IEBlog

Hopefully the next IE 8 final version comes out early november.

I work as a web developer for a lining, and I can't see why Microsoft is even bothering fixing it's extremenly buggy browser... if they would just switch to the Gecko engine they'd save money, and make it easier for us all to develop webpages... ;)

Good, Better, Best!

The Microsoft IE Team rocks! :-)

If I can help at all please let me know.

 Salute,

   Mark Wisecarver

   Technical Evangelist

How nice that InternetExplorer is finally getting the press that it deserves... Perhaps now, it will gain a real percentage of the market share.

All joking aside - I have high expectations for IE8. Keep up the good work.

Great to know, however if you still have time to fix a CSS 2 bug...

In previous versions of IE, you could define a negative text indent on elements (negative indents are valid in CSS); this worked on elements which didn't have layout (hasLayout=false) but not on elements which had one (hasLayout=true).

Now in IE 8, hasLayout is said to be gone, and the problem isn't visible anymore...

Except on pseudo-elements containing images.

If I do the following:

mytag:before {content:"[my block]";display:block;text-indent: -5em}

then the generated content will be indented. However, if I do:

mytag:before {content:url("myimage.png");display:block;text-indent: -5em}

Then the image won't be indented, although it should be.

I know that generated content couldn't contain images in IE8b1; it was said to be fixed in IEb2.

Obviously, you forgot something.

>> I can't see why Microsoft is even bothering fixing it's extremenly buggy browser

Good point, but let's be honest - we spend too much time hacking our clean (compliant) code to make it work for those who just don't get the whole windows update thing and are still using IE6.  At least if there is a standard-compliant IE on the market (wink wink MS) then there is a *chance* that one day our lives will be simple and happy... :-)

>> Perhaps now, it will gain a real percentage of the market share.

If it truly does deliver a safe browsing experience twinned with standards-compliance, then it will deserve good press and the market share.  The IE team are under a lot of pressure to deliver!

I think that's the nicest thing I've said about IE all year....

Thanks for sharing! And the comments still don't work properly. I always get an error when trying to post a comment!

Canvas support in IE8 would be nice...

...IE8b2 doesn't support negative bottom margins on generated content either (interestingly, it does support negative top margins - it reminds me of IE 6's top/bottom/left/right positioning bug), whether it contains text or images is irrelevant here.

A practical example can be found on http://moneyshop.perso.cegetel.net/moneyshop/index.php

Compare with Firefox's, Opera's or Safari's output, you'll notice:

- top right corner is missing: text-indent doesn't work when content is an image (which, as an inline box, should follow the rule)

- bottom corners don't appear: margin-bottom isn't respected when margin is negative

- a piece of generated content is missing altogether: the small hand after clickable headers (h2 a:after) won't appear, even if its content is replaced with text

Note: rounded corners will seem to work in IE 7 and older, as I added a workaround for them.

On Page 8, Doug Stamper states:

"Microsoft Internet Explorer team's goals for final...   including early HTML 5 support..."

Just what features of HTML 5 are you planning to include? This is all news to me.

Or is this Pie-In-The-Sky Marketing schlep?

Now Hold On Just A Minute Here!

From Page 10: "Developing Standards Mode Pages"

1. Author your pages according to the latest Web standards as supported by Internet Explorer 8 beta 2.

WTF??!?!?!?! Are you serious?  No, No, NO Freaking way No!

MICROSOFT DOESN'T OWN THE WEB - Web Standards are just that, Standards.

Microsoft doesn't own them, nor make them.

Developers should code against the published standards of the W3C, HTML, XHTML, CSS, ECMAScript, etc.

They should then adjust their sites to provide a downgraded but usable experience where a browser doesn't support a given standard. (e.g. native SVG, styles and events on Select lists etc.)

I Hope that the author of this content wasn't on MSFTs payroll.

- Total Public Relations Fail

@TakeWithSalt: IE8 includes support for DOMStorage, postMessage(), Connection Events, Window Targeting rules, and a number of other minor (stable) elements of the HTML5 spec.

@EricLaw [MSFT] - ah, ok, so not actual markup support or XHTML, gotcha.

I didn't read about the "Window Targeting rules" (going to right now), but is this what broke frame navigation in IE7/IE8 beta 2? where every link targeted at a frame that is opened in a new tab still thinks it is in the parent frame structure and crashes hard when it tries to navigate up in the structure? (spawned links in new tabs always "think" they have a parent when they don't)

The user agent sniffing code listed as the "correct way" on page 17 is by far some of the worst agent sniffing code I have ever seen.

Lets not even mention the complete lack of other better browsers from the sniffed applications, but the alert message of: "You should upgrade your copy." is not how to inform users.

If I'm on Expedia.com and this alert pops up, what is it telling me? Upgrade my trip info? huh?

Then of course the issue is that you're trying to tell an IE user to upgrade, when so many of them are locked down on a corporate environment where they can't choose a better browser (Firefox, Safari, Chrome, OR IE8).

$SaltGuy, HTML5 explicitly is *not* XHTML.  XHTML is basically a failure which is why HTML5 even exists.

Where's an example of your crashing page?

$Bryce, it's pretty obvious that some business-type is going to write the strings that the user sees; the point of their code sample is to show how to detect the IE version.  

As for your claim that it's "the worst you've ever seen"--- well, what exactly is wrong with it?  The point is to detect the IE version, which it does just fine.

Is IE8 going to fix the issue with SPANs inside an A element not showing the correct cursor?

<a href="foo.html"><span>Go Foo</span></a>

When a user hovers over the "Go Foo" text, or any part of the A element, the user should see the default cursor for a link (e.g. the pointer).  In IE, the user sees the "I bar" thing.

Since this is a beyond obvious fix for child elements of an A element can we get a status on this issue being fixed in the next beta?

IEBlog is quite pointless. Most post are here boring as hell. Firefox, Opera, Safari and Opera keep innovating and IE8 folks here are tooting about technical nyances. Who cares to be honest. Many on "new" IE8 features are already here in other browsers. Adhering to web standards cannot be feature, it is a requrement. When you are goign the release the damn thing? Meanwhile post something really useful and interesting.

