I found a web page that doesn’t work, now what?

We’ve previously blogged about Compatibility View; this post is a follow-up on compatibility and Internet Explorer 8 RC1.

As you are browsing the web with Internet Explorer 8 RC1, you might come across some web sites where everything is not displayed correctly:

snippets of broken websites including sites in multiple locales 

These issues can be fixed by clicking on the Compatibility View button on the address bar:

Picture of the end of the IE address bar, the compat view button is highlighted.  It is located to the left of the refresh button. 

This button draws the page the same way that Internet Explorer 7 would - allowing content designed for older web browsers to still work well in Internet Explorer 8.

By default, Internet Explorer 8 displays a site using its most standards compliant mechanism. In the majority of cases, this works out just fine. However, every once in a while, a page that says “display me using modern standards” really means “display me like Internet Explorer 7 used to display modern standards pages”. This is where Compatibility View comes in.

Compatibility View is domain specific – if you click the Compatibility View button on abc.example.com, all pages on example.com will be rendered in Compatibility View. Internet Explorer 8 remembers the setting – next time you come back to example.com, we will render it in Compatibility View.

There were a number of cases in Internet Explorer 8 Beta 2 where IE8 would not render a standards-based page like other browsers, and the user would use the Compatibility View button to fix the issue. We want Internet Explorer 8 to render standards-based pages like other standards-compliant browsers, so the IE team investigated thousands and thousands of IE8 rendering issues – Issues found by end-users, our test team (testing thousands of the most popular sites in the world), or cases where the ‘Report a Webpage Problem’ add-on button was used.

A large number of sites (through user agent detection) were actually relying on Internet Explorer 7 rendering behavior. In these cases, the IE team does outreach to the website, asking them to make the site work with IE8 rendering, or to use the IE=EmulateIE7 header.

In cases where Internet Explorer 8 is not rendering a page (with valid markup) like other standards-compliant browsers, the IE8 team would reduce the problem down to the minimal amount of HTML and CSS that reproduces the issue, and fix the rendering issue. While standards-based tests are a great way to ensure that IE8 renders content like other browsers, they do not cover all rendering scenarios that exist on the web – we wanted to do a real-world sanity check to find any remaining rendering issues.

To help the IE8 team investigate and fix rendering issues, please use the Report a Webpage Problem Internet Explorer 8.0 Beta Add-On to report page issues. The IE team will investigate the issue, and fix any bugs that prevent IE8 from rendering a page like other browsers.

If you like, you can also point out (in the comments) any sites where IE8 RC1 does not render a page the same way as Firefox, Safari or Opera, and we will investigate the issue.

If you are a web developer (and you have a case where IE8 is not rendering standards-based markup like other browsers), you can also point to a url with the reduced repro, or copy-and-paste the minimal set up html that will reproduce the issue.

In addition to standards-based rendering compatibility, the IE Compat team also investigated a large number of popular IE add-ons, to ensure that no bugs were introduced in IE8 that will stop popular add-ons from working. In cases where breaking changes could not be reverted (i.e. we made a change for security or stability reasons) the IE team would contact the add-on maker to update their software.

Additionally, because IE8 renders all intranet content in IE7 mode (unless specified otherwise by the user, admin or web page) the IE Compat team also spent a lot of time double-checking that IE8 (in IE7 rendering mode) renders content exactly like IE7 does. The Microsoft corporate network was a great test case here – a company the size of Microsoft has thousands of Line-of-Business applications that were designed for IE7. We ran multiple test passes on LOB apps, to find any changes in IE8 that would break these applications. Also, we’ve been reaching out to major LOB application vendors, to get them to test their applications with IE8 – so that, by the time the final IE8 version is released, IE8 is an easy drop-in replacement for IE7.

Thanks for helping make Internet Explorer 8 better!

Frank Olivier
Program Manager

Comments (47)
  1. sroussey@network54.com says:

    http://www.localsa.com.au/ie8/ and hit login.

  2. James Snape says:

    What about sites where the rendering is fine but Javascript fails? e.g. http://brightkite.com and http://getsatisfaction.com?

  3. Paul Hill says:

    And what happens if/when the site designer fixes the site to work in IE8?  Does compatibility view work forever or is there an update sent out to IE8 users saying "These sites are now fixed, turn off compatiblity view"

  4. boen_robot says:

    I’d like to see the the bugs at


    fixed, plus this regression


    None are really fixed in IE8RC1. Is it too much to ask that those get fixed in the IE8 time frame? They appear to be the only thing standing to a really complete (and bug free) CSS 2.1 support.

    @Paul Hill

    If the user chose to use compatibility view, it’s also their choise to turn it off. If your site starts to look bad in IE7 mode, then you can bet users would turn off compatibility mode. Until then, it shouldn’t matter to you anyway.

    If the user didn’t chose to use compatibility view (but instead, you were on the MS list), then you can email MS to remove you from the list at the next list update cycle. BTW, I failed to find the email (or whatever) where you should opt-out from/report compatiblity stuff.

  5. Max says:

    So if a site contains svg or canvas elements is there a magic button I can press to see them?

  6. boen_robot says:


    If it worked in IE7, it’s the same button, though it didn’t, so no, there isn’t.

    (I know you were sarcastic… I hope we can see some native SVG in IE9)

    For SVG, there IS a magic *link* that you can click though. See:


    (on Windows Vista, make sure you run it as an administrator)

  7. jasper says:

    It should be noted in the documentation or this post that:

    Compatibilty View: Changes the page rendering (including (HTML parsing & rendering), (CSS parsing & application) & (JavaScript parsing & execution) on the page.

    This means that in Compatibility View things like document.getElementById(id) and .setAttribute(name, value) are STILL horribly broken.


  8. bjp106 says:

    A browser SHOULDN’T have to have a ‘compatibility mode’.  Even though it fixes its own mistakes, the button should be taken out.  It should just work.  B/c of all the bad implementations in the past with the IE browser, you guys are definitely paying for it now.

    I’ve definitely seen some improvement on RC1.  But it still isnt there.  You guys have been working on this browser for years now, and its still not ready.  Will it ever be ready?

  9. Neal G says:

    I’ve noticed on my own site that IE8 chokes on overflow: auto; by adding scroll bars to elