Announcing Internet Explorer Developer Channel | Hacker News Reader

Announcing Internet Explorer Developer Channel

blogs.msdn.com

159 pointsrobin_reala92 comments

The really amazing thing here is that they've disentangled IE from the base Windows. The key comment:

  This release of the IE Developer Channel uses a combination of code changes
  and App-V Client to virtualize and run alongside IE11.

This is pretty huge, because it means that the support requirements for IE can be drastically reduced. That means faster releases without harming backwards compatibility. One IE for the server and OS platform, one IE the application platform (modern apps), and another for the user. Or even instanced per-user IE, or who knows what they could do. It's a big deal that they've gotten IE to the point that it can run in an App-V container, and the result will be a faster development cycle and earlier access to new features in the client, without imposing a huge support matrix on application developers.

This could pan out into versioned webview controls and all sorts of things. I hope Microsoft is willing to take this as far as it can go.

Hopefully it won't end up in a situation where a lot of apps bundle seriously outdated versions and any given system has a dozen old dlls strewn all over. Didn't that kinda happen with bundled flash players - even Adobe shipped exploitable players for years in other products?

We're already there on Android. The only reason we aren't there with Apple is because they release so infrequently.

The new WKWebView seems to solve this problem. It's bundled with the WebKit project, not iOS or OS X.

Hopefully they do what Firefox tried (and failed) to do: ship a stable, enterprise version and an advanced version. No locking to arbitrary versions, but able to say, support "IE11 and IE Latest" in your product. Hopefully they allow developers to opt into a 30-60 day preview window such that they can remain a bit ahead of their customers.

and failed? Mozilla’s ESR releases are definitely a thing.

I'm using one right now.

Pretty sure they back port security/bug fixes as I have to keep updating it.

The Firefox ESR branch is on the same release schedule as the Firefox 'Stable' branch. So, you get a scheduled release every 6 weeks. But ESR just gets the security fixes whereas Stable also gets the new features. When there's an out-of-band security update on Stable (a .1 release), ESR gets that, too. ESR is pegged to specific releases of Stable and operates in parallel for a few releases so there are two different ESRs... the older one and the newer one... so organizations can transition from one browser engine to the next over a couple month timespan and ensure corporate apps work on both.

Oddly, some non-organization people want ESR because they think it's updated less often. It's not.

They do. That’s the support in ESR.

I would not suggest MS use App-V to do that though.

It still required a system reboot after installation (Win7). I'd assume it's not so disentangled yet.

That is one of the things that still make me wonder how an installer (or installer maker) decides retart is needed or not. I mean, take Visual Studio for example: it seems an order of magnitude larger and more complicated than a web browser, comes with prerequisites and whatnot and the last 2 or 3 release I installed didn't need a restart.

http://msdn.microsoft.com/en-us/library/aa372059(v=vs.85).as...

I didn't need a reboot after the install, but I'm on Win 8.1.

> The really amazing thing here is that they've disentangled IE from the base Windows

Calling bull. MSIE has never been "entangled" with the OS and this has in fact been proven in court ( http://www.justice.gov/atr/cases/f2600/2613g.pdf pages 283 through 288 are especially noteworthy). The integration was merely a scam to defeat Netscape.

Here's a bunch of standalone versions from 1996 to 2002 (made by MSFT themselves) which can all run side by side: http://browsers.evolt.org/browsers/archive/ie/win32/

On the same site, you can find MSIE for HP-UX and Solaris.

IE has been a system-level service in Windows for a long time. Naturally, you could boot the kernel without it, but a bunch of included Windows software (by Microsoft, as part of the distribution) and even system libraries rely on parts of IE.

Even now, there are weird dependencies on IE components that you wouldn't expect: http://randomascii.wordpress.com/2014/03/31/you-got-your-web...

It's quite possible it was all bullshit at the time of the antitrust case, but it's been true for quite a while now. In practice, 'removing' IE from windows at this point would just mean removing iexplore.exe, because literally every native .dll and COM .dll bundled with IE is probably leveraged by some application or component somewhere in the OS. The whole IE object model and programming interface are documented on the web and exposed, so there are programs using them - Valve's Steam game management/storefront app used to use it before they moved to their own embedded version of WebKit.

