14 replies to this topic

[alphadog]

http://www.scirra.co...rmance-analysis

[fireside]

Interesting. I still won't use IE9, though.
From what I've tried, I don't think WebGL is quite ready for prime time. It doesn't even work for me with Chrome. With Firefox, it locks up once in a while, but works decent most of the time.

[touch_the_sky]

Quote

I still won't use IE

I only have one thing to say - THANK YOU.

Unfortunately, there is 'couple' more users in the world who wouldn't agree for one reason or another (people have some mercy and stop using IE! -> Future... ;);))

Article seems interesting though, I hope I'll be able to clear my decks at work this week so I can dig deeper into this

[fireside]

Keeping old browsers around has become dangerous. With Microsoft, they have been dragged kicking and screaming into open standards and I doubt they will ever be standards oriented. The latest is WebGL. They won't support it unless they have huge losses in market share. It will always be something like that for them. When they actually decide to support one, they do a good job eventually, though.

[alphadog]

touch_the_sky said:

Future... ;))

Well, I think that article is wrong and stems from the fact that web designers/web developers seem to be one of the worst groups in terms of wanting to play with shiny new balls of foil. You can see that when someone says something like "IE7 is holding back the web". Really? The whole web is held back by IE7? The *whole* web?!?

(Or, just your target market can't be bothered to upgrade because your app isn't worth it? :sneaky:)

We're shoving so much into browsers that were creating a fast-moving, pseudo-OS over a base OS. And, like having to support multiple OSes, configuration pains now come into a space where most people are "multi-platform virgins".

But, anyways:

A) Most developers use a client-side framework, so there isn't as much "support" as you think. Pick a good framework, and the cross-browser compatibility is built-in. Also, IE7 is IE7; it isn't going to change radically.

The constant and rapid upgrade cycle of Chrome, and now Firefox, is a new, different kind of irritant... one I consider worse. Once I build for IE7, I know I won't have to fix it, but damn, every once in a while there goes some new whizbangfeatureupdate on Chrome, shoveled out to the masses without pause, lighting up my support center.

Everyone seems obsessed on A), but real comments about seem far between. I guess it doesn't matter, as long as you get HTML6 to play with, huh?

<engage type="escape pod" direction="To The Cloud!">

[touch_the_sky]

It's not quite like that alphadog, you're assuming some things. Web developers want and DO play with shiny new tech. (not to mention that from gamedev point of view it may seem that canvas, webgl, etc. are the only things conisdered 'new balls of foil' - which is not true).

The whole IE holding back web - it doesn't work quite like that, it's not video games, with typical, everyday webdev bread and butter (which usually brings the most money) there is no 'people upgrading browsers cause you did some worthy stuff, etc.'. In the real world you can't build for one specific browser / version, and IE, yep you guessed it, is always the least pleasant browser to work with (differences in rendering between various versions are the biggest too). I don't know a half-decent frontend dev who would have much of an issue with rapid upgrade cycles of Chrome / Firefox either.

Anyways, the article of course was not to be taken dead serious. Which doesn't change the fact that if some skiddie from Syberia made a virus which erased all IEs from the planet - many frontend devs would be grateful ;)

Peace!;)

[fireside]

Microsoft does it's own thing and ignores web standards.

"Microsoft, developers of the Internet Explorer (IE) browser, said that Acid3 does not map to the goal of Internet Explorer 8 and that IE8 would improve only some of the standards being tested by Acid3.[34] IE8 scores 20/100, which is much worse than all relevant competitors in their versions from the test's release, and has some problems with rendering the Acid3 test page. On 18 November 2009, the Internet Explorer team posted a blog entry about the early development of Internet Explorer 9 from the PDC presentation, showing that an internal build of the browser could score 32/100 for the Acid3 test.[35]"

http://en.wikipedia.org/wiki/Acid3

[Achtzehn]

fireside said:

Microsoft does it's own thing and ignores web standards.

"Microsoft, developers of the Internet Explorer (IE) browser, said that Acid3 does not map to the goal of Internet Explorer 8 and that IE8 would improve only some of the standards being tested by Acid3.[34] IE8 scores 20/100, which is much worse than all relevant competitors in their versions from the test's release, and has some problems with rendering the Acid3 test page. On 18 November 2009, the Internet Explorer team posted a blog entry about the early development of Internet Explorer 9 from the PDC presentation, showing that an internal build of the browser could score 32/100 for the Acid3 test.[35]"

http://en.wikipedia.org/wiki/Acid3

Well, a few years ago they could do pretty much everything and set the trend. But time is over. IE will die slowly and I think there's nothing they can do but producing a better product finally. But this is something Microsoft is not able to do so they shall fail ...

[alphadog]

fireside said:

Microsoft does it's own thing and ignores web standards

That's old news.

