OSNews

Microsoft are focusing on more than just raw JS execution speed as most of the browser’s time is spent outside JS doing DOM and layout. In this respect, IE9 is faster as everything is hardware accelerated right down to the text.

IE9 is neither released, nor finished but it shows an approach that is ahead of the other vendors. I doubt it will last though, hardware acceleration is planned to arrive with the other browsers who will still ship first.

Why should one need hardware acceleration in order to display text and pictures, which the current Web essentially is from a display point of view ? I know that current code is unoptimized, but I didn't know that it was *that* bad...

Edited 2010-06-26 13:22 UTC

It’s not about _need_, it’s about availability. If a GPU is there and it’s “only 14x faster” than the CPU, then why not use it, for everything; even text—which is a whole lot more taxing than most give it credit.

Yes, that's called Aero and Windows Presentation Framework.

Seriously, even after all these anti-trust sanctions, people still think MS is using hidden APIs for non-core OS functionality. The APIs are openly documented, they - legally - cannot use hidden ones.

Why doesn't Microsoft just write the browser for them? Seriously? A system-wide canvas library? What on earth?

Microsoft does the best that they could, Direct2D is a relatively simple API (though still a COM API, but ugh), but simple nonetheless.

HW Accelerated canvas drawing is simply a consequence of the entire renderer being written in Direct2D.

What a silly conspiracy theory.

The hyped hardware acceleration in IE9 is nothing more than IE changing from using the GDI scanline renderer to the Direct2D renderer they added in Windows 7. This is a public API and so any competitor can also use this for their rendering if they so desire.

Just like any other modern windowing system, Windows Vista and Windows 7 stores each window in a texture on the GPU, regardless of what technology you use for the rendering. The windowing system then offers a series of different technologies to fill that texture with contents. In Windows those are GDI, DirectDraw, Direct3D, Direct2D and OpenGL.

What we are talking about here is simply a new interface that is more compatible with the way a modern GPU works. The original GDI graphics API makes some assumptions about the graphics card that isn't true anymore and therefore virtually everything in GDI has been running in software. Microsoft gave up on accelerating it and instead wrote Direct2D and now are bragging how fast IE gets if they use that instead.

If you wonder what is wrong with GDI, then its small subtle things like being able to render directly to the screen (which doesn't make sense when your display window manager does that) and the entire way bitmaps were designed.

Then you design YOUR UI with wrapper code that is able to reference two different underlying libraries and rely on build scripts to know which to compile into the application. There are ways to do this on the developers' end and once the work is done it's done.

Hopefully you will have designed it with the future in mind and the ability to add to the API when needed and not to have to rewrite large chunks as things change underneath.

I'm sorry, but what the hell *are* you ranting about? You seem to be upset that Microsoft has APIs at all, as if this is some cardinal sin. Yes, it's true that if you use an API to write your program, you will be stuck with that API if you are not careful when constructing your program. This is not news and it's just as true on Linux, OS X, BeOS, OpenBSD, etc. as it is on Windows. If you use Xlib, then you will be stuck with Xlib. Who cares?

And also, why should MS write all of their programs to run on every GD operating system out there? It's not in their interest and honestly, it's a waste of time.

Okay. It's just that when I hear about browser hardware acceleration, I see in the future Canvas-powered websites which use the GPU for stupid things, just like Flash currently does. I see laptop battery life continue to go down. I see a Vista effect where developers require ultra high power from the user's machine because their own machine is powerful. I see users which don't have fast 3D acceleration being penalized because they use open-source drivers or AMD's crappy proprietary drivers, or just because their old laptop includes a crappy intel GMA chipset.

If you make the feature available, you can safely assume that people will use it. Hardware acceleration in browsers is just like composited desktops : it's fine as long as it's optional and people can survive without it. Sadly, the web is not a desktop, where disabling compositing will just make windows look ugly. If developers require hardware acceleration and it's not available, the site won't work. That's the kind of things which I'm fearing.

Edited 2010-06-26 14:46 UTC

-

Edited 2010-06-28 12:36 UTC

Using GPU accel isn't straight performance win: http://blogs.gnome.org/otte/2010/06/26/fun-with-benchmarks/

If you look at the HTML behind Web 2.0 applications, you'll find that they're not simple anymore, and the complexity can bog down a software-based renderer. Being able to offload the layout processing to the GPU helps a lot.

FireFox has hw accel on Linux since Version 2.0

It’s not a browser yet (IE9pp doesn’t even have a back button), it’s just a Trident instance. Microsoft still have every capability to totally screw up IE9 with the GUI. Almost nothing could be worse than IE8, but we will see…

Yes... let's hope they stay focused on end features and functionality and also manage to get this thing completed in some reasonable amount of time.

I haven't used IE regularly(except for unit testing at work, or in the past where sites were designed for IE) for at least a decade. Maybe MS can do something right here that fits nicely with Win 7... or they can drag their feet or even cancel the project... that wouldn't surprise me either.