> In practice, 'removing' IE from windows at this point would just mean removing iexplore.exe

And this is exactly what happens if you don't install IE with Browser Choice in the EU. The rendering engine and all the DLLs are still installed, It's just iexplore.exe which isn't.

I has been, what, 15 years from those court proceedings. Don't you think it is possible the MSIE has been tied to the operating system in subsequent releases: Windows XP, Windows Vista, Windows 7 and Windows 8?

Sure it's possible, but then it wouldn't be so amazing that they had un-done it. Microsoft was, in those court proceedings, claiming that being integrated with the OS was a very fundamental part of IE that was not trivially undone.

AFAIK nothing much has changed since these days, but MS did add a option to remove the Internet Explorer directory in Program Files in Vista and note that HTML Help for example already depended on IE components even back in 1997.

What is funny is while the FAQ says a lot about past release versions of IE not supporting this, it does not say anything about whether future release version of IE (like IE12) will support this.

IE also shipped for Unix and Mac OS. So I doubt the entanglement claim.

The fact that it's an installable application for other platforms doesn't mean that IE on Windows isn't used by system services... Quicktime is installable on Windows, so would you say Mac OS doesn't use it for anything? Because that's not true either.

IE has been used extensively by a bunch of Windows components and software for a really long time. Sometimes in unexpected places: http://randomascii.wordpress.com/2014/03/31/you-got-your-web...

> The really amazing thing here is that they've disentangled IE from the base Windows

You realize that this is something other browser makers have done since, oh, the invention of the browser? Perhaps it's amazing that Microsoft, of all companies, finally did it, but let's not get carried away declaring a historic moment.

Everything in context. It is huge for Microsoft specifically because it's previously been so entangled. Progress is good on all fronts. For the very large portion of the world that uses IE by default and will in the future get faster/better upgrades to be more secure and more full-featured, this is great.

Entangled is a very kind word, almost sounds accidental, when the anti-trust stuff found it was all very much on purpose.

Exactly, IE has come a long ways in the past few versions, but it's still behind other modern browsers. It makes me wonder why MS is still pushing it. Why not just bundle another browser with Windows and save themselves some effort?

Probably for the same reason Apple and Google both produce their own browser - it's too important a part of the user experience to delegate responsibility for it to a third party.

It's unlikely Google would suddenly turn around and say they're no longer producing Chrome for Windows, for example, but it's not impossible.

Also, IE, or rather Trident, is a core OS component. Applications rely on it for displaying web pages and HTML. Applications can also be written in HTML and JS, both HTML Applications (since... Windows 98, I think?) and WinRT applications (since Windows 8).

I never understood why they simply could not use Trident for OS calls and let IE have its own engine. It seems like such a straightforward solution.

It is like building a computer and soldering in the GPU, just because you want to reference the PCI port statically.

The main IE components has been in System32 since IE3 I think.

Yeah you're probably right. They want the experience to be cohesive with the rest of Windows. Last time I looked Chrome still doesn't have a metro mode besides just opening up in full screen mode, so there's one reason to keep IE around

This is exciting news. I switched from Firefox to IE11 at home, and find that IE is as good as any other browser for regular web users today - this after using Firefox for almost 10 years.

I am also forced to use IE8 at work, as we are still using Windows XP - a non-software company is really really slow to upgrade their software - we were on IE6 till 2012.

As an aside, can someone let me know why IE11 still scores so low in http://html5test.com . Does this website test features which IE11 and future versions have refused to implement at all?

It varies by feature, but due to IE's long release cycle, it's generally about a year behind Firefox and Chrome on most things.

While basic sites work, some things just aren't possible to do in IE yet. For example, http://www.wavesurfer.fm/ works in every recent browser other than IE 11.

It varies by feature, but due to IE's long release cycle, it's generally about a year behind Firefox and Chrome on most things.

That's not entirely fair.

In terms of ticking boxes for new features, yes, IE certainly trails Firefox and Chrome by a few months, sometimes even a year or two.

On the other hand, when IE does claim to implement something, generally that implementation is fast, robust and stable. That certainly can't be said for Chrome or Firefox. Some obvious examples are the awful font rendering in Chrome (only just being fixed at the moment), lack of H.264 support for <video> elements in Firefox (still not fixed, just avoided on Windows with a workaround), and performance problems rendering SVGs that really undermine things like animations and using lots of SVG icons on a page in both browsers, just to pick three widely used recent features that all major browsers now claim to support.

