Linked by Thom Holwerda on Wed 11th Mar 2009 09:11 UTC, submitted by poundsmack
Internet & Networking Apparently, Internet Explorer is on its way out. JCXP.net is saying that Internet Explorer 8 will be the last traditional version of Microsoft's web browser, and that Microsoft's next web browser will be based on a promising Microsoft Research project dubbed "Gazelle".
Order by: Score:
Very vauge, don't you think?
by TLZ_ on Wed 11th Mar 2009 09:43 UTC
TLZ_
Member since:
2007-02-05

The mentioned linked seems to follow this pattern:
"This is something amazing/special - very uncertain, don't blame us if we are wrong".

I'm sceptical.

Reply Score: 3

latex
by evert on Wed 11th Mar 2009 10:10 UTC
evert
Member since:
2005-07-06

Exciting. It just looks like they are still using Latex for their papers :-)

I had a look to the PDF properties - indeed, TEX!

Gazelle looks OK to me.

Reply Score: 3

Comment by flanque
by flanque on Wed 11th Mar 2009 10:28 UTC
flanque
Member since:
2005-12-15

I know a lot of people will just say it's gonna be crap because it's from Microsoft, but if these things hold true then this is very good news from a competition and security point of view.

It seems to look like a stand alone product as well, which is excellent news too.

Reply Score: 5

RE: Comment by flanque
by Clinton on Wed 11th Mar 2009 10:32 UTC in reply to "Comment by flanque"
Clinton Member since:
2005-07-05

People wouldn't say things like that if it hadn't been true so many times over the years.

Reply Score: 9

RE[2]: Comment by flanque
by flanque on Wed 11th Mar 2009 10:34 UTC in reply to "RE: Comment by flanque"
flanque Member since:
2005-12-15

No it's not true. Yes some aspects of Microsoft's software has been crap, but overall they offer excellent software.

It's a stupid generalisation.

Reply Score: 2

RE[3]: Comment by flanque
by darknexus on Wed 11th Mar 2009 10:42 UTC in reply to "RE[2]: Comment by flanque"
darknexus Member since:
2008-07-15

It's not true now. Trouble is, a bad reputation tends to stick around longer than the actual problems, and there's no denying that back in the days of Windows 9x a good deal of Microsoft's consumer software was a bloated, bug-ridden mass of bits. 9x itself, Office 97, etc.
Personally, I don't use Microsoft's software because I'm sick of being treated like a thief, it's the innocent until proven guilty attitude (activation), and the fact that even if I need to reinstall on the same computer I must call them to get an activation code. It's similar to WMA authorization schemes, no control granted to the user at all. Honestly, people complain about Apple's controlling methods, but even they don't put people through that crap. I also have come to dispise a lot of their business practices and artificial software limitations--Vista starter, anyone? So, I choose not to use their products.

Reply Score: 9

RE[4]: Comment by flanque - "back then"?
by jabbotts on Wed 11th Mar 2009 12:44 UTC in reply to "RE[3]: Comment by flanque"
jabbotts Member since:
2007-09-06

They skipped over Vista pretty darn quick to start pimping (er.. prepping) Win7 for release. I'll have more faith in the company's products when "back then" stops happening so frequently. Until then, I remain respectfully skeptical.

Reply Score: 4

redviper Member since:
2006-11-28

Don't be ridiculous. Vista came out end of 06, Win7 will be out end of 09.

In contrast, Win2k came out in 2k and WinXp in 2001.

Reply Score: 1

jabbotts Member since:
2007-09-06

The marketing hype is all about Windows7 and it's impending release. After an average of two to five years between major releases, that seems pretty darn short. The fact that they pretty much all but dropped premoting Vista and even the new MS figure head said "uh.. you can skip Vista, we're going to have Windows7 out for you really soon"...?

I will admit that the windows7 beta is far ahead of where Vista betas and pre sp production versions where. Hopefully the don't bloat up the userland around the kernel and it retains it's current snappy responsiveness. That plus IE9 being a truly third party application to the OS could mean some very nice benefits to the end user.