This compatibility mode thing isn't working.

There are 2 scenarios where this fails.

Scenario 1.) You have multiple applications on a given domain, running on DIFFERENT PORTS.

:80 Main End User Application

:7000 Application Server Interface

:8000 Application Database Interface

In most cases, the End User App is up to date, running standards based code (e.g. IE8 Standards mode), but the App server and database server interface are running legacy code in IE7 mode.

Since the Compatibility Mode is set per domain (not domain + port), this scenario fails in IE.

Scenario 2.) You have multiple applications on a given domain, running on DIFFERENT SUBDOMAINS or DIRECTORIES.

www.uberwidgets.com

www.uberwidgets.com/daves/ultramapper/

www.globotools.uberwidgets.com/

www.blog.uberwidgets.com/

You might have a blog running (like the one that runs this site) that can't handle modern standards, but the main application can.

Or there may be many applications (by different users/companies) running on a subdomain or subpath for a site. This would be quite common for a widgets based site.

Since the Compatibility Mode is set per domain (not subdomain + domain OR domain + path), this scenario fails in IE.

Finally, since these scenarios (as well as the IntRAnet/Localhost) issues keep causing so much pain, can you make the Compatibility Mode button either (ALWAYS VISIBLE), or an option to make it so. (I don't care if you disable it when a site sends in a META tag to force IE7 mode), but it is a royal PITA when I can't change it to test something because one of the (4 scenarios) above causes IE to fail.

IE8 Beta 2 (standards or IE7 mode) as well as IE7/IE6 in a very interesting way.

when a form button on pageX.html does this:

<input type="button" value="Go to Y" onclick="location.href='pageY.html';"/>

If a user clicks the button, on pageY.html the following behavior is observed in PHP.

Firefox, Safari, Opera, Chrome:

$_SERVER['HTTP_REFERER'] = pageX.html

Internet Explorer:

$_SERVER['HTTP_REFERER'] = undefined

Now is this an issue with the way PHP (v5) handles IE? or does IE not supply the referer info?

Thanks

@webster - yeah its a bug in IE (not PHP).  Using Fiddler you can see the HTTP_REFERER being sent in if you use a browser like Opera (Firefox uses its own client, not wininet) but when IE (any version) uses the location.href syntax to go to a new url, IE DOES NOT Pass the referer (in the miscellaneous section of fiddlers headers).

Which is really odd because if you click on a link with this syntax, it works just fine:

<a href="pageY.html">Go to Y</a>

Though if you do this, you will not get the referer either (#2nd copy of bug)

<a href="javascript:location.href='pageY.html';">Go to Y</a>

this fails too:

<a href="#" onclick="location.href='pageY.html';">Go to Y</a>

Actually Eric L. (the guy who deals with a lot of this HTTP traffic stuff might be able to either fix the bug or explain why IE behaves differently than every other browser)

Gerald and webster, why would you write code like that? That's just silly! If your reason has to do with what shows in the status bar, it doesn't the fact that's silly!

1. Anchors do what you want to do without the redundant JavaScript.

2. Why are you placing events inline?

3. Why would it matter what shows in the status bar? Seriously?

Why you're at it, why not slap a form on the page, change the action, and use form submittals to navigate everywhere (hint: try doing that and see if resolves your referrer issue)?

@Glen Fingerholz - You are missing the point completely.  Using an Input button to provide navigation is as common place as any other technique and was common long before the web.

Wrapping the button in a form doesn't help either, because you have to submit the form, and ALL of its contents to use form based navigation.  Since forms can not (by spec, and enforced by browsers) be nested, you can't have 2 or 3 buttons on the same form navigate _CLEANLY_ to multiple URLs by using submit.

IE's lack of referer on location.href changes is just silly and is a well known, never addressed (or acknowledged (AFAIK - 5min on Google)) bug.

My guess is that now Microsoft has a catch 22 issue here.  If they fix this in IE8, then they have to admit it was a huge bug in IE6, and IE7.  So far Microsoft has been very careful to phrase their "support of web standards" as just that, "support", not as "we fixed a huge pile of massively broken DOM implementations, a.k.a. bugs"

Now for the good news!

There is a hack to workaround this bug in IE.

Regardless of where you call location.href to set a new url, call location.setHref(url); instead...

Then define this function in your global library:

(note, 'IE' is a boolean indicating if the user agent is MSIE - use whatever detection you want/need accordingly)

/* location.href FIX for MSIE */

location.setHref = function(url){

 if(!IE){

   location.href = url;//Standards based browsers

 } else {

   var lha = document.getElementById('_lha');

   if(!lha){

     lha = document.createElement('a');

     lha.id = '_lha';

     document.body.appendChild(lha);

   }

   lha.href = url;

   lha.click();

 }

};

So how does this work?

1.) If the browser is a standards based browser it just uses the location.href sytax and all is well.

2.) If not, then we do something special for IE.

a.) Check for a bogus link in the page, if not present, create it, assign an id, and add it to the page (note it doesn't have any text content, thus is effectively invisible)

b.) Set the url of the link to the desired url

c.) Call .click() on the link, which has the same effect as a user clicking on a regular hyperlink, navigating to the page.

This is obviously a bit of a cumbersome workaround for a simple bug, but seeing as this wasn't fixed by IE8 Beta 2 yet I don't see this fix being high on Microsofts priority list.

Well its November, any chance of a release date???

@Gerald and @Webster

I agree with Glen Fingerholz.

A link should be used to load another webpage into the same window instance, not a command button.

"javascript:" pseudo-protocol-links should be banned everywhere with the unique exception of bookmarklets. This recommendation is now a widely admitted, known one. "javascript:" pseudo-protocol-links always give a variety of problems (accessibility, usability, etc).

@Scott Dickens [MSFT]

The "Making Your Web site Compatible Across Multiple Versions of Internet Explorer" article has numerous errors and is definitely NOT recommendable. The code examples have validation markup errors, uses a non-recommended (by W3C) doctype declaration, there are also incoherences in the code.

