Internet Explorer 9: Platform Preview | Hacker News Reader

Internet Explorer 9: Platform Preview

ie.microsoft.com

86 pointssid080 comments

Certainly looks as though Microsoft are finally willing to try to take on the competition (they are actually flaunting the progress of ACID3, and posting benchmarks).

In fact, its interesting that their new engine is so fast. Its going to get very interesting, especially since personally, I couldn't give a damn about supporting unfinalised standards the ACID3 tests (in fact, one of the ACID3 tests can't be passed 100% of the time anyway).

Speed is important, and since a lot of applications use the IE engine, maybe this will speed up some other applications as well.

Btw, I can confirm their sunspider benchmark.. Their preview does run faster than Firefox on my system (although, unsure if my addons are having an impact on that, or external causes). As its just a preview, it should only be taken with a grain of salt anyway..

I don't think anyone has ever questioned Microsoft's ability to take on the competition - or even beat the competition as they did in the Netscape wars. Rather, I think we're permanently jaded against the IE browser not because of its technical merits or flaws, but because they let it go stagnant.

Knowing that people using Windows 2000 are stuck with IE6 - despite the fact that their machine still suits their needs - is why we need to continue to push people away from IE. Even if IE9 is amazing.

Firefox 3.6 STILL works on Windows 2000 (http://www.mozilla.com/en-US/firefox/system-requirements.htm...). Until Microsoft releases a browser we know isn't going to put us in the "I can't upgrade past IE6" situations that we hear about, I'll proudly support everything that isn't Internet Explorer.

(Note: Has anyone ever tried to submit a bug report to Microsoft? It is impossible! I even put together http://crashie8.com/ as a proof of concept, and I still can't get anyone to acknowledge the problem!)

I think we should all take a moment to thank Google for making Chrome and get all browser developers to get out of their collective ass and do something to fix their broken browsers. The amount of improvement and works going on the top 4 browsers since Chrome was released is really amazing. IE 7-8 were half-assed job to counter Firefox's popularity.

Even Firefox is getting their act together and their next release is very much inspired by chrome.

Agreed. But I find it amazing that a browser that has roughly a 5% user share has had so much impact. I surmise that it's really the weight of the company behind Chrome that has scared its competitors more than the adoption of the browser itself.

The marketshare in the general public (the 5%) is almost inconsequential. The general public doesn't build websites for a living. The marketshare amongst tech professionals is more significant, and there Chrome's numbers are much higher. More so, Chrome's marketshare, even among the general public, is growing quickly, only a fool would ignore the significance of that.

Also, marketshare is not necessarily the primary motivator for every developer. A lot of the Firefox developers are motivated to put the best browser possible out there, regardless of marketshare, the fact that there is a leaner, meaner, faster browser even exists (even if the marketshare was less than 1%) might be enough motivation for them to innovate.

I seem to remember scaring the competition into movement was one of their stated goals. They wanted their sites to be more usable.

I'm actually amazed at how quickly Chrome gobbled up 5% of the market.

I think a lot of us question "Microsoft's ability to take on the competition." Maybe not from a technical point of view, but certainly from a bureaucratic, entitlement and motivational one.

>> Rather, I think we're permanently jaded against the IE browser not because of its technical merits or flaws, but because they let it go stagnant.

Partly - I think we are jaded because of technical flaws as well, even in their later versions. IE8 was supposed to be their modern browser, and it still had no canvas support and horrible JS performance.

Otherwise, I agree with you - even if IE9 is a huge hit and awesome, a lot of developers supporting IE won't really have a better life.

I think there's more to it than that. Somehow Google managed to build a top tier, highly standards compliant browser "from scratch" in a short matter of time. Much less time than IE development has not been stagnant, for example. And yet the work and the decisions of the IE team continue to disappoint. They have made some good choices and improved IE a great deal, but they are catching up slower than the other browsers (FF, Chrome, Safari, Opera) are pulling away.

More so, there doesn't appear to be any significant sense of urgency in the IE team related to just how far behind they are. Firefox has revolutionized its Javascript engine, massively increased standards compliance, and added support for next generation html5 and css3 features. We'll be lucky if IE9 manages to support css 2.1 correctly, let alone implements html5, css3, or comes out with a competitive javascript engine.

Microsoft seems to be pretty complacent in regards to IE because it still dominates marketshare. However, the decline of Myspace shows just how quickly marketshare in the online world can change.