Reply Score: 2

RE[3]: Comment by flanque
by Kroc on Wed 11th Mar 2009 11:00 UTC in reply to "RE[2]: Comment by flanque"
Kroc Member since:
2005-11-10

Yes, overall, those 5 years of total neglect were alright. It only cemented the spyware and malware industry we have today.

Reputation is one thing, the damage created is another.

Reply Score: 3

RE[4]: Comment by flanque
by iptables on Wed 11th Mar 2009 17:40 UTC in reply to "RE[3]: Comment by flanque"
iptables Member since:
2009-02-26

Yes, I think they have cornered the market in this area.

:)

Reply Score: 1

jabbotts Member since:
2007-09-06

.. I'll be right back after I fix another of the office workstations..

Reply Score: 3

RE: Comment by flanque
by Laurence on Wed 11th Mar 2009 16:08 UTC in reply to "Comment by flanque"
Laurence Member since:
2007-03-26

I know a lot of people will just say it's gonna be crap because it's from Microsoft, but if these things hold true then this is very good news from a competition and security point of view. It seems to look like a stand alone product as well, which is excellent news too.


Not just their webbrowser either.
Windows is (finally) going modular:
http://news.bbc.co.uk/1/hi/technology/7932149.stm

Reply Score: 2

RE[2]: Comment by flanque
by n4cer on Thu 12th Mar 2009 02:30 UTC in reply to "RE: Comment by flanque"
n4cer Member since:
2005-07-06

The capability discussed in that article already exists in Vista (and Server 2003 +). The primary difference with 7 (vs Vista as it's been available in Server) is that MS exposed the option to remove IE and media-related features within the GUI rather than you having to perform the removal via Package Manager (pkgmgr.exe) or the Optional Component Setup (ocsetup.exe) commandline tools (or via unattended install).

This is a screenshot of the UI in Vista for which in 7 they've now added entries for IE, media, and new 7 features.

http://www.activewin.com/winvista/images/Windows%20Vista%20...

Reply Score: 3

webstandarts
by Arno on Wed 11th Mar 2009 10:34 UTC
Arno
Member since:
2006-01-10

If they would use webkit it would give web designers finally the possibility to use css2 to its full potential. It will still be a long time before the old IE-browsers will be gone tho.
The use of webkit is not certain according to the source but I could only hope!

Reply Score: 1

RE: webstandarts
by Kroc on Wed 11th Mar 2009 11:06 UTC in reply to "webstandarts"
Kroc Member since:
2005-11-10

Depends who you are targetting. In some countries FF usage is >50%. Home user usage is higher than corporate usage. Mobile devices like the G1, iPhone and Nokia N-Series all have good CSS support (webkit).

IE6 usage is 20% and dropping rapidly. FF3 usage alone is 20% worldwide. Can you see where this is going?

The point is that you should start developing for standards, HTML5 features and more *now* so that you can hit the ground running. It is not unreasonable to drop support for IE6 and ask home users to upgrade to Firefox if you're targetting home users.

IE should not be the millstone around your neck when it comes to personal development work.

Reply Score: 4

RE[2]: webstandarts
by werpu on Wed 11th Mar 2009 11:26 UTC in reply to "RE: webstandarts"
werpu Member since:
2006-01-18

Problem is, ie6 still is a huge burden for everyone not doing personal development. If I would make a private site I would probably even not support IE7 so much distain I have for this browser. The problem stems from the fact that in public sites even with a market share of 10% you still have to support that utter garbage IE 6 and 7 in reality are once it comes to standard support!

Not sure if ie8 is that much of an improvement, it will help that they seem to support CSS 2.1 properly now, but things like SVG and javascript to the latest level are still a huge issue!
But there wont be any wide adaption of that browser before 2010 or 2011 anyway given the slowness of typical Microsoft herds!

Problem is, that many corporations and banks curently are slowly migrating towards ie7 and some still will be stuck on ie6 for the years to come (seriously, some even run on ie5.5). So if you run a site or page targetting exactly those people you cannot ignore it no matter how hard it is to still support all this. If I had a choice probably Firefox3 and Safari4 as well as Chrome would be the lower limit (sorry opera but the 9er release is still way too buggy for many things dhtml wise) and I would not support anything below!
And nothing from Microsoft as long as they want to push their proprietary SVG copy instead of the real thing!