The example on automatic counter and numbering is simple but when one tries bug #96 at my webpage, IE 8 beta 2 fails miserably when all other web standards compliant browsers (Firefox 3, Opera 9.50, Safari 3.1.2, Konqueror 4.1, etc) succeed. So even the provided example in that article is not at all convincing.

The whole section on user agent string detection should be removed in my opinion.

User Agent string detection

---------------------------

A wide majority of expert web designers will recommend to use object/method detection support first before resorting to user agent string detection. Conditional comment (when used for judicious targeting like working around a particular bug, a specific IE7 incorrect implementation) should be used before even thinking about using user agent string detection. This is not the first time this has been said in IE blog.

Regards, Gérard

@Scott Dickens [MSFT]

The W3C markup validator will report

"

the Document Type (-//W3C//DTD HTML 4.01 Strict//EN) is not in the validator's catalog

"

The possible valid doctype declarations are here:

www.w3.org/QA/2002/04/valid-dtd-list.html

On user agent string detection:

A Strategy That Works: Object/Feature Detecting by comp.lang.javascript newsgroup FAQ notes

jibbering.com/faq/faq_notes/not_browser_detect.html#bdFD

Browser detection - No; Object detection - Yes by Peter-Paul Koch

www.quirksmode.org/js/support.html

Javascript Best Practices by Matt Kruse:

Feature-Detect Rather Than Browser-Detect

www.javascripttoolbox.com/bestpractices/#detect

Regards, Gérard

@JP Gonzalez-Castellan[MSFT]

Text-selection with keyboard (with caret browsing) and text selection with mouse-dragging can still be very difficult or impossible to perform in IE 8 beta 2. One example is bug #151 in my webpage.

Your "Listing 1: Sample Web page to understand better the new Zoom" lacks a doctype declaration and that triggers quirks mode along with the broken CSS 1 box model.

If you and all of your MSFT/MSDN colleagues would always use a W3C recommended doctype declaration, preferably with a strict DTD (to ensure clear separation of content and presentation) and with a full URL, then MSFT/MSDN would be in fact promoting what should have been promoted, recommended to web authors more than 9 years ago to trigger standards compliant rendering mode and compliant CSS 1 box model.

Regards, Gérard

Its not that i dont like it, its just a beta i want internet 7 back until all the bugs r fixed but i cant uninstall it, i checked in my program removel and my update history its not there, if someone could help that would be great

@gerald:

You can use Fiddler with Firefox easily, just read the docs: http://www.fiddler2.com/Fiddler/help/hookup.asp#Q-NonIE

Sometimes, I'm really surprised that those who cry for standards the most, are the ones working against them the most as well.

This applies to anyone, really.

Its November, is there any kind of release update

There are folks posting on this blog as if there will be a final RTM this month (November), folks that think an RC will be out this month, and folks that believe there will be a beta 3 out this month.

Since I am firmly seated in the last group, e.g. I'm in the group that thinks IE Beta 2 is NO WHERE NEAR READY to release, I'm getting very worried that one of the folks in the first group might actually be correct (that transparency thing hasn't worked yet for the IE Team)

So, what is the deal?

1.) Is there going to be a release of any kind this month.

2.) What will the next release be? Beta 3? RC1, or RTM?

Since many of us have sites that will fail miserably if IE8 Beta 2 is retagged as RTM I desperately want to know what I am going to have to do.  I certainly don't want my end users to have to use compatibility mode as that would indicate to the world that we are just shy of incompetent, but at the same time we have no idea if the existing set of bugs in IE8 beta 2 (new ones and old ones) are here to stay or if they will be fixed.

Our hands are severely tied since MSFT makes NO efforts to lets us know what their delivery schedule is like or what it will entail.

If there is going to be an RC before the RTM, great, we certainly need one, but if there isn't going to be one, we certainly need to know which beta is the last.

I'm tired of pulling out my hair with IE and losing sleep over having to do all the bug discovery/tracking myself.

I am 110% serious when I say that Microsoft's disclosure of the IE8 delivery process is beyond pathetic and highly un-desired in this modern age.  We understand that all bugs can't be fixed, but we do need to know what to build our code to work against, and this wall of the unknown is really starting to irk us.

Seriously, get a public bug tracking system online ASAP!

Seriously, get a public release schedule online ASAP!

Seriously, get a public roadmap online ASAP!

Seriously, provide info for developers that need to build code to work in IE... we can't deal with not knowing when IE will ship!

For the love of (deity-du-jour), please at least indicate if there will be an RC release before the RTM.  I can tell you right now that I'm not making a single line of code change until I know what the $#@! I am coding against.

For all the readers of this blog, is there a definitive list of the known bugs in IE8 at the moment, indicating which are new, which are regressions, and which old bugs are only partially fixed.

I'm mostly concerned with the list of regressions, since my sites just tank on IE8 beta 2 right now, and if I have to hack through the night when IE8 ships, I'd like to know where I have to focus first.

Worried Sick

I downloaded IEv8 and apparently the Internet Explorer cannot display some the webpages! No fix for this issue so far!

I have 4 website to my home page tabs, The first tab is not colored "default color" and the 3 tabs is colored "red".

I downloaded IE 8 Beta 2. no change performance between IE7. it strain for new tab.

why save page is slow and when ie is saving a page we cant see other tabs?

I hope release version of IE8 will be adequate of Windows 7.

Watching the videos of PDC 2008.  Nothing is stated about the next release of IE8 (is it beta 3 or ?)

Can someone verify?

@Pete LePage: Just heard you jump in on the PDC2008 video with Alex M.

I haven't laughed so hard in a long time...

"H" "H" "T" "P" "colon" "whack!" "whack!"...

OMG, you are a PM on the leading Web Browser?

man, I've got coffee everywhere now.

The key you are refering to is known to the rest of the world as the "slash" key.

helen-- Technically, "whack" is a shorthand for "the backslash key", not the slash key.  Of course, if you type http:\\www.ebay.com it works anyway.

And Pete (http://blogs.msdn.com/petel/) is a product manager, not a program manager.  very different at MS.

Saad Sh-- Run IE without addons.  See that it's no longer slow.  Now go disable the slow addon.  Notice that IE8 is fast.  FTW!