While basic sites work, some things just aren't possible to do in IE yet.

That's very loaded terminology. Almost all real sites work just fine in any recent version of IE. The few that don't are mostly things like web design blogs and demos that push the bleeding edge because they want to.

Often these sites use features that aren't standardised yet and maybe still need browser prefixes to access the feature at all. These are exactly the kind of features that Chrome and Firefox tend to half-implement and then regress or change the spec several times over the next year or two before finally dropping the prefix. I doubt any professional web developer would actually use that kind of feature on a production site today unless they had very unusual requirements to meet. The example site you gave is another example of this.

You can compare the items tested against what is being developed here: http://status.modern.ie/

Have just installed the Dev Channel version of IE11 and noted that the dev channel version contains additional ticks for:

Input (was 5/20, now 7/20) * Gamepad control

Storage (was 26/30, now 30/30) * Database storage - Objectstore Blob support - Objectstore ArrayBuffer support

So, previous IE 11 score: 372/555, IE 11 Dev Channel score: 378/555.

I wonder if they purchased Custom Support for WinXP.

In today world using any version of IE as a main browser is a bad idea. Why? Because browsers evolve very quickly and IE doesn't. It might look at first that IE is running quite nicely and even fast. But soon you start to notice all of the glitches. Especially if you're web developer. As long as there is no auto update, IE is always a threat for web developers as IE6 can happen again..

IE10 introduced auto-update. Once 11 came out, computers silently updated to it. No big windows update prompt asking them if they're sure they want it.

IE9 to IE10 was aggressively launched as well. I believe users got a basic prompt and IE turned on auto-updating by default. According to some of my users, IE10 just appeared, so my guess is that MS may have the prompt time-out to 'yes' instead of infinitely holding the computer.

Here in corporitstan, I had to put in extra blocks to keep 10 and 11 away, due to our CRM being suck on IE9 for the time being. Trust me, MS is trying real hard to get everyone to the newest version of IE.

You see I do understand that they try and I do applaud that. yet on my windows 8 machine I still haven't received IE11, even though I just recently did got a notice that IE has been updated, it is still IE10. To make matters worse even if I visit their website to download IE11 manually [1] they're telling me that I already have IE11.

So that experience alone tells me that not everything is done right yet. And even if everything would work, they have to keep doing things right for some time to regain trust.

[1] http://windows.microsoft.com/en-gb/internet-explorer/ie-11-w...

IE11 didn't ship for Windows 8.0 it only shipped for Windows 7 and Windows 8.1.

Windows 8.1 is a free upgrade on Windows 8.0, more like a service pack release (when you look at the changes). Why not port it to Windows 8.0, it didn't make sense, 8.0 should fade out as people install the free upgrade that comes via Windows Updates (or the store, I forget which) and is a mandatory update. Opting out of mandatory updates means you opt out of associated upgrades.

This isn't a new or unique policy in Windows, look at iOS, you won't get WebGL on Mobile Safari without iOS 8, OSX doesn't back-port Safari and Andriod has a lot of versions with terrible browsers.

>even though I just recently did got a notice that IE has been updated, it is still IE10.

Win8 ships with IE10. If it did an upgrade, then you should be on 11 now.

IE11 is supported on Windows 7 and 8.1, but not Windows 8.0. So on 8.0 the auto-update feature is ineffective. According to StatCounter only 5% of 8.0 users have updated to 8.1. There must be something blocking people there, not sure what it is, but end result is the auto-updater doesn't actually auto-update everyone and fragmentation gets worse.

Why blocking IE10/11 in a corporate environment instead of simply updating it and pushing a GPO that forces IE9 mode for your CRM?

Those modes fail testing. They're not replacements for the real browser. Helps with trivial things like formatting issues, but for the handful of things I've tested that work in 9 and break in 10, the legacy modes have largely the same issues.

Because compatibility modes are imperfect.

Well, I am not a web developer - which is why I mentioned "regular web users".

Also, most of the WebGL demos around the net do not work in Firefox for me - it is possible that it is something to do with the installed drivers. However, they work perfectly work for me in IE.