But reality is different to what is possible, unfortunately. I am just glad that the Firefox users are saner and migrate normally within 6 months to the next release so dragging around stone old firefox versions never has been an issue, same goes for Safari.
(I really love the 4 beta for all its capabilities)

Reply Score: 3

RE[2]: webstandarts
by google_ninja on Wed 11th Mar 2009 12:41 UTC in reply to "RE: webstandarts"
google_ninja Member since:
2006-02-05

Personally, I can't wait for min-width/min-height to finally be available to us.

I work for a company that has 35 moderate to very high traffic public facing properties, we just did check of our stats recently when the topic of dropping IE6 came up, and it turns out that it accounts for about 45% of our traffic.

Our typical strategy has been to support latest browsers -1 major revision, but we can't really do that if half our visitors are on an older browser.

Makes it a huge pain for us and our designers, because we have to fire up virtual machines to test for it.

Reply Score: 4

RE[3]: webstandarts
by Kroc on Wed 11th Mar 2009 13:10 UTC in reply to "RE[2]: webstandarts"
Kroc Member since:
2005-11-10

So have you found out how many of those IE6 users *can't* upgrade?

Don't just sit there and assume the marketshare problem is going to sort itself out, as a web developer we all have the responsibility to educate users.

Reply Score: 1

RE[4]: webstandarts
by google_ninja on Wed 11th Mar 2009 14:10 UTC in reply to "RE[3]: webstandarts"
google_ninja Member since:
2006-02-05

...and as an employee in a company providing a service to people and ultimately looking to make money from them, we have a responsibility to provide the best experience we can to them, not become an IT department they never asked for.

If a significant portion of them used lynx it would be silly in todays day and age, but we would work on providing a good text only experience. If there was a significant amount of mobile users, even though it is very painful to do we would work on good mobile interfaces. The same applies to horrible and obsolete browsers.

Reply Score: 5

RE[3]: webstandarts
by TBPrince on Wed 11th Mar 2009 14:30 UTC in reply to "RE[2]: webstandarts"
TBPrince Member since:
2005-07-06

Yours is the only sane approach for a business. When those IE6 users will stop existing, I'll be more than happy.

But many people thinks supporting standards will solve all of their problems, which won't be true. When IE8 will be here and FF3 and Safari4 will be here (not to mention Opera), designers will still be fighting to achieve a good effect on all of them. Unfortunately, IE6 is a relic of an acient (and fortunately died) Microsoft strategy.

But if people thinks software makers will stop fighting about introducing changes inside their Web platforms (of which, browsers is the key part), that's just ingenue.

People often forget when Netscape was the bad guy and it was breaking compatibility and IE was the god-sent software supporting all standards. Ironically, IE became so widespread because it was supporting standards better than Netscape...

Reply Score: 1

RE[4]: webstandarts
by google_ninja on Wed 11th Mar 2009 14:40 UTC in reply to "RE[3]: webstandarts"
google_ninja Member since:
2006-02-05

People often forget when Netscape was the bad guy and it was breaking compatibility and IE was the god-sent software supporting all standards. Ironically, IE became so widespread because it was supporting standards better than Netscape...


I don't know if I would go that far, but they were both non standard in different ways. In some ways IE was dumb (document.all), other ways ended up getting rolled into the specifications because they were either useful or made more sense (element.innerHTML, element.style.*, XmlHttpRequest), and some things they still do more intelligently then the standard way (the ie box model)

Reply Score: 2

RE[5]: webstandarts
by TBPrince on Wed 11th Mar 2009 14:56 UTC in reply to "RE[4]: webstandarts"
TBPrince Member since:
2005-07-06

I don't know if I would go that far, but they were both non standard in different ways. In some ways IE was dumb (document.all), other ways ended up getting rolled into the specifications because they were either useful or made more sense (element.innerHTML, element.style.*, XmlHttpRequest), and some things they still do more intelligently then the standard way (the ie box model)


But IE was supporting Netscape custom extensions while Netscape wasn't supporting IE custom stuff. If you think about that, IE deserved its widespread usage. Then Microsoft tried to be the Web itself and failed.