>> we spend too much time hacking our clean

>> (compliant) code to make it work for those who

>> just don't get the whole windows update thing

>> and are still using IE6.

I'm sure this has nothing to do with the fact that IE7+ isn't runnable on anything below WinXP.   No sir.

@Stan - Helen never said he was a Program Manager? she said he was a PM.  Regardless of the title, someone in the browser world that doesn't know how to spell a URL out loud needs to reconsider speaking publicly about a web browser.  I don't know what side of the moon you come from but the keys are slash and backslash.  In 20+ years with computers I have NEVER heard anyone call it a "whack".

Its as anoying as those people that think that this is somehow easier/shorter to say, and just end up looking like tools.

"dub" "dub" "dub" dot google dot com

instead of

triple W dot google dot com

I'm pretty sure anyone that walked into our dev shop for an interview talking about pressing the "whack" key wouldn't be getting a call back, thats for sure.

trevor: you must be new here.  in use on this blog, pm means "program manager".

Suggesting that you, personally, are aware of

all terminology in use worldwide arrogant to the point of silliness.

everyone *I* know says "whack" instead of "backslash"-- it saves syllables.  In case you're wondering, yes, everyone also says "dub" instead of "W", for the same reason.  Personally, I think anyone who says "triple W" is a fool, but hey, do whatever you want.

Hmm... your "dev shop" vs. the world's most successful software company.  Based on results, it sounds like your hiring criteria need some work.  Of course, you'd have to be smart enough to realise that.

Whack is pretty common for forward (or even backward) slash.  Even wikipedia has it in the summary paragraph for both slash and backslash.

http://www.computerhope.com/jargon/w/whack.htm

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

:)

I personally use slash and backslash, but I kind of like Hack and Slash because it reminds me of those Reboot cartoons from when I was little.  I also use the wordy "Exclamation mark" instead of "Bang!" when I can't get away with "not".

I have to say, that's the first time I've seen "triple w" as a short form for how I say it (double-u double-u double-u).

Depends on how you consider Microsoft "successful"... If it's in software quality, although huge laps (in a good way) were done these last few years, unfortunately success doesn't equate to quality; if IE beta 2 is any indication, I already stumbled upon 3 (three) large regressions and major bugs that make my website unusable, while other beta stuff like Chrome, or developer's build like Shiretoko worked perfectly the first time.

For the bugs, see my other posts above.

Some people enjoy using proper terms for things; slash and backslash, being the proper term, are unequivocal, while "hack" and "slash" may not work in an international environment where slang words don't translate that well. The same can be said about "www", which may be shortened to "w3" (wee-three), "triple doubleyou" or "doubleyou three"; it remains that the only unequivocal way to say it is "doubleyou-doubleyou-doubleyou", as "w3" may actually be a proper subdomain (or www3), and that, in short, relying upon slang terms and shortcuts is the best way to blow up a project.

Ask NASA how much it cost it to not put the measure behind a drop-off altitude. Sure, it saved 3 letters and a pair of syllables, but it also cost a few million dollars...

Personally, I'd rather work with someone not using slang terms, thus judged "outta teh hype LOLz", but who knows his job, what AJAX really means (no, it doesn't mean you gotta load 250 Kb of gzipped Javascript libraries, but that you know how to dynamically modify a DOM from XML-encoded data retrieved with Javascript through the XmlHttpRequest object - use libraries to do so if you wish, but understand how they work), and what is actually hiding behind "Web 2.0" (no, it doesn't mean Flash nor Silverlight, it means that the browser can be used as a platform - regardless of what language you'd rather use to do so; for some it's ActionScript, for others it's Javascript, SVG and CSS).

Maybe I'm wrong; maybe, now, AJAX means that you gotta use one of these libraries, otherwise it's not AJAX it's only asynchronous Javascript and some XML thingy. Maybe Web 2.0 means you gotta have videos, feeds, blinks all over the page, and that has to be done through a plugin (nevermind HTML 5, SVG and JIT Javascript).

Strangely, when I ask hipsters what AJAX means, not many of them can explain; when I ask them to explain "Web 2.0", they go off in long tirades on video stuff with buttons you can click to rewind or include in your blog, and use your webmail off-line.

Smells like Marketing. And the 2k bubble.

@Stan - you live in a special place then. I too have never heard anyone calling it whack.  In fact I think calling it that, is whack.

Just where abouts in the world do you live? I live/work in NYC and travel to Chicago, Denver, Miami, and San Francisco every year.

In our dev group you wouldn't *not* get hired because of it, but we would certainly have a good laugh after you left, and if you joined our company you'd either re-learn "slash" and "backslash" or face giggles every time you say "whack".

@Mitch74

Can you create testcases of each single problems you experienced with IE 8 beta 2? At this point, the only thing which will help relevant people is by providing *_reduced_*, minimized testcase of each of the problems you experience.

Again, a real webpage is not what IE developers are looking for but minimized testcase with a clear pass condition, with as little code as possible.

Then upload those testcases on a web accessible site and then report the URLs with the "Report a webpage problem ..." IE Addon or post the URLs in this blog.

Regards, Gérard

Finally Figured it Out!

Finally Figured it Out!

Finally Figured it Out!

Finally Figured it Out!

IE8 Standards Mode REGRESSION BUG.

IE8 in Standards Mode will **NOT** (I Repeat, NOT) fire the resize event on the window element, this is on ALL window objects including iframes and frames.

I consider this to be a fairly critical regression bug, since there is no workaround for this.

Arghh No wonder sites look so messed up in IE8... glad I finally figured this out.

Please advise if Beta 3 has a fix for this.

steve

Steve that is an IMPORTANT BUG!!!!

@steve_web

> IE8 in Standards Mode will **NOT** (I Repeat, NOT) fire the resize event on the window element

Steve, You may want to validate and to vote for bug 364571

connect.microsoft.com/IE/feedback/ViewFeedback.aspx?FeedbackID=364571

> this is on ALL window objects

> including iframes and frames.

No, not a regression for frame objects. IE 7 was not firing resize events for frames (don't know about iframes).

URL: www.gtalbot.org/BugzillaSection/FramesetMetaTestCase.html

Steps: resize top frame and the counter won't increase.