I understood that. Yet you're here and you're familiar with IE6 situation. You know that while it might have looked like developers only ones who got hurt by that situation, actually everyone, including regular web users were hurt. As many hours were spent just to provide support for IE rather than creating something more useful. In other cases support for IE was dropped, which hurt users IE right away. I provided developers view, so that you could see how regular web user choice makes an impact to the whole industry.

I highly recommend FireFox, Opera or Chrome instead of IE. IE is perennially around 2 years behind (or more) every other browser (overall - in places they are on a par - and very occasionally ahead).

IE 11 is evergreen, I believe. It's auto-update. So hopefully in the future, IE 6 won't happen again, except for those enterprise customers who turn off the auto-update.

IE6 has already happened. It is called Native Client in Google Chrome.

>As long as there is no auto update, IE is always a threat for web developers as IE6 can happen again..

They have auto update for more than a couple of years now.

http://www.geek.com/news/microsoft-decides-to-automatically-...

Yes, but it doesn't work. It might be because people opt-out, they disable updates, or maybe IE is mostly used in corporate envs where people don't autoupdate anyway. I don't know WHY, but numbers show that it doesn't work.

IE11 has been released on October 2013 (8 months ago), but IE10 market share is still 6.77% today (compare with IE11 at 17.03%). I'm quoting NetMarketShare (but it doesn't really matter, all browser stats agree on this topic, with slightly different numbers).

For comparison, Chrome and Firefox versions usually go under 1% after no more than 4 months after a newer version has been released.

People who use IE are part of the problem, not the solution.

You're now part of a statistic that will be (sometimes wilfully) misinterpreted to force devs to support IE6-10 and to not support all the exciting technologies that IE doesn't support.

Please reconsider what you're doing, it's bad for the web ecosystem and therefore society in general.

I don't get it. So by using IE11, you're part of the IE6-IE10 statistic? How so?

This is an interesting announcement. I wonder how far away the IE team is from open-sourcing the browser on top of all this.

There's been a sea change under at Microsoft over the last few years: support for web standards (WebGL!) and less of the "old Microsoft" anti-OSS messaging. I could easily see them taking this next plunge. There's still a lot of work they need to do to win me back to the platform, but I like what I've seen over the last few years.

Now if only we could get a proper unix subsystem on Windows....

Sad thing is they had a product like this, Services for Unix but it was a server product that never got much love or attention. IIRC, it extended the POSIX personality to give you a true Unix environment running on the NT kernel.

A lot of people don't know that Cutler's team designed NT to support different personalities/subsystems and it originally shipped with a POSIX one, an OS/2 (text mode) one, and the Win32 subsystem.

If you put in the effort, there's no reason you couldn't use the NT kernel as the basis for a Unix OS without any Windows API whatsoever. It's a single root filesystem/object system with devices mounted in directories, etc. It goes even further, things like synchronization primitives are also mounted there.

If there was a new unix-y personality on top of NT, I'd be so happy. Win32 is saddled with such crazy baggage like MAX_PATH being 260 characters, even though the underlying OS supports 64kB paths. Arrrrgh.

Correct me if I'm wrong, but isn't all of this openness a result of Microsoft realizing it's loosing, and trying to win back market share by niceing up to OSS developers?

Feels like they're flogging a dead horse to me a little. IE is constantly playing catchup to the other browsers, and MS gave us the web development darkages thanks to version 6/7/8. I kind of wish they would just kill it off and offer the completion up to the users.

I truly hope these are the first steps towards an evergreen Internet Explorer.

When that day comes, webdevelopers world wide will have one hell of a party.

It will still be unacceptable for development purposes because of the poor quality of their utilities. Does this affect non-dev users? I would argue yes, because web apps will have the fewest bugs and run the best in the browser they are developed in. All work done to make the app "cross-browser" will be blamed as compatibility issues with other browsers. This is why IE, no matter what version, will always be seen to have compatibility issues (unfairly in my opinion - IE is a fine browser and all browsers have some differences) until it has development tools on par with Firebug or Chrome.

Have you looked at IE's developer tools recently? They are way ahead of FF, and in some aspects even Chrome. I'd use IE11 for debugging if it would run on OSX :)

I like the memory profiling, way better then the others.