Hardly from scratch at all. The base work was done by the KDE team as KHTML many years ago and then Apple put the finishing 50% in with WebKit. By the time Google came along, WebKit (aka Safari) was already a force in the market. Google's biggest contribution thus far to the current browser wars was the V8 engine, which significantly raised the bar for JavaScript performance.

You may have noticed the quotes I put around "from scratch" in an effort to forestall responses just such as yours. Yes, I realize that they built on a substantial amount of existing work, but every project does. The point is that they did nothing that nobody else could do, there's nothing stopping MS from using WebKit in IE9. The fact that a relatively small number of total developer-hours compared to the efforts having been expended for development of IE7 and IE8 can translate into a top-tier web browser is significant.

The Internet is such a terrible medium for communications subtleties. I agree though... but imagine how that would look from a PR perspective. Microsoft is, essentially, the anti-thesis of open source.

It's not so much the stagnation, or even the technical flaws, it's the lock-in. That's the ultimate source of most people's problems with IE, it's the tight integration with IE6 and enterprise apps. That's why dealing with IE6's technical flaws is still a problem, it's difficult for users to switch due to lock-in. That's why MS's browser development stagnated, they had obtained the browser share and got lazy. And that's why standards are important.

And the solution to that is pushing out a security update that turns IE6 into something like Mozilla Prism, where you have to set it up for each individual enterprise site that relies on it, and can't use it for "free-range" browsing. Pack that update with an alternate-browser-download tool as well, with a system-tray alert saying "Your computer will no longer be able to connect to the World Wide Web until you have selected a browser to install" every hour or so. I think they just don't want to start up the pre-XP updating machinery again.

Why would they? Windows 2000 is ten years old, that's longer than the security patch support period of RedHat/CentOS, Ubuntu LTS, Debian, Solaris (and Windows 2000).

Why would or should they go back to developing a free browser for an out of support system which will mess up the user's system as a deliberate side effect, for the benefit of some arbitrary web developers?

It's not really support for their old browsers, though—it's User Experience support for their new ones. Banishing IE6 allows all websites more freedom to use modern formats and specs. Modern formats and specs give them something to compete on in IE9.

To put it another way—if there was an update they could release to Windows 2000, at this point, that people would install voluntarily, but would make more people buy Windows 7 as an effect, don't you think Microsoft would be all over that?

  I couldn't give a damn about supporting unfinalised
  standards the ACID3 tests

MS too sometimes speaks of Acid3 in dismissive terms as having little relevancy. Let's see and then decide by yourself, is this stuff important:

  DOM2 Core
  DOM2 Events
  DOM2 HTML
  DOM2 Range
  DOM2 Style (getComputedStyle, …)
  DOM2 Traversal (NodeIterator, TreeWalker)
  DOM2 Views (defaultView)
  ECMAScript
  HTML4 (<object>, <iframe>, …)
  HTTP (Content-Type, 404, …)
  Media Queries
  Selectors (:lang, :nth-child(), combinators, dynamic
  changes, …)
  XHTML 1.0
  CSS2 (@font-face)
  CSS2.1 (‘inline-block’, ‘pre-wrap’, parsing…)
  CSS3 Color (rgba(), hsla(), …)
  CSS3 UI (‘cursor’)
  data: URIs
  SVG (SVG Animation, SVG Fonts, …)

(Taken from: http://www.webstandards.org/action/acid3/)

Looks important to me.

CSS3 isn't a finalised standard, and the test only tests a few things. You do realise there is a lot which this doesn't test too right? That's why some browsers took so quick passing ACID3 (because they focused on fixing the cases tested by ACID3 only)

And as mentioned, one of them is a speed test which is affected by external circumstances.

I'm not saying its irrelevent, but I am saying that a 100% pass rate doesn't mean its 100% compliant with the things being tested. Furthermore, we shouldn't be punishing companies which aren't willing to implement support for standards which are drafts.

ACID2 was genuinely useful because it was all finalised standards. ACID3 though I think could have done better.. So its nice to pass, but a 100% pass just says you are compliant with 100 tests. I'm betting the compliance testing suites used by browsers are a LOT more comprehensive..

  CSS3 isn't a finalised standard, and the test only tests a
  few things. You do realise there is a lot which this
  doesn't test too right?

That's exactly my point: Acid3 test the important bits, the parts of the spec which are useful. This test is web apps oriented (hence a strong focus on DOM manipulation and dynamic behavior) and thats where web is heading now.

        I couldn't give a damn about supporting unfinalised standards the ACID3 tests
    ...
    HTML4

Saying HTML4 is important in reply to a comment about unfinalized standards is a bit of a non sequitar, no?