Anyway, what I was stating is the custom browser battle goes far beyond CSS standards support. The battle has moved from supporting your own extensions for HTML/CSS to being able to connect HTML and browser context to my specific platform.

This way I can inject my platform into HTML context, whatever it might be (.NET, Android, those FF extensions and so on...).

Reply Score: 2

RE[3]: webstandarts
by andrewg on Wed 11th Mar 2009 19:09 UTC in reply to "RE[2]: webstandarts"
andrewg Member since:
2005-07-06

min-height is not a problem since height in ie6 can't handle *> selectors and it treats height like min-height. So you would do something like:

#container {
height: 400px;
min-height: 400px;
}
*> #content {
height: auto;
}

I have never used min-width.

Reply Score: 3

RE[3]: webstandarts
by proforma on Thu 12th Mar 2009 08:39 UTC in reply to "RE[2]: webstandarts"
proforma Member since:
2005-08-27

You can have IE6, IE7, Firefox, Chrome, and Safari on the same machine without having a virtual machine. You can have all of them on Windows at the same time.

http://browsers.evolt.org/?ie/32bit/standalone

Download: ie6eolas_nt.zip

This gives you the real rendering engine of IE6 while having access to IE7 at the same time. I have used it so I know it works great.

Edited 2009-03-12 08:54 UTC

Reply Score: 3

gustl
Member since:
2006-01-19

The authors also state that they have made no concessions to backwards compatibility in favour of security.


I would assume from past Microsoft behavior, that the browser they will be putting out has to be backwards compatible no matter how much that may compromise security or speed.

Otherwise it looks like a promising idea.

Reply Score: 4

jabbotts Member since:
2007-09-06

They should stick to the HTML standard. Continuing to support broken IE only html makes any standardization compliance about as effective as the security-theater we all get to play a part in at the airports now. Why implement true travel safety when you can give the apearance of security; why implement good html when the broken website still displays fine in IE.

Baby shuffles are all well and good but eventually they'll have to take some baby steps.

Reply Score: 2

Not another rendering engine!
by vasko_dinkov on Wed 11th Mar 2009 11:44 UTC
vasko_dinkov
Member since:
2005-09-13

If this is true, I just hope it doesn't feature a new rendering engine but uses Webkit, Gecko or Presto as otherwise hell will break loose again for web developers.

Reply Score: 0

jabbotts Member since:
2007-09-06

Using an existing engine or writing something in house is just fine provided they actually stick to the HTML standards. I'm less concerned with how it's done just as long as the web content is independent of the browser used to view it; the same page should display the same way across browsers.

Reply Score: 3

Say.. there's a thought..
by jabbotts on Wed 11th Mar 2009 12:37 UTC
jabbotts
Member since:
2007-09-06

"Web content does not interact with the actual operating system at all"

That has to be an improvement over the current IE/OS intimacy. Here's hoping IE9 shows an honest go at unbundling the browser and protecting the end user.

Reply Score: 2

ClockEndGooner
Member since:
2009-03-11

The link to the paper from Microsoft Research below makes for interesting and thought provoking reading:

http://research.microsoft.com/pubs/79655/gazelle.pdf

I can definitely see the need to improve overall web browser and underlying platform security, and extending this security to the web browser, its plugins and the underlying native platform. From the paper's conclusion:

"We have presented Gazelle, the first web browser that qualifies as a multi-principal OS for web site principals. This is because Gazelle’s Browser Kernel exclusively manages resource protection, unlike all existing browsers which allow cross-principal protection logic to reside in the principal space. Gazelle enjoys the security and robustness benefit of a multi-principal OS: a compromise or failure of one principal leaves other principals and the Browser Kernel intact.

Our browser construction exposes challenging design issues that were not seen in previous work, such as providing legacy protection to cross-origin script source and cross-principal, cross-process display and event protection. We are the first to provide comprehensive solutions to them.