That dev tools screenshot looks pretty awesome. Seems to have even more detail than Chrome's timeline view. Gotta try it sometime.

Anyone else getting "application server errors"? No other details (for security reasons), but reloading keeps showing various error pages.

This is my first release in the JavaScript team. Here I optimized the JSON parser to work 30% faster for escaped strings. @cshung

Does this developer channel include WebRTC support?

It looks like the WebRTC Object API might happen for IE13, but no support for the full API is planned.

It looks like IE12 should have the Audio API and getUserMedia, though. That lets you do some interesting things with audio, at least.

I wish it did. I'd also like file API and directoryReader support, and the web audio API. The truth of it though is I don't care anymore if someone's browser doesn't have WebRTC. If you don't have it my site won't work for you, tough luck. WebRTC capable browsers are available on every platform. IE and Safari (on a mac) are like the shitty bundled browser you have to use to download a real browser at this point. Firefox, Chrome and Opera have all this stuff, except Firefox doesn't have directory uploads :(

if you take a look at http://status.modern.ie/ you can see which major standards / features are in flight. or not, in the case of WebRTC

In my design these days I specifically write down: test the HTML/JS/etc code in FF and Chrome only and forget about IE fully.

I mean really, why do I need care about IE at all these days?

Everybody that targets business users.

Funny thing: I have the opposite problem right now, I'm using a site that is targeted strictly towards business users and I don't have IE anywhere here so it's been nothing but trouble for me over the last couple of days.

You can get free VMs at https://modern.ie/en-gb/virtualization-tools . And also a free 3 month BrowserStack trail from the same site.

Thank you very much. Unfortunately the job is of such a nature that I won't be able to use this but the pointer is much appreciated.

Does this mean that Internet Explorer can, in fact, be run completely independently of the core operating system?

Doesn't seem that simple:

> This release of the IE Developer Channel uses a combination of code changes and App-V Client to virtualize and run alongside IE11. This virtualization creates a small performance hit, so we don’t recommend you use this version to measure your site’s performance.

Of course, they would highlight the complexities of the process wouldn't they? (cough - antitrust - cough)

It'll be using a veritable pile of standard Windows libraries no doubt, so you'll not be able to just run it on Linux or OSX if that is what you are asking.

By "able to run independently" I assume they are meaning that it won't have to become the one and only version of IE on your machine as other versions do meaning you can run it along side other IE versions (unlike with 11 and prior which essentially require you to have a separate Windows setup, usually in a VM, for each version you want to test against).

It was a tongue in cheek reference to the antitrust trial of the 1990s where Microsoft spent millions arguing that IE was inseparable from the core OS.

On a sidenote, why use JPEGs for screenshots? It makes text look awful, use PNG already.

http://blogs.msdn.com/cfs-filesystemfile.ashx/__key/communit...

I see this problem in almost all MSDN blogs, must be part of a convoluted CMS workflow or something.

I assume that their blog app just compresses everything as a JPEG and nobody ever bothered to change it.

Maybe because it's resized down. In contrast to what others have replied, I personally find PNG file sizes to be horrible for thumbnails - even after running through optipng.

File size I imagine. Typically, a PNG will be 4 or 5x the size of a low/mid quality jpg. 31kb vs 150kb when you're delivering a bazillion pages makes a difference if you don't have the bandwidth.

Granted, that seems quaint nowadays, but maybe MS doesn't give two shits about making sure the blogs have the bandwidth and resources they need for nicer looking graphics.

Actually, for pictures of text, a PNG is usually 1/2 or 1/3 the size of a jpeg. Try taking a screenshot of a hacker new's comment section and saving it as PNG or jpeg and check the results.

Actually, for images with very few colours (such as screenshots), PNG usually has significantly smaller file sizes than JPEG...

Or they could use WebP and get the best of both worlds.

O wait...http://status.modern.ie/?term=webp

or an ISO/ITU-T standard: JPEG-XR, which they do support already?

http://ie.microsoft.com/TEStdrive/Graphics/ImageSupport/Defa...

Too little too late.

Finally, the entire source code!

Sorry MS astroturfers, I thought this was a thing for developers!

Reply on news.ycombinator.com