Personally, no, I don't give a damn about SVG, but I'd be pretty upset if my browser didn't support iframes. Does the Acid3 test weight its tests appropriately?

From the FAQ:

Does Platform Preview replace my current Internet Explorer?

No, it does not, though it does share some settings with your existing Internet Explorer intsallation. One of the best features of Platform Preview is that it installs side-by-side with earlier versions of Internet Explorer and any other browser(s) on your computer.

While I'm glad to see this, it flies in the face of their previous position that side-by-side IE was impossible.

Maybe their previous position was entirely accurate for older versions of IE.

http://tredosoft.com/Multiple_IE

It is possible, but a bit of a hack.

better version

IE 1,2,3,4,5,5.5,6,7,8+Dev Toolbar

http://finalbuilds.edskes.net/iecollection.htm

this is slick. thanks for the link; i have never seen this before!

Finally, somebody implemented border-radius without throwing a prefix in front of it. I get tired of all the -moz and -wekit prefixes.

I suppose you hate namespaces in your code too?

The vendor specific prefixes allow a playground for vendors to define and test things their own way without setting expectations for things that are not standardized, are being standardized, or have not even been conceived of yet.

It's like the X- prefix on mail and HTTP headers.

I think Opera beat them to it.

They did. Opera has this since Presto 2.3 (Opera 10.50 comes with 2.5)

Does this mean that webkit and mozilla are going to drop the prefix? What's the point of the prefixes to begin with?

The point is that those things are/were still in "draft" specification and so it's asking for trouble to implement them in a way that will prevent changes to the final spec. I'm sure once the standards are finalized the prefixes will be dropped (or really, deprecated).

Because you get yelled at for doing a blink tag, but it's fine to do a moz-blink tag.

The problem is when browser #2 wants to implement blink. They shouldn't use your prefix, so they now have webkit-blink. And the HTML writer now needs to sniff the browser.

Progress?

These are CSS properties, not HTML elements. There is no need to sniff; you just provide multiple properties.

p.wibble { -moz-blink: slowly; -webkit-blink: 10%; -ie9-blink: "Microsoft.XML.Autocompositor(alpha=99,blinkrate=14692,emulate-old-ie-versions=false,true,true,false)#ie9" }

or whatever. (In practice everyone who provides these properties does so in a very similar manner, so it's easier than the above pseudosnippet suggests.)

ms demo of border radius works in Chrome as well

Aside from all the meta-discussion about MS and IE in the abstract, my opinion on this platform preview:

I'm extremely unimpressed. Judging by the absence of a location bar and settings menu this appears to be more of a demo than a pre-release application preview. This is very disappointing when measured against MS's other previews, betas, etc. (Windows betas and RCs, Visual Studio CTPs, even previous IE betas) and outright pathetic compared to, say, Chrome's dev-channel releases (which are release quality by the standards of this industry).

Speed was unimpressive on my system compared to Chrome, Safari, and Firefox. Acid3 performance was also unimpressive as I had to force IE9 document mode via the debug menu in order to achieve even a score of 39. One of the important aspects of the acid tests is that the browser be in the default configuration when visiting the test page. Just because there's some magical "improve standards compliance" button on IE doesn't mean that the browsing experience for everyday users will increase, somehow every other browser manages this problem just fine, yet IE always struggles with it. Perhaps they'll fix it by release, if we're lucky.

The supported features are refreshing but nothing that multiple other browsers haven't supported for a while.

Once again IE is trying to play catch-up and once again they are doing so slower than the other browsers are pulling ahead. It's good to see IE improving its feature-set, standards compliance, and performance, but at this point IE9 is racing the clock against irrelevancy. By the time IE9 is released its implementation of several years old technology may seem quaint.

Anyone know what the word is on WebSockets?

As far as I know Chrome is still the only browser that supports WebSockets. Can anyone refute this? FF 3.7 maybe?

FF 3.7 is the targeted release for WebSockets.

Isn't there a release candidate out for FF 3.7? If so can you confirm that WebSockets support is in there?

There's an Alpha3 'technology preview' of Fx3.7: ftp://ftp.mozilla.org/pub/mozilla.org/firefox/nightly/3.7a3-candidates/

I don't know if WebSockets is included or not.

Websockets are something I long dreamed about : bidirectional HTTP, no more stupid AJAX polling tricks.

Yeah now you just have to deal with this stupidity:

Send the following bytes:

           0D 0A

did that work? ok send the following bytes:

           57 65 62 53 6F 63 6B 65  74 2D 50 72 6F 74 6F 63
           6F 6C 3A 20

etc: https://tools.ietf.org/html/draft-hixie-thewebsocketprotocol...

