HTML5 Web Standards Tests for IE9 | Hacker News Reader

HTML5 Web Standards Tests for IE9

samples.msdn.microsoft.com

48 pointsnikz37 comments

This is a very carefully selected set of test cases that only cover a small subset of HTML5. For example, most of their CSS test cases focus on border-radius (mostly obscure corner cases), but what about shadows, gradients, independent transparency, etc? I appreciate Microsoft's efforts to support HTML5 in their next release of IE but this page is simply ridiculous.

It seems like they picked a few features of HTML5 and decided to support only those but support them well down to the most obscure corner cases, as opposed to the "let's do everything now and then we will fix it later if it's not working properly" approach of open source browsers.

The main problem is that IE release cycle is considerably slower than other browsers'. If they only implement a few features, we have to wait several years to other features, if they implemented all features, we would have to wait several years for Microsoft to fix the bugs. The web these days is just too fast paced for that. It would be a lot better if they switched to a more frequent release cycle.

So the countermove is for other browser vendors to go through the list and fix these "small" cases. This takes away any of IE9's bragging rights and we'll have 5 browsers that pass on this exact subset of tests. Microsoft is forced to be more compatible and html5 arrives on the scene sooner. IE9 takes us to 2014 or 2015. Let's push Microsoft to do even better.

It won't matter. For them to have constructed such a biased test shows how much in denial they are. It's pretty depressing actually.

Release cycles matter: just randomly clicking through the tests, most of the reported failures for Safari or Chrome already work using WebKit or Chromium dev channel - both of which are likely to make it into official releases before IE9 comes out.

Here's a response to the web browser tests recently performed on the Internet Explorer Testing Center:

http://www.freeciv.net/yet-another-html5-browser-test/

They missed the most important categories:

Works on millions of existing Windows XP installs:

  FAIL  Internet Explorer 9
  PASS  Mozilla Firefox 3.6.3	
  PASS  Opera 10.52	
  PASS  Apple Safari 4.05	
  PASS  Google Chrome 4.1

Supports CANVAS:

  FAIL  Internet Explorer 9
  PASS  Mozilla Firefox 3.6.3	
  PASS  Opera 10.52	
  PASS  Apple Safari 4.05	
  PASS  Google Chrome 4.1

While lack of canvas support is definitely a noteworthy omission, I say people running an OS almost a decade old can't really complain when the newest software doesn't work.

MSIE 9 uses hardware acceleration for faster rendering which involves APIs and interfaces which are not found in Windows XP. If you want the world to move forward, you will occationally have to leave someone behind.

That time has long come for Windows XP.

Since DirectX 10 has been demonstrated as portable to XP, the lack of acceleration is not a technical or administrative problem - it's a financial decision by Microsoft to force people to upgrade.

But this will backfire on them because it's MUCH easier and cheaper to use another browser than to upgrade your OS.

By the very same logic there is no reason (apart from investing time & money) that Firefox shouldn't be able to run on 16-bit systems. But you don't see anyone complaining about that.

You have to draw the line somewhere. XP is old. There are a lot of decent OSes and upgrade-options out there. I see no problem ignoring XP for any new, major development projects, especially if new APIs will need to be developed in the process.

There are millions, possibly a billion XP installs out there (I'm one of them). They can run all the newest software still, it's 99% the same API as Vista and W7. Plus XP is still faster than Vista or W7, uses less memory and FAR less disk space to do the same functionality.

Don't confuse the need to sell more software with the previous version not actually being outdated. There's a huge difference between Windows 3.1 being extinct and XP.

Good for you. You are obviously a power-user and any restriction placed on MSIE will not really affect you as I assume you are using some other browser anyway.

The fact remains that XP is almost a decade old, and just because your OEM paid $20 for XP 9 years ago, does not entitle you to having free support, patches and new APIs backported to XP forever.

Microsoft has officially ended regular support for Windows XP and are now only providing patches/bug-fixes while on extended support until 2014.

Windows XP might work for you, even under these conditions, but don't delude yourself into thinking it isn't outdated just because it works. When the vendor stops supporting it, it is outdated.

Unless you are running on a machine with less than 1GB ram, I see no reason to cling on to XP.

I'm deeply suspicious of any page provided by any vendor in which all of their tests pass and everyone else is missing some.

I think it's really great IE9 is doing HTML5/SVG/CSS3/etc full steam, but I'd much prefer to see where they are failing too.

http://caniuse.com is much more trustworthy and has broader coverage.

Hmm.. I find it interesting that they have kept IE6, FF 3, Chrome 3, Opera 9.6 and Safari 3 in far-past category. Doesn't IE6 is Far-far past then?

"Google Chrome 4.1", sure let's test the development version of IE, but not the dev version of chromium (chrome).

No mention of workers, canvas, audio, video, <device>, web databases, or advances in javascript speed improvements. Just talk about separate specs, CSS3 and SVG.

Anther waste of time, presented by Microsoft...

Although I love the effort put forth by MS for even thinking about HTML5 but the webpage is silly. Firstly, why are they even testing other browsers. Should I thank MS for testing out everyone else? Secondly, as mentioned already, someone should send them a memo on <canvas>. Lastly, the first thing I saw was that border radius with dotted style is not supported in chrome so I tested it and it works flawlessly. This was the first and only test I conducted from their data and it was enough for me to dismiss this article.

No mention of canvas. Doesn't the IE team realize that developers are reading these blog posts?