MS did goof up with IE8, but I think they realized it and have turned that lumbering ship around. Not entirely their fault, as too much "enterprisey" software was built within (and pushing) the boundaries of IE7/8.

IE9 is better and IE10 already shows lots of promise in ECMAScript tests. I think MS was scolded, re-geared and back with a vengeance.

However, their lack of support for WebGL is annoying. If there's one old-MS-style fly in the honey, it's WebGL. You'll have to use the MS-specific JavaScript bindings to Windows Direct graphics APIs, and distribute your app as Windows app not as a true web app. I would like to believe there's a growing swell of complaints that MS will not be able to ignore forever, but only time will tell. Their defense line is unsupportable.

[.oisyn]

Yeah, MS is learning that in these days standards compliance is what is demanded by their customers. They learned that in C++ earlier on, and they're accepting it for IE as well. Why they refuse to support WebGL is beyond me, however.

Also, what I find extremely annoying about MS is their way-too-long update cycle and their refusion to make an effort and fix simple issues. For example, today I was hit by a bug in the VC++ compiler that has existed for ages (VS 2003 was the earliest repro case I could find on the interwebz).

struct __declspec(align(16)) struct S { };

void Foo(S s) { } // error: formal parameter 's' won't be aligned.

Ok, fine, I'm willing to accept that. But then:

template<class T> void Bar(T t)
{
    std::cout << &t << std::endl;
}

int main()
{
    S s;
    Bar(s); // no error
}

Yes, that's right. Template instantiations are not covered by the alignment check. It compiles fine, and the 't' in Bar<S> will actually NOT be aligned. Your code will crash if it just so happened to be a SSE type and you're going to do an aligned operation on it.

The bug has been reported, but the VC++ team flagged it as 'won't fix', 'because it doesn't meet the triage bar'. And even if they do fix it, you usually have to wait 2 product cycles (if you report an issue with 2010 that they will fix, you won't see the fix earlier than the release after VS 2012).

Or, on Windows Phone, they use a fork of IE9 now. That's fine, way better than the old IE7 renderer. But unfortunately it lacks support for anchors in URLs. That's pretty damn annoying when browsing the web, but don't expect that to be fixed any time soon.

MS needs to drop their old behemoth of an update cycle and start being more flexible.

[fireside]

Quote

That's old news.

I guess you are right, IE9 passed the Acid3 with 95/100. I haven't kept up with their changes. I think the main reason was that they were losing market share, though. If they get it back, they will return to their old ways and try to lock in customers. That's their schtick.
They also don't even offer IE9 for XP, even though they still support XP. Just like their DX10 nonsense. They're no worse than Sony, Nintendo, or Apple. A commercial company is going to try to lock in profits.

[alphadog]

fireside said:

They also don't even offer IE9 for XP

That's a common complaint, but then IMO, I don't think it's unfair that they don't offer a new product for an OS that came out in 2001 and that, at this point, should be phased out of the digital ecosystem.

[Nick]

fireside said:

From what I've tried, I don't think WebGL is quite ready for prime time. It doesn't even work for me with Chrome. With Firefox, it locks up once in a while, but works decent most of the time.

Feel free to file a detailed bug report and I'll see what I can do. :blush:

[TheNut]

Hmm, that list is growing Nick. What's going on :lol:

I don't find it surprising that WebGL is leagues better than the 2D canvas renderer. Even if they could level the playing field, having full control over how things are cached and rendered will always pull out with better performance than some general implementation. It's part of the reason I don't bother with the 2D canvas even for simple 2D games. The visual quality you can pull off with shaders just blows it out of the water.

.oisyn said:

Why they refuse to support WebGL is beyond me, however.

I think it has to do with the MS mentality that they need to control everything. Even their other products like WP7 and Silverlight lack sufficient D3D control. Ultimately the demand and future of WebGL will dictate terms, but I think if MS opens up the D3D API using JS bindings then developers will be content enough with that. Supposedly Windows 8 has something in the works, but I haven't played with their preview yet so I dunno.

[Nick]

TheNut, on 26 November 2011 - 01:40 PM, said:

Hmm, that list is growing Nick. What's going on

It's actually going really well. We've just officially achieved OpenGL ES 2.0 conformance (http://www.khronos.o...ormant-products) and bumped the version number to 1.0.


The issues list is mainly growing because there are lots of feature requests which are not a priority just yet. In particular the demoscene people are pushing the limits (yes, their shaders are more complex than those in Crysis)! Also the project is now being used for validating shaders on Android and possibly other things I'm not aware of. Furthermore some bugs appear to be caused by DX9 driver issues so there might not be a lot we can do about those. Anyhow, overall the project's robustness and performance is steadily improving and it's used successfully by ever more people.