One last thing. Giving a meaningful, helpful and sensible title to a blog post message can and will help IE team people when they could (in the future) look for various kinds of bug reports or various types of enhancement requests.

Regards, Gérard

* Still no tool for debugging memory leak issue which is still a problem in IE8

* Combobox/dropdown list still have the fixed width problem - when it is set to fixed width, the dropdown menu width is fixed as well!

So I'm looking forward to a new IE yet. please work on those!!!

* Not passing Acid 3 tests. WebKit, Firefox, Opera and even Chrome are miles ahead.

* No Canvas support.

* Still leaking memory.

* Still not taking this browser serious.

do i just delete the webslice shortcut in Favorite Bar to unsubcribe? how do i unsubcribe the a webslice?

thank you

www.dorra-2009.blogspot.com

Final release is in November right

@lezby: No, no one at Microsoft has provided a release date.

<<Still not taking this browser serious.>>

Then you're not a serious person.  By this time next year, IE8 will be the most popular browser on the planet.  For serious web developers, that's plenty of cause for serious consideration.

Stan: you're deluded, it'll take longer than that for IE8 to spread. It's not as if IE7 became the most popular browser within a year after its release.

I notice that Windows Explorer also have information bar when opening network and the highlight is light blue when mousing over the information bar and much better color than the IE 8 Beta 2. Change the blue color highlights to light blue when mousing over the Information bar(Similar color when mousing over refresh/stop button)

<<It's not as if IE7 became the most popular browser within a year>>

Data for your claim, please?

Here's mine: //en.wikipedia.org/wiki/Internet_Explorer#Market_share_by_year_and_version

It's pretty annoying that anything with a hyperlink in it is treated like spam by this comment system.

@Stan - keep a few things in mind.

1.) Pre-WinXP users still can't update IE

2.) Many IE6, IE7 users will not move up

Between Firefox 3, Firefox 2, Chrome, Safari, Opera, IE7, and IE6... I don't see IE8 stealing enough market share even from "within" the IE install base to garnish the #1 spot.

Based on the number of issues with IE8 at the moment I do not see upgrading being a mass-appeal thing when it goes from beta2 to beta3, beta3 to rc, or rc to rtm.

Most developers are not going to change any of their code until IE goes to RC after the next beta so they know what they are coding against, and many won't update their old code due to lack of access, time or lazyness.

I wish IE8 lots of success, but not for an RTM before 2009.

Errr... yes, there's a number of users that are still on the internet using Windows 2000 and lower.  Of course, when rounded to the nearest %, that number is 0%.

Chrome has <1% share, as does Opera.  Safari has somewhere in the neighborhood of 5%.  Firefox2 falls out of support in a few weeks, and has virtually no share anyway.  Firefox 3 has plenty of share, but it's not meaningfully growing.

By the end of 2009, IE8 will be the most popular browser in the world.

On the second page of the "Making Your Website Compatible ...." article, it says to use X-UA-Compatible: IE=EmulateIE7, but everything else says to use IE=7. Which one should be used? I thought I had it all figured out, even with the changes between beta versions, but that just confused me.

@rpgfan3233: The X-UA-Compatible tag differences are as follows:

"IE=7" means "Render this page in IE7 strict mode."

"IE=EmulateIE7" means "Render this page in IE7 strict mode UNLESS it's a quirks mode page, in which case, render it in quirks mode."

Using the latter helps ensure that your page works as it did in IE7, by preventing quirks pages from being forced into IE7 strict mode.

Number of Win2000- systems is closer to 3%, if http://marketshare.hitslink.com/report.aspx?qprid=10 is to be believed. That is still vastly lower than I expected.

@Stan - I'll take that bet. Nov 5th, 2009 IE8 will not have the #1 share.  It will be either IE7 or Firefox 3.

The biggest issue is lock down enterprises that won't upgrade their systems due to I.T. depts that won't move forward.

The good news is than many advanced companies now allow users to upgrade their own apps.  So when a user wants a better browser than IE6 or IE7 they can upgrade to Chrome, Safari or Firefox.

So far the number of users that upgrade to a better browser and then downgrade back to IE7 or IE6 is near zero.

The best thing about an open web is that users get to pick the best tool for the job.  One day that might even be IE.

Firefox 3 biggest feature was the awesome bar but IE team really did a good job on the smart address bar by making it organize and adding a autocomplete suggestion "shift+Enter" Another reason to use IE again. My only complaint in IE 8 beta 2 right now is stability and performance issues sometimes the address bar would not show the website title and url link.

<website has not set a title> or .url or the registry location would show up or the windows explorer location of the shortcuts. The IE 8 beta 2 browser crash 13 times compare to IE 7 5 times shown by reliability and performance monitor.

Stan: in pretty much all of the public web browser stats I check out, as well as the stats of my own stats, I can tell you IE7 only surpassed IE6 a few months ago. Meaning it took almost 2 years rather than 1.

Check the archives for these sites for example, if you will (like through http://www.archive.org/ or whatever):

http://www.upsdell.com/BrowserNews/stat.htm

http://www.w3counter.com/globalstats.php

http://www.w3schools.com/browsers/browsers_stats.asp

@Gérard: my bad! Most of the problems come from an obvious reflow problem in IE8, as reducing the page to its simplest form show pseudo-elements' content:url("image.png")... Sometimes. Pressing F5 repeatedly will display the image in one pseudo-element, then in the other, then in both and then in none.

The more complex the page, the less likely is generated content to appear. This may just be fixed along with the huge reflow bug: reset scrolling on DOM modification.

As soon as I've put together enough test cases, I'll publish them here. For now, this one is:

http://moneyshop.perso.cegetel.net/moneyshop/testcases.html

Precision to the above: another thing that triggers this peek-a-boo bug is resizing the window enough to trigger a reflow.

@Gérard: there! I finished the test cases, all in http://moneyshop.perso.cegetel.net/moneyshop/testcases.html

To see the errors, load/reload (F5) the page's content, resize the window, refresh when scrolling is at the page's bottom...