MS5 was at the time, by far the best browser for Windows, especially compared to new version of Netscape that had just escaped from some mental institution.

Which is why they stuck with it for far too many years.

I tried using 3.7a but the 64-bit version for Windows is still buggy. Screen goes black when you hover over certain parts of pages. Otherwise it looks nice (FF overall, that is).

HW acceleration doesn't work in Chrome (or if it does, very badly). Not that I/we NEED it yet. Anyway, when FF stabilizes a little more I will check it out again, it is looking pretty decent.

Difference between doing standards correctly, versus just doing standards.

Check the CSS tests on all of the major browsers, it offers a much clearer picture of the actual situation.

WebKit (the darling of the rendering engines) is particularly notorious for buggy or lacking implementations.

CSS 3 selectors fully supported in IE9 http://www.quirksmode.org/blog/archives/2010/06/ies_big_leap_fo.htm...

I'm actually really happy to see Microsoft having improved in this area so much. It would be nice if it is a precursor to their entire organizational mindset.

And no, there is no sarcasm in this. I'm genuinely glad.

Firefox 3.7a5 already supports Direct2D on windows.

Any idea if they are going to support HW acceleration under Linux (or other OpenGL-supported platforms) and if so, when? Or will this again be a Windows-only feature?

It's just a simple app to make use of the rendering engine. It is the rendering engine we are talking about here - not the the functionality of the application as a whole.

You CAN navigate to other sites - there is a primitive pop up text box where you can enter URLs. But this isn't meant to be a complete and useful application, it is meant to showcase the most important part of the browser.

I would think t3h int4rweb would know by now that there are two main branches of English today (en-gb and en-us).

And everyone knows that GB-ers drive their cars on the wrong side of the road, too.

It is just an odd notation when referring to things as impersonal as large corporations. The singular collective usage seems to be more frequent than plural collective in most British books and publications I read. More interestingly, I've noticed OSNews will sometimes vary its usage within a given article, which makes it stand out more.

Typically a collective plural verb seems to be best applied when trying to emphasize the individuals within the collective.

Classic British English or some new bastardized form?

Try finding an old British headline that reads Germany have invaded Poland.

Single organizations use the singular form.

The use of "Microsoft have" rather than "Microsoft has" is because of the way collective nouns are treated. Read this: http://en.wikipedia.org/wiki/American_and_British_English_differenc...

The logic is that you wouldn't refer to Microsoft saying, "He has released IE9PP3." Instead, you would say, "They have released IE9PP3." The British take it one step further. They stick to the plural sense at all time, out of consistency:
"[The people who develop IE] have released IE9PP3."
"[They] have released IE9PP3."
"[Microsoft] have released IE9PP3."

Or in short, every time you refer to something that describes several people, even when the noun is singular (band, corporation, team, gang, fanclub, OSNews), you can treat it as plural. So, OSNews have done well. Kroc has done well.

______________
EDIT: I accidentally wrote "He have released IE9PP3." in the example line. So much for confusion! Haha.

Edited 2010-06-26 17:17 UTC

The neuter form is one of the strange parts of English. I suspect "they" is often used since most people seem to prefer it to "it" or the "generic masculine." Hence, people will speak of a singular person using "they" if they do not wish to specify a gender -- even though that is incorrect.

Probably the best usage would be "Microsoft has announced... It will make the updates available..."

Interestingly, surveying the BBC, they are inconsistent in usage. I found some variation even within a single piece. Sloppy.

I do see your point. "Microsoft Corporation" is a single legal entity or juridical individual. So in that context, I think it is even better to refer to it in a singular sense.

It is these types of ambiguities that have warranted the use of the singular form, which is also a form of consistency. The rule of thumb I use myself is: "if I would replace the noun with a description of what it conveys, is the description singular or plural?"
Example:
"[The legal entity] has issued more shares."
"Microsoft has issued more shares." (singular)
"[The group of IE devs] have released IE9PP3."
"Microsoft have released IE9PP3." (plural)

The use of a singular sense has even expanded to plural nouns, such as "data", the plural form of "datum". Merriam-Webster provide a clear explanation with "usage": http://www.merriam-webster.com/dictionary/data

Dude, you're missing the entire point!

No you're missing an English lesson.

Single entities always use the singular form.

No sentence should begin with Microsoft have.

Edited 2010-06-29 04:53 UTC

So you're actually saying Kroc needs a lesson in English, and I should be in the same class?

Wow! Amazing.

10 year old operating system. There's more than IE that needs upgrading.

It's only 2 years old if you count SP3 and 6 years old if SP2 (a major rework). It works very well and upgrade is not an option for many reasons. Sorry, XP is here to stay.