The implementation and evaluation of our IE-based prototype shows promise of a practical multiprincipal OS-based browser in the real world. In our future work, we are exploring the fair sharing of resources among web site principals in our Browser Kernel.
"

Also, since Microsoft has come under criticism from some investors for still re-investing revenue in R&D in a very tight ecconomy, I wonder if this effort and publicity, even with its tremendous value, is an effort by the company to backup its approach to its business?

Reply Score: 2

Has anyone even read the Gazelle paper?
by ba1l on Wed 11th Mar 2009 15:05 UTC
ba1l
Member since:
2007-09-08

Specifically, the one ClockEndGooner linked to earlier?

http://research.microsoft.com/pubs/79655/gazelle.pdf

Gazelle is not a new browser. It's a prototype of a potential way to isolate multiple instances of a browser's rendering engine.

(Long post - the summary is basically that Gazelle still uses IE's rendering engine, and the paper even states that it'd be stupid to throw that away and start again. IE's rendering engine is going nowhere.)

At the moment, it consists of a basic "kernel" application, and a basic browser instance (running in a separate process). The browser instance actually embeds IE 7's rendering engine, as-is, with virtually no modifications at all. So anyone hoping for a new rendering engine is out of luck.

In the current prototype, it uses the COM interfaces to IE's browser engine to redirect all network access and display through the browser kernel using IPC. The browser instances will eventually be running with no access to OS system calls, and will be required to do everything through the browser kernel, but they've not got that to work yet.

Basically, it's almost the same thing as the existing OP browser, which runs on Linux, uses KHTML for rendering, and split the browser into far more pieces (kernel, UI, networking, and then per-page instances of HTML, JavaScript and Plugins), using SELinux (or was it AppArmor?) to isolate those processes from the operating system by restricting the system calls they're allowed to use.

The difference is basically that Gazelle uses a single process for each distinct origin (using the Same Origin Policy used for AJAX requests), while OP used multiple processes.

While I agree that splitting the JavaScript runtime off from the HTML renderer will cause a lot of performance degradation with no security benefit, I'm not convinced that the same is true of plugins.

Chrome's security model, which the paper spends a long time discussing, is a simplified version of OP's security model, that doesn't split the JavaScript and HTML rendering components into separate processes. Gazelle differs from Chrome's security model in only three ways.

First, Chrome has the option to use less secure process models, while Gazelle does not. There is no support for sharing a single browser instance between multiple origins, as Chrome can. I think Chrome only implemented this as a performance feature anyway.

Second, Gazelle strictly separates content from different origins, including scripts from another domain - a single page could potentially include object from multiple origins, which requires multiple processes. In Chrome, any embedded content is loaded into the same process as the page that contains it, which could potentially allow a security breach.

Being so strict could potentially break websites that rely on JavaScript code to be served from a different domain, which is obviously why Chrome doesn't do it (after all, Google's ad system relies on it).

Third, Chrome considers subdomains to be in the same origin as the parent domain, while Gazelle does not. Again, this could cause security breaches - you would only have to compromise one subdomain to be able to (potentially) affect all of them.

Of course, Chrome actually exists as a real-world product. Gazelle doesn't even implement half of the security features Chrome does yet, such as isolating the browser instances from the OS.

The key things from the research paper were:

- Being this pedantic with same-origin policies, and strictly separating content from different origins, can cause compatibility problems with existing websites, but not as many as you might think.

- This kind of split architecture can be made fast enough to use as the default operating mode of a web browser.

- This kind of architecture can be implemented without significant changes to an existing browser engine. In fact, the prototype didn't require any.

- There's no point in throwing away the existing investment in browser rendering engines, since they can be so easily retrofitted with this kind of security model.

- Microsoft really don't seem to like Chrome very much.

- Microsoft's research department apparently don't have access to the source code for Internet Explorer's rendering engine, and many bugs and limitations of the prototype (including, apparently, the lack of OS-level isolation) are caused by limitations in what they can do with the public COM API.

Reply Score: 6

kaiwai Member since:
2005-07-06

Specifically, the one ClockEndGooner linked to earlier?

http://research.microsoft.com/pubs/79655/gazelle.pdf

Gazelle is not a new browser. It's a prototype of a potential way to isolate multiple instances of a browser's rendering engine.