It seems there is 'only' one bug at play: pseudo-element's image content isn't reliably loaded nor displayed. It is not a resource loading problem though, as there are only 2 different images in this, and they are several times - but not always displayed.

It seems the upgrade from beta 1 wasn't complete.

@Mitch 74

I created a much simplified, minimized testcase from one of your tests:

www.gtalbot.org/BrowserBugsSection/MSIE8Bugs/negative-margin-generated-content.html

and I'll report it to connect's IE beta feedback

A few notes regarding your webpage:

- text-indent should be used only for the right purpose, for the best element (one with text, obviously) and in the most adequate code context and not for positioning purposes, layout control

- your CSS code is definitely over-constraining a couple of elements and my experience taught me that this sort of coding approach is bound to break in browsers, in different context (text size increase|decrease, viewport dimensions modified, etc). Overconstraining, over-declaring and over-coding is most often the sign, the path of a code that will eventually fail and that is over-excessive for the real goal

- if all your code is aimed at compensating for absence of border-radius support in IE 8, then your "solution" is over-excessive in comparison to the problem (absence of rounded corner). Just live with such absence of support: no IE 8 (and lower) user will ever complain about such missing rounded corner. And your code maintenance will be a lot more easier.

Regards, Gérard

Stifu, the fact that you even mention the w3schools data shows how little research you did.  Even THEY admit their stats reflect nothing about browser share on the web at large.

blog comments borked again, submits to the index page, not the post page.

@Mitch 74

I posted bug 380412:

[GC :before content url()] Negative margin is not applied to generated content

connect.microsoft.com/IE/feedback/ViewFeedback.aspx?FeedbackID=380412

Regards, Gérard

@Compatibility Mode Fail: It's even worse than that.

If I go to a site such as www.ssc.govt.nz and add the domain to the compatibility list, it adds the entire govt.nz second-level domain. If I then go to another government site such as www.justice.govt.nz then it inherits the settings from the other site.

So far I haven't heard anything about whether this will be fixed in a future release, or whether the entire govt.nz second-level domain will be affected by this issue.

Hi IE developers,

Please tweak IE to pass ACID3 Test. Please focus on how you fix this test first before anything else. For those who doesn't know about ACID3 Test, please search them with your favorite search engine. :)

Nene-- I prefer the IE team work on things that matter to normal users like me.  All this acid stuff is just web developers doing drugs.

is compatibility mode going to be fixed in the next release?   there seems to be so many problems with it and with IE8 not being ready for prime time the button is very important.

also can has someone explain when the next beta of IE8 is going to be?   i will be taking holiday for next 2 weeks i don't want to take my notebook but if there is release i would like for test.

much thanks

hav

IE8 is the slowest Next Gen browser out there...

Compared to Safari, Chrome, Opera, and Firefox 3, especially 3.1 with tracemonkey, IE 8 gets owned.

I hope this gets remedied, I'm not sure how it will, but I hope it does.

hrothgar, IE8's Javascript engine is between 5 and 20 times faster than IE7's.  

Sure, the other browsers have faster script engines, but there's more to performance than just script performance.  Given that most sites spend maybe 200ms in the script engine, the fact that it's perhaps 100ms faster in some other browser is hardly something a normal person would ever notice.

Hey IE team, what is this I'm reading: www.appleinsider.com/articles/08/11/06/microsofts_ballmer_considers_using_webkit_within_ie.html Does this mean WebKit will take over in IE9? I will *NOT* accept such a browser after so much development and investment of my work, time and money around Trident. Tell that to Mr.Ballmer.

I reallly hope this works;it's 6 in the morning and to be honest I'm tireds of all this stuff, so PLEASE SIGN ME IN!!!

I reallly hope this works;it's 6 in the morning and to be honest I'm tireds of all this stuff, so PLEASE SIGN ME IN!!!

anonimous: this is just stupid. Whether they use Webkit or make Trident better, the point is to improve the rendering and make things better for developers. The fact it was worse before is totally irrelevant, we *don't* want things to stay the same, we want things to be cleaner and easier.

Sure, they could have switched to Webkit earlier then, but that's beside the point. We want IE to improve, and its old bugs to be gone.

@anonimous - I read that link too!  Yeah, glad to see that Steve Ballmer has a sense of humor and is aware that IE isn't the greatest rendering engine out there.

As for MS using it, why not?  I've spent tons of time tweaking for Trident, true, but I would drop it all in a heartbeat if they swapped it out for Gecko, Webkit, or whatever Opera uses.  in a heartbeat!

MS could keep all the other stuff that they might want, but with a better rendering engine, we would have 1 less level of testing/debugging our content for IE.  That's what I love about Google Chrome, its a new browser, but if your stuff worked in Safari, it pretty much works perfectly in Chrome, with no tweaks at all.

Stifu, you're right, we don't want to let things stay as they are.

As I found out in this "magazine" (?) IE8b2 includes a 4th mode (Almost Standards Mode). This is the worst decision MS made for IE8 so far, and I hope they'll undo this for the final release. See Bug 379343 (https://connect.microsoft.com/IE/feedback/ViewFeedback.aspx?FeedbackID=379343)

The problem is:

* The only bug emulated by this mode is easily fixed and doesn't need a workaround on the engine.

* This is so inconsequent. This is one of the smallest problems websites would encounter, when they switch from IE7 Standards mode to IE8 standards mode.

There are so many ways for web developers to mitigate this problem, yet the IE devs seem to feel it'simportant.

* Sad is, were MS not implementing this unecessary mode, Mozilla, WebKit, Presto and whonotall could also remove their Almost Standards Modes. But maybe that's nothing MS likes to think about?

I hope, and pray, that this functionality will not make it into IE8, for a better future of the Web.

Is beta 3 coming this month? I'm looking forward to fixes for all the issues in beta 2 so I can start adjusting my production code to handle IE8.

I would hope that after beta 3 that there wouldn't be (m)any more significant changes... right?

I'm not going to code against beta 2 of course because it is still to far from ready but I'm anxious to start finalizing code to handle IE8's changes (for good or bad).

thanks

randal

If they fix all issues in Beta 2 then final version will come out this month

Brez: most useless post here. :p

It's not like any browser vendor, let alone the IE team, fixes all issues before releasing a new version.