The problem is IE6, which is still hugely popular in some countries or organizations and its quirks are still shaping the web. I hoped IE9 would nudge at least some of IE6 users to use a standard compliant browser. Apparently that wasn't the goal for Microsoft (not really a surprise, why to fight a won war if there are challenges elsewhere).

Yes and no. My website still doesn’t work, but it has progressed. According to the HTML5 spec, the <html>, <head> and <body> tags are optional, and the <body> should be implied when the browser hits a tag that isn’t specific to <head>; IE doesn’t do this right, and in pp3 it does it, but incorrectly puts the HTML5 <header> into the <head> and not the <body>.

Before, I was not convinced they were serious about standards, especially since their announcement about H.264 was so vile (the wording was dishonest to the last period), but this new preview shows firm, measurable commitment to HTML5 and standards. The H.264 thing is a whole ’other battle that goes beyond the scope of just the web and I don’t see Microsoft moving away from H.264 any more than Apple would. For now, WebM will be available for use if installed by the user, and that will be increasingly likely given Google/YouTube.

I’ve been using HTML5 since 2008 and I’ve been advocating to people to get learning it _now_, because by the time IE9 comes, those developers that have sat around waiting for IE will already be left in the dust by developers like me who have been making the most of HTML5 for years already in better browsers. Microsoft will have a struggle getting comfortable IE developers to get a clue.

I firmly believe that HTML5 will be the new base standard by which things get done on the web and IE9 could therefore be a runaway hit.

I don't really like that aspect of the spec. I prefer it to be done explicitly. :/

Do you complain that the image tag allows multiple codecs? If it stuck with one when it was invented, then we’d all still be using XBMs for images.

No, just not specifying <body> and <html>. That sort of thing. I can't help it, I am a programmer and I like to be explicit.

HTML itself is a simplified form of SGML. *shrug* I’m just pedantic by leaving them out, I doubt most developers will.

I am not sure I understand your reply. Pedantic?

In the general world of development, I know that few developers document, design, write to save memory, write to garner speed through efficiency, etc. It irks me that no one seems to care. All the memory, fast chips, the fancy interpreters and byte compilers, etc., seem to have made everyone sort of soft.

Sorry, not trying to poke at you. It's just one of those things with me... like when someone touches my screen.

I'm going to defend IE a little bit here I must be ill.

iPad Safari: 127 + 7
iOS 4 Safari: 185 + 7
Firefox 3.6.3: 139 + 4 (may not be latest I rarely run it so it may not have updated yet)
Opera 10.52: 129 + 4
Safari 5: 208 + 10

So out of those browsers it is only the latest versions of Safari (I don't have chrome installed) that are more than 2* what IE gets. Yes I should be trying out nightly builds of everything but I'm using what is directly at hand.

Edited 2010-06-27 11:07 UTC

Funny thing

I get 1 FPS with 4 satellites using Google Chrome 5.0.375.70 on Windows Server 2008. With Firefox 3.6.4 on Windows I get 11-13 FPS. Looks like the hardware acceleration in Windows is rather spotty.

A browser prefix is the correct way to introduce new proposed features into CSS.

The browser prefix is a way to test the implementation practically, to see if it fulfills the need it was created for. For example, Webkit & Firefox (Gecko) implement background gradients with different syntax. I think it was first proposed and implemented by the webkit team, but the Firefox team decided that they preferred to implement it in a slightly different and simpler way.

After this is done and they get feedback, they discuss and decide which is the best implementation, and once it it finalized the prefix is removed.

Which is what was done in the latest webkit version, where the -wekit- prefix is no longer necessary to use border radius.

The prefix is used so that it is clear that this is not the final version, and to use with care.

BTW, just so you know, all the rendering engines have a prefix not just webkit.

Webkit: -webkit-
Gecko: -moz-
Opera: -o-

Don't mean to burst your bubble but I read somewhere that WP7 will have a variant of the IE7 rendering engine.

Link: http://www.neowin.net/news/windows-phone-7-browser-is-based-on-inte...

But yeah maybe future version of WP may benefit from all this.

Yeah I read about that somewhere, but I don't think it's been confirmed yet.

But it just makes no sense to me. Why go to all this trouble to prepare IE9, and then ship Windows Phone 7 a month or two early and not include IE9.

IE7 is slow, and non compliant, and will get loads of bad reviews. If MS is stupid enough to release Windows Phone 7 with an ancient browser, I'm pretty sure they will get a lot of bad reviews, and they will deserve each one.

I'm definitely not looking forward to start debugging even mobile browers for IE bugs :-(

One more thing. Right now web designers & developers need to spend time fixing bugs in IE7 because it has such a large share.

But I can't see that happening for mobile sites, where the good mobile browsers (webkit & opera) are already compliant and have a majority of the share.

The result will be that users browsing from IE7 on WinPhone7 will see a large majority of mobile sites broken, and they will rightly blame MS for that.