 will be hardware accelerated...
I hear the sound of virus/ worm writers rubbing their hands in glee.... :P
Video is accelerated but the rest isn't.
It's amazing what happens when there is serious competition. It just goes to show how important for consumers and manufacturers the relevant anti-monopoly committees and EU regulators are.
As soon as you have viable competition then the products get better for everyone.
They just need to speed up the process of discovery, trial and 'punishment' so that smaller players aren't killed off before the decisions are made.
Fast IE is still IE. It still drags around the rotting horse meat that is ActiveX, and it still scares the willies out of me.
«Anonymous coward» hit the nail on the head -
serious competition in the web browser field during the last few years (thanks, Mozilla !) has led to incredible improvements in browser performance. Now, if only we could see the same sort of development in the OS field, presently inhibited by an all-pervasive quasi-monopoly supported by anti-competitive collusion - hopefully Google's Chrome OS will be the impetus needed to bring this about....
A world record?
13 posts in a row about MS, and IE to boot, without any flaming, just fairly objective arguments. Something's not right, has an England victory put everyone in a good mood... probably not, proper nerds despise "a group of overpaid thugs kicking a ball around" don't they?
Did they run the Firefox tests on a typical setup with multiple tabs open which causes it to come to a standstill and needs to be forcibly killed? Firefox is terrible these days.
The only possible explanations I can think of for the dross you just posted is that you either have a complete POS machine, or one which is riddled with crapware (or both).
I have never experienced an issue like this with Firefox. Nor has anybody I know.
Firefox is bloody marvellous these days.
Re: Firefox tests
No, it was a benchmark test. Do you really need that explaining to you?
My worry is that the marketing people will force the developers to optimise their browsers for tests rather than real pages. A bit like how I often wonder whether some ISPs tweak their network so that commonly used speed test sites get priority over real websites.
<a href="http://www.socialnews.biz/tag/Microsoft">Microsoft</a> is making strides. But still, this demo was tuned for the software tested. No wonder the other browsers did not do as well. When MS rolls out the full beta version, that will be the real test.
Can't wait to code for it... really I can't...
I mean, it'll be great having to support 4 iterations of internet explorer, due to the usual glacial uptake - the stats speak for themselves, developers are *still* lumbered with the ie6 legacy, 9 years after it's release.
It took microsoft 6 long years to release internet explorer 7, which is what really screwed things for developers, amongst other obvious cock-ups.
You could go so far as to say Microsoft, due to it's massive browser market share a few years back, has help up rapid adoption of new tech.
As for Flash, it's time it was taken outside and shot, or locked away under quarantine.
Paris, because she should be too.
I do wish media outlets would stop parroting Microsoft's disingenuous statement that they'll support WebM / V8 "if a user has that codec on their machine". The whole point of a web standard is that it works across all browsers. If Microsoft is refusing to install WebM / V8 along with IE9, they are not supporting it. Full stop.
Are MS are doing their best to piss of Opera? If you run the Amazon Shelf demo - in most browsers it says which browser is running in the bottom left hand corner. For Opera though it says "Other browser" - or are they just reading the user agent string wrong?
Opera reports UserAgent differently
for some reason Opera formats the userAgent slightly differently to everyone else, so if you are parsing the browser name out using a lot of the common recipies or have an older browsercap you end up with a less than helpful string of text.
they could have done a UA.contains and checked for Opera explicitly or, like a lot of developers, decided that it's %age of the market didn't really justify extra work and testing :)
Why do they bother?
Really, why is it so important for Microsoft to have a browser?
The days of "Designed for Internet Explorer" are largely confined to the past. Microsoft spends lots of money on a browser when they have more pressing issues, like having a decent mobile OS and actually trying to make money outside of Windows and Office.
I think Safari is a waste of time too. While I don't want all my browser eggs in one basket (Firefox), there is Opera and others. too.
@But still, this demo was tuned for the software tested.
Actually, the tests run just fine in other browsers. Opera does spectacularly well on them, Firefox 3.6 does decent on them, Minefield a lot better, Chrome performs them quite nicely except for the Asteroids demo, which it runs abysmally.
And these are good things, they give the other browsers makers something to go "wait, why is IE better than we are at these?" at, which will further improve them as applications.
So... the cynical part of me wonders: will this break the workarounds I've made to get Canvas alternatives to work in IE?
Simples: They just don't need to. They still have a commanding market share towering over Firefox and the speed at which they're losing market share dictates how swiftly they develop their own product. Firefox is over 5 years old now and whilst it has gained a good market share, IE is still the daddy by a long mark, and will continue to be for a long time yet.
There are plenty of office intranets and corporate/industrial systems based on IE that will keep its market share high for years to come.
JIT will come to IE eventually, but they're not in a panic to roll it out like Opera were because they don't need to. They're not as fast as Opera or Chrome, nor support as many standards, but their product works and they make money. They'll continue to hack away at IE over the next few years and manage to complete what Opera, Apple or Google do in a year, but that's fine by them.
You can be sure that if their market share suddenly dropped and corporations started developing their next generation intranets and control centres for Opera, Chrome or Firefox, they'd soon get into the same gear as those companies and we'd probably get a major new version every year with massive improvements in performance and features.
Now, cue a Firefox retard to start foaming at the mouth and compare the combined market share of all versions of Firefox ever made with one specific version of IE, possibly even one that doesn't even have the largest slice of the IE pie. IE6 lives folks! And 7! And 8! And soon 9! It's alive! Aliiiive!
The pioneers of JIT compiling?
You might want to let John McCarthy know, since when he implemented and documented it in 1960 he seemed to think it was a new idea. Even the ACM are fooled. Ditto those hacks at Sun who managed to steal it from Microsoft and put it in Java three years before .NET was released.