Although the tests only cover a tiny subset of HTML5 cases that IE9 supports, I like the approach. If each browser had a test suite that they passed 100% and could run against the other browsers, it would be an interesting way to move towards compatibility.

Better yet, make automated test cases part of the standard.

The pages used for this actually are test pages they developed with the W3C, which is almost making them part of the spec.

Only Microsoft would be able to come up with test results that show Internet Explorer as the perfectly standards compliant browser while all the others aren't.

Their test for "Set selectionStart and selectionEnd on a text field" tests that selection start and end can be set which should pass in Safari but their test fails because they are asserting that window.getSelection().anchorNode == null after they've set the selection start and end points. This doesn't make sense to me and I don't get that from my reading of the spec. Am I missing something?

On most compatibility tests, IE9 always seems to come up way short. It looks really good on their tests. Where does IE9 really stand? What are the chances of getting <canvas> support? IE9 could be the key to taking the web to the next level.

In comparison on the Acid3 test...

  IE9 Platform Preview 68/100
  Firefox 3.6.3 94/100
  Opera 10 100/100
  Safari 4 100/100
  Chrome 4 100/100

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

Don't worry. It will default to 'compatibility mode' and there won't be any way to force the browser to use the new engine. It will still be full of security wholes, take forever to parse simple javascript and no doubt offer an overall inferior browsing experience. I don't think Mozilla/Apple/Google have anything to worry about.

As a web developer, I'm quite happy with the way Firefox's rendering capabilities are progressing, despite what these phony tests might suggest. And seriously, how can they compete with Firefox's 'AwesomeBar'? I seriously couldn't live without it.

I just hope these make believe test pages don't fool the end users :(

I just hope these make believe test pages don't fool the end users :(

While these tests are certainly biased, they do show that MSIE is at least progressing now (as opposed to standing completely still for years and years) with regards to standards compliance and HTML5.

And really... Would you want end users on MSIE6/7/8 or MSIE 9? For the group of users with no technical interest, sadly that is usually the choice you have.

The problem is that it isn't a choice of MSIE6/7/8 or MSIE9, it is MSIE6/7/8 AND 9. Maybe I've just been burnt one too many times, but they really should just give up. Sure, it might pass a few tests here and there, but MSIE9 will just be ANOTHER buggy web browser whose idiosyncrasies web designers will have to consider when they build someone a web site.

Anyone can make software to pass standardized test, but take my word for it, MSIE9 is just another MS browser full of wildly inconsistent renderings of simple HTML, a half baked Javascript implementation and fruity vendor specific extensions that developers will have to cater for.

MSIE6/7 and 8 aren't going anywhere soon.

The problem is that it isn't a choice of MSIE6/7/8 or MSIE9, it is MSIE6/7/8 AND 9.

While it hasn't always been that way, at least lately (since MSIE7 and 8) Microsoft has been working hard to get rid of the older versions as well and the newer versions (while certainly not perfect) most definitly improves on the shitty standards support found in MSIE 6.

In MS's latest technet newsletter there is even a link to a paper on virtualizing older versions of MSIE so you can upgrade to something modern while retaining compatibility with older legacy apps requiring MSIE 6. [1]

So while MS can be blamed for producing MSIE 6, the blame for that MSIE is still around lies entirely in corporate IT and them not using the options they most definitely have for getting rid of it for mainstream use.

[1] http://www.dabcc.com/documentlibrary/file/Solutions_for_Virt...

I strongly disagree with this HTML5 standards test from Microsoft. They have included only a small subset of the HTML5 standard, and have excluded important parts, in particular the HTML5 canvas element. As a result, Internet Explorer 9 will lead to an even more fragmented web where browsers interpret standards differently.

This is why we started the Internet Explorer 9 HTML5 canvas campaign: http://freeciv.net/internet-explorer-html5-canvas-campaign.j...

Fine, it's bias.... but I like the attitude and love the fact that a primary competing factor between vendors is how compatible with the HTML5 umbrella they are! It's win all round.

I'm so very confused by the chosen subset of html5 and css3, it looks like they implemented a few things and ran like 20 tests on those few things that other browsers have yet to adopt while the actual things that other browsers are accepting aren't even shown.

I really hope they IE team continues to work on putting html5 and css3 in their browser and if it turns out to be the most standards compliant browser kudos to them but this was disappointing.

MSFT has a history of spouting out bs claims about its own browser superiority over others. They have recently going after Google Chrome at full stream[1]. They obviously feels threatened by the new comer. As for me, I can't wait for them to die.

[1] http://geektechnica.com/2010/04/microsoft-continues-its-trad...

> MSFT has a history of spouting out bs claims about its own browser superiority over others

who doesn't?

Yeah, but at least most other browser vendors merely distort the truth; here Microsoft purports to do an HTML5 feature comparison, and they don't even get around to mentioning <canvas>! That's flat-out lying about what it means to be an HTML5 Web browser.

Sigh... part of me really hoped that IE 9 would finally give Windows a standards-compliant built in Web client. But perhaps something about the culture surrounding the IE team is so fundamentally broken that we'll never see such a thing short of a major Microsoft reorganization.

> Yeah, but at least most other browser vendors merely distort the truth

promotion, marketing, ads, they are not meant to tell the truth. If you want truth, use your own judgement.

"The table at the bottom of this page provides details and links to each of the new test cases we submitted to each appropriate W3C working group"

Translation: cherry-picked for maximum green in their own column.

ill believe it when I see it.

Reply on news.ycombinator.com