IE8 may be a "rendering engine", but it's not a "browser".

Browser lets users browse the web sites. IE8 just crashes 5 times/minute instead.

Please rename IE8 to the more appropriate "Windows Crash Exploder".

IE8 has a really great AJAX support. It's true.

I can't say I won't be surprised if there is no beta 3.  I don't think IE 8 is ready for prime time so to speak.  Then again to quite a few beta testers of Vista it wasn't ready yet either. My scenario is Microsoft may give up on IE 8 and release it and concentrate on IE 9 to address the myriad of issues that seem to be plaguing IE 8.  Off topic here, Windows 7 still carries on the locking up of the UI, can't move the menu, address bar, tool bar, this stance started with Vista/Longhorn ver 4074 and carries on to this day. Four years of the same complaints and still no end in sight.....

(hey do you think I would not throw in a jab at the pathetic UI)

@Eghost: No one's "giving up" on anything, and the suggestion is ridiculous.  The IE team remains hard at work on IE8.

@Fduch: I assume you've tried running IE without addons to eliminate those as a source of problems?  (www.enhanceie.com/ie/troubleshoot.asp#crash)

Do you see the same crashes when in Compatibility View?

We've made tons of reliability improvements since beta-2, so rest assured that the next public build will be far more robust.

@Gérard: the code I'm using is here essentially to cover a problem not in IE, but in CSS in general: there's no way (for now) that allows you to say that an element will represent 100% of the width of its parent+10 pixels! The way I chose was to say that the box _is_ 100% wide, and its content's overflow goes a measured way to the left. Then all I need to do is shift the box with negative margins (for browsers that go the CSS 2.0 way, instead of CSS 2.1: can't set its position as relative). What remains though, is to make sure the box and its content have the same height, and while it's easy to set it up for the box, its content (being text) is much more finicky: gotta get creative with line-height and font-size if I don't want the layout to go bye-bye on browsers I can't test...

Yes, border-radius is the thing I wanted to emulate here (I can't even use the CSS 3 syntax, it's not final yet) - but more than anything, I wanted to draw something that wouldn't overlap the box's background color (I needed to make a grandparent's color visible), and that wouldn't go crazy on browsers using alternate box sizings (IE 5.x will annoy us a while still). My method allows me to create _ANY_ set of border I'd want, it resizes to quite an extend, it degrades quite well (no strange oversized padding on the borders) and it works over any background.

Of course, if I could merely set an SVG file as a box background and have it grow or shrink with the box (what fun!), set a different image for each box corner or border, there would be no need for such CSS  kung-fu...

But still, a bug's a bug. And thank you for reporting it for me.

I like your test case BTW: it is indeed clearer than mine, I hadn't considered using a border's color instead of a box's background.

@EricLaw

Sorry poor choice of words, let me rephrase, Microsoft will possibly ship IE 8 with out having a third beta, issues that people are having could be addressed at a latter date either in the form of a service pack or possibly in the next revision.

Eric, let me make this clear I have no doubt that IE team works hard, They have done some great improvements in IE 8. Yet for me the lack of a fully customizable UI is a major flaw and impediment.  I want to use Microsoft products, I have for years almost exclusively.  Like it or not since the early betas of Vista Microsoft has either given up, ignored, or just doesn't feel as though there is enough consumer demand to warrant addressing the  UI customizations.  The pre beta of Windows 7 and the beta of IE 8 exacerbates this belief because neither one is fully customizable, the address bar has become forbidden ground. Yet no one from Microsoft will say why? Or, "The reason why it's better this way is because..."  In closing, I'm just disappointed, I've seen the comments in  the betas of vista, and IE 8, for the most part they have been condescending, "it's no big deal", "only 20% of users move the tool bars around"  to me 20% when it comes to the windows user base is a significant number. I hope the IE team addresses this, because then possibly the Windows team will also. To you it may not be a major improvement, for me it will be the tearing down of the final wall that prevents me from moving on from IE 6, Firefox,and Windows XP. Again great work, but I am sorry it's just not quite there, for me....            

In the past months. I heared that final release of IE8 will b in Nov 2008. Almost half of Nov is gone.

So i dn't think any other Beta 3 version before RTW of IE8 is Their.

Waiting for next Build of IE8..

You guys are really holding back the web here. All I really want from you is SVG support - I can code around all your other bugs - I'm used to it anyway, but lack of SVG is really disappointing.

Why go to all this effort just to release  browser less capable than

Firefox, Opera or Safari? Even the friggin iPhone supports SVG now.

Can we at least get an honest answer about when you expect to support it or why you won't?

@Nikhil - They've hinted at a November release... but did __NOT__ say that it would be the final.

EricLaw[MSFT] said: "We've made tons of reliability improvements since beta-2, so rest assured that the next public build will be far more robust"

If they've made "tons of changes" since Beta2, you can bet that the next release isn't a Final version - that's way too risky.  Their next release will either be a Beta 3 (the last Beta - to let devs sync up), or they may be overly confident that all the fixes are in, and jump straight to an RC (Release Candidate) which is also fine, as long as the major outstanding bugs are fixed (cough, cough, resize events and XP theme issues).

MSFT won't go RTM until the big issues are fixed.  Beta 2 was close, but still had significant issues.  MSFT is well aware that they can't release IE8 until it is ready since this release is such a major upgrade from IE6/IE7.  If sites fail to render/work in it, in STANDARDS mode (with code to suit) MSFT will be in major damage control mode. This time around, MSFT has to make sure that "both" browsers work (~IE7 & IE8), and better yet, the compatibility mode handling, switching and configuration options need to work perfectly out of the box (they are currently about 75% complete) or this release will be a flop.

We're all excited about IE8 releasing, but both developers and MSFT are well aware that this release must be rock solid, or MSFT will have 3 unstable products in the marketplace (ie6, ie7, ie8), and will not succeed.

@Jake: Which XP theming issue are you referring to?

@eghost: Who said it was 20%?  I seriously doubt it's anywhere near that figure (although I was one of them).  As has been discussed here before: the reason that customizability was lost for IE7 was due to the heavy rewrite of the UI chrome.  