(Long post - the summary is basically that Gazelle still uses IE's rendering engine, and the paper even states that it'd be stupid to throw that away and start again. IE's rendering engine is going nowhere.)

At the moment, it consists of a basic "kernel" application, and a basic browser instance (running in a separate process). The browser instance actually embeds IE 7's rendering engine, as-is, with virtually no modifications at all. So anyone hoping for a new rendering engine is out of luck.


Awesome post - I was too lazy to read the actual pdf so you've given me a good over view (and happy I didn't waste my time reading something that will yield no real improvements in regards to rendering and compliance).

I think the most telling sign of what is wrong with Microsoft was when a manager (I think it was either the Grand Poobah of Windows development or some team manager) referred to years of crap in the operating system as an 'asset'. When management is so out of touch with reality that they can't differentiate between a liability and asset, you know they've lost the plot entirely.

Microsoft is a compulsive hoarder and its time that the company is taken to therapy to get this addiction sorted out because eventually they're going to drown themselves in the thing they label as an 'asset'.

Edited 2009-03-11 22:58 UTC

Reply Score: 2

JLF65 Member since:
2005-07-06

Being so strict could potentially break websites that rely on JavaScript code to be served from a different domain, which is obviously why Chrome doesn't do it (after all, Google's ad system relies on it).


Now we finally see the heart of the matter.

MS Client: Your new browser breaks our Google ads!

MS: Well, maybe you should switch to OUR ads! After all, you don't want to compromise security just for Google ads, do you?

Reply Score: 3

Funny ...
by inetman on Wed 11th Mar 2009 15:30 UTC
inetman
Member since:
2006-05-30

... in a MSFT research paper you find nytimes and google als references for an overloaded and a very slim website...

what about live.com? Oh wait, they needed a slim website ;-)

Sounds promising anyways!

Reply Score: 1

usr0
Member since:
2006-10-27

...because of its different implementations. The second big issue is CSS. But here the problem is that many features are not implemented by IE at all and not because of a wrong implementation.

ALL alternative browsers support many CSS3 Draft features: hsl, hsla, border-radius (FF only). IE7 does not even support the CSS2 Standard set ;)

Reply Score: 1

BLAH BLAH VAPORWARE
by bandido55 on Wed 11th Mar 2009 22:54 UTC
bandido55
Member since:
2006-10-02

MS has not released IE8 and they are already leaking information about the next browser? It is obvious that when a company instead of promoting a soon to be released product, instead chooses to leak information about a "future" product is because the product they are going to release is below par. It is like MS telling the world, "forget about IE8 (which we all know is a dud when compared to FF3.1, Safari4 and Chrome2) because the next generation is going to be very innovative. I am selling a bridge in NY. Any takers? I am selling it cheap. Vaporware crap

Reply Score: 1

RE: BLAH BLAH VAPORWARE
by Alex Vancina on Thu 12th Mar 2009 18:41 UTC in reply to "BLAH BLAH VAPORWARE"
Alex Vancina Member since:
2006-09-24

MS has not released IE8 and they are already leaking information about the next browser? It is obvious that when a company instead of promoting a soon to be released product, instead chooses to leak information about a "future" product is because the product they are going to release is below par.


I think you're missing the point. It's a research project, not a product. Microsoft Research publishes papers about things they’re working on all the time. You might as well also say that Microsoft thinks Windows 7 is dud simply because Microsoft Research is experimenting with a totally new operating system, Singularity.

You, sir, are a troll.

Reply Score: 1

LOL... Net+
by Kancept on Thu 12th Mar 2009 01:20 UTC
Kancept
Member since:
2006-01-09

Man, how funny is it that I just read that article in BeOS R5 running Net+? Couldn't post a comment, but I bet it'll make for some neat server logs reading...

Reply Score: 2

OMG another instance of wasted resources
by Shagbag on Thu 12th Mar 2009 19:57 UTC
Shagbag
Member since:
2009-03-12

MS use its own software rather than someone else's standard? I shake my head but I'm not surprised. MS wants to own everything, but is what they're doing really the most efficient allocation of resources? As a trained economist I can tell you that one of the downsides to monopolies is that resources are wasted in an economy.

Reply Score: 1