No, you don't need to. Try putting this in a hex editor. You'll see a http handshake. The first part you need to send is "GET " which is the same as sending the following bytes: "47 45 54 20". It's just made a bit "simpler" so people don't get the encoding wrong (non-ascii characters, etc...).

long polling is nice because you don't have to implement a streaming parser on the client side.

I believe Microsoft has to ask msxml and WinINET team to handle this.

I hate to be conspiracy-theory-guy, but deep down, I feel that the rate at which IE becomes compliant is tied to Silverlight's market penetration. The move to open standards on the web will only truly come when MS feels like it has sufficient control over a different proprietary technology.

Uhm, off-topic much? I mean, Silverlight isn't linked to IE9 in any way. Silverlight is a plugin, like flash. It will work in Firefox too, but isn't installed by default. I don't see the link between IE9 (which gives web developpers the possibility to make interactive websites WITHOUT plugins) and Silverlight (a plugin).

All this highlights is their complacency up until now. Competition should never have been the thing that stirred them into making a browser that wasn't dire, it should have been a matter of course.

Imagine if they had to fight for operating system dominance.

IE 9 will not support XP. Now take that to the IE 6 user-base. Ref: http://blogs.zdnet.com/microsoft/?p=5572&tag=col1;post-5...

No mention of WOFF? So we're stuck with EOT?

Here's the developer guide: http://msdn.microsoft.com/en-us/ie/ff468705.aspx

Here's what would make my day: an announcement by MS that they will force-upgrade all users to IE9 from IE6,7 and 8. Goodbye clunky images + css + javascript hacks, hello lovely design.

It'll be nice to see independent verification of their SunSpider benchmarks but if they hold up that's very good news.

hmm, in viewing HN from the IE9 platform preview it centers all the text which is promising ;)

This is a step in the right direction but it obsoletes any windows/IE computer earlier than Vista SP2.

Considering the amount of resources used by Vista, I think all speed improvements of this browser would only be realized if you're already dedicating lots of processing power to OS overhead.

While the hardware acceleration will only be available on modern Windows, there's a good chance IE9 will come out for Windows XP too. (The lack of hardware acceleration on non-modern Windows is something that Firefox will face too once it flips on the D2D switch by default.)

> I think all speed improvements of this browser would only be realized if you're already dedicating lots of processing power to OS overhead.

That's really not how operating systems work. CPU and RAM are rather orthogonal.

That's really not how operating systems work. CPU and RAM are rather orthogonal.

1) click on start menu 2) wait 3 seconds for it to respond and open 3) click on icon for browser 4) wait 15 seconds for it to respond 5) wait 5 seconds for it to render the default home page 6) type in desired url and wait 3 seconds for it to download and render.

It doesn't matter if the new IE browser is faster than firefox, since any computer that is powerful enough to run Vista in the first place will probably not take a noticeable amount of time to render a page in Firefox.

The above is an example of the vista user experience on a slower computer. That same computer could (running linux) load google chrome and open the desired page in 1 second each.

> It doesn't matter if the new IE browser is faster than firefox, since any computer that is powerful enough to run Vista in the first place will probably not take a noticeable amount of time to render a page in Firefox.

Probably not, but hardware acceleration makes a ton of difference in responsiveness (it pushes a lot of things below that 100ms barrier where you feel that things happen instantaneously). Try out a Firefox 3.7a3pre nightly on a Windows Vista SP2 or 7 computer, enable Direct2D, and see for yourself.

How about making a Mac version?

I probably wouldn't use it, but the more browser choice, the better (healthy competition pushing everybody to do better).

Apple has Safari for windows, it would make sense for Microsoft to do something similar...

Why would it make sense? Would any Mac user use IE? How many Windows users use Safari? Not many I'd wager.

The whole point is that browsers should stand on their own and not be part of the OS, no?

If IE becomes really good, it should be judged on its own merit and maybe some people on the Mac platform would want to use it.

I dislike Microsoft as much as the next guy, but I know that when there's healthy competition on a platform, everybody wins. I'd love to see the IE guys try to make a better browser than others, and not just on windows.

Nice work zaatar and team!

It's certainly fast, but the lack of <canvas> is disappointing.

But it should support SVG. I'm not convinced canvas is the greatest in-browser rendering option for many of the things web devs want to do anyhow like click handlers and simple animations. There is a reason why many browser-based vis libraries are switching over to SVG (and VML in older IEs).

There's always been separate use-cases for pixel and vector-based graphics.

Even so, canvas is useful and I'd really like to have it available.

lol Microsoft.

Interested in checking out the new Web platform capabilities of Internet Explorer 9?

Honest question?

No. Not at all. All I am interested in is: Can you please ship it with chrome frame pre-installed.

Browser competition is a good thing. If Microsoft released the fastest standards compliant browser and became the industry leader (from a standars perspective) tomorrow I would be happy. I don't care what OS you run on your computer or which browser you choose. All I care about is my bits are displayed as intended and I find that IE makes that task significantly harder to accomplish. If my life could be made easier by a new MS browser, and other's browsing experiences improved, I'm all for it.

Not sure why I got bumped to -4, did we suddenly fall in love with IE?

I'm also not sure how you fit "browser competition is a good thing" and microsoft in the same sentence. I still have scars from IE5-7, and currently collecting new ones from IE8, which is less prone to screwing up, but also more subtile when it does (intermittent white page due to js bugs? lovely..).

As far as I am concerned I stand by what I wrote: I wish it would just go away (pipe dream) and I'm not interested in the slightest in what new bugs and misfeatures they introduce this time. I'll have to deal with them soon enough anyways, unless someone finds a way to trigger a force-install of chrome frame without user interaction...

(I'll take a gecko-frame or even opera-frame as well, if those exist - none of them have ever given me nearly the misery of trident)

Sure. But there's a much simpler and more foolproof way to get your bits displayed properly to everyone who buys a new windows machine (which of course are just about the only non-developers who will get ie9):

Scrap IE, start shipping windows with Opera, Chrome, Safari, or Firefox. Let the crappy existing versions die off slowly and focus on improving the mediocre operating system rather than reinventing the wheel with a consistently inferior and uncooperative browser.

I think Windows 7 is a pretty solid offering.

> (which of course are just about the only non-developers who will get ie9):

Why so?

Historical evidence. Firstly, IE users are less likely to care enough to upgrade a browser, for the same reason they didn't care enough to look for a better one. Namely, it doesn't really matter to them and they're unaware of why it should. Secondly, Microsoft is a lot less pushy about software updates than other browser vendors.

IE7 came out 4 years ago (and then IE8 two years later), but despite what an enormous improvement it was to IE6, we're only now seeing IE6 finally start to fade away. IE6 usage didn't drop below half of all IE usage until around June 2009. By comparison, Firefox 3 had more than double the usage of FF2 within 3 months of its debut. Chrome shows even better (in fact, absurdly fast) adoption rates.

All this source StatCounter: http://gs.statcounter.com/#browser_version-ww-monthly-200807...

Add in corporations hesitant to update IE in the workplace, and you've got the IE stick-around-forever problem.

Since those not using Chrome probably don't know about it, it's worth noting that Chrome updates itself whenever you start it. That explains the absurdly fast adoption rate of new versions.

But that still doesn't provide evidence that people buying a new computer are the only non-developers who will get IE9. You need to demonstrate that upgrades from IE6 to 7/8 are non-existent.

In fact, I'd say the fact that IE7 has dropped below IE8 is evidence that people on IE do upgrade, and that the people on IE6 are only on it because they need it.

I'm counting on common sense interpretations here. Obviously it's not the case that there are NO EXAMPLES of 7->8 upgrades. I'm suggesting that it's very much an exception rather than a rule. I don't have the energy to dig up the stats demonstrating a correlation between purchases of new computers and browser version market share. It's probably there, maybe it isn't. Whether that's the cause is immaterial though because the point remains: new IE versions tend to be adopted very slowly.

And I would point out the enthusiastic reception of Win7 (comes with IE8) compared to the chilly reception of Vista (came with IE7) as evidence that browser adoption is correlated with sales of new computers (or more specifically, new OS's. Keep in mind a lot of Vista machines were sold with downgrade packs).

And yes, Chrome auto-updates. That's part of my point about Microsoft being less pushy with their updates than other vendors.

> I'm suggesting that it's very much an exception rather than a rule.

I think the opposite is true. I don't have hard data to back up my thoughts, but the fact that IE8 is above IE7 seems suggestive.

> I don't have the energy to dig up the stats demonstrating a correlation between purchases of new computers and browser version market share. It's probably there, maybe it isn't.

So you're retracting your statement, then? I'm asking because you seemed to lack any sort of doubt when you made it, going to the length of using of course.

> Whether that's the cause is immaterial though because the point remains: new IE versions tend to be adopted very slowly.

I agree (relative to other browsers at least), but this is certainly not what you stated in your original post.

Reply on news.ycombinator.com