The address bar is "untouchable" in part because it's a primary security indicator, and thus it's forced on for all windows and popups.  For IE8, we did make the bar resizable within a range; trading space with the search box.

I reallly hope this works;it's 6 in the morning and to be honest I'm tireds of all this stuff, so PLEASE SIGN ME IN!!!

What is all this fuss over IE8 natively supporting SVG? How could that possibly make IE8 better?

I very rarely ever EVER see SVG's on web pages. I only ever see them on wikipedia.

And there is already an SVG plugin from Adobe for IE that I hardly use because I don't ever see any SVG's.

What is all this fuss over IE8 natively supporting SVG? How could that possibly make IE8 better?

I very rarely ever EVER see SVG's on web pages. I only ever see them on wikipedia.

And there is already an SVG plugin from Adobe for IE that I hardly use because I don't ever see any SVG's!

But hey, while were at it, why don't we complain to Mozilla to natively support every plugin it has for FireFox?

"Holding back the web". Really Geoff? That's gonna make things better?

Okay IE team, heres the secret to make IE successful, support SVG. When you do that, nobody will complain about IE anymore about standards, slowness, lack of this or that, it won't matter because it now supports SVG and that makes everything better!

Look, I'm all for suggestions to make IE a better web browser, but complaining about not natively supporting SVG, it's just nonsense. There are much more important things for the IE team to focus on then supporting SVG. I can already tell I'm gonna get a lot of hate comments on this, but thats just the way I see things.

What is all this fuss over IE8 natively supporting SVG? How could that possibly make IE8 better?

I very rarely ever EVER see SVG's on web pages. I only ever see them on wikipedia.

And there is already an SVG plugin from Adobe for IE that I hardly use because I don't ever see any SVG's!

But hey, while were at it, why don't we complain to Mozilla to natively support every plugin it has for FireFox?

"Holding back the web". Really Geoff? That's gonna make things better?

Okay IE team, heres the secret to make IE successful, support SVG. When you do that, nobody will complain about IE anymore about standards, slowness, lack of this or that, it won't matter because it now supports SVG and that makes everything better!

Look, I'm all for suggestions to make IE a better web browser, but complaining about not natively supporting SVG, it's just nonsense. There are much more important things for the IE team to focus on then supporting SVG. I can already tell I'm gonna get a lot of hate comments on this, but thats just the way I see things.

@Eric Law - "@Jake: Which XP theming issue are you referring to?"

Uhm, the hugely obvious one where any change to a select list's border or background styles cause a de-theme on windows XP.

Has been reported since day 1 of the Beta 2 release.

@static - yeah, where to start....

SVG is important and the reason why it isn't so widespread is BECAUSE IE doesn't support it natively, therefore we CAN'T use the standard.

We can tell our end users to get the adobe plugin, then when they do they'll see the "hey, this ain't supported anymore" warning, which makes them wonder if they are doing the right thing - so yes, IE IS holding back the web.

As for the more important stuff that the IE team could be doing? yes, Im sure there is, but MSFT doesn't give us any insight as to what they are working on, so we have to ask about everything.

For starters we still don't have a clue when IE8 Final will be released, or the RC(s), or Beta 3 - cause there ain't no roadmap.

No roadmap == Developers out of the loop

Developers out of the loop == Developers losing interest

Developers losing interest == Developers focusing on browsers and technologies that are open, standards, documented, with public bug tracking

I'd love to do some real cool stuff in the browser (and I do) but it sure as heck don't work in IE, cause IE doesn't support the stuff that Firefox, Safari, Chrome and Opera do - out of the box.

jake... it's only "hugely obvious" if you happen to set a border or theme on a select control, which is far from common.  

because this limitation has been there forever (IE6) with no broad outrage, it's hard to see how any serious person could consider this a "major outstanding issue".

A better question is what exactly you'd WANT to happen in this case?  Your page explicitly tries to style a SELECT element-- why would you expect XP themes to be applied?

@Joe - dude you are way off base! You ***CAN*** style the border AND background of a select list in IE5, IE6, IE7, IE8-CompatMode, but NOT in IE8 Standards Mode.

The bug is unique to IE8 Beta 2 - in Standards Mode.

Second - the issue is VERY common.  If you like pure black and white controls in your applications then hey go for it! but for the other 90% of the web developers out there we like to add a little style, even if it is only on focus so the user knows where on the form they are.

Again, I repeat, this worked just fine in EVERY other browser that runs on WindowsXP.  Only IE8 (in standards mode) can't handle the styling without messing up the native chrome.

It is what us professional developers refer to as a MAJOR REGRESSION BUG.

Just how reliable are web slices?  I just created some and set the reload timer to 15min.  I made 3-4 changes on each over a 1 hour period.  None of the slices went bold to indicate a change (nor did the titles change)

Does IE cache any of the page content? does it ignore HTTP Headers / always request a fresh copy?

I'd post a url but the site is on a local Intranet.

Well you don't see much SVG on the web yet because IE doesn't support it. Which is precisely why I think Microsoft is holding back the web.

And honestly, If you don't think vector graphics are important and will change the way web developers do things, why do you think microsoft is putting so much effort into Silverlight and XAML?

Microsoft is being completely hypocritcal in saying they are working towards supporting web standards while ignoring the bits that compete with their own technologies.

And yes, I do want Mozilla to support all *open* web standards natively - I want all the browsers to.

@EricLaw

Sorry I do not call being able to make the search bar bigger or smaller customization. Now all of a sudden it's a security issue for moving the address bar around?  Please explain to me how it's a security issue to being able to relocate  the address bar. Your biggest competitor Fire Fox has no security issue with this ability. In my opinion Microsoft is not changing their stance on this is because quite simply they don't want to. Microsoft has decided this is the way we want the UI to look and we don't care we will not change it. This is not just a flaw it is a huge corporate wall, that Microsoft refuses to tear down. I strongly remember this type of arrogance from the Beta's of Vista.  I can not believe Microsoft stance on this matter, it makes no sense, and then you wonder why people don't like Microsoft...

Recently someone asked me about the best practices for AJAX performance. Though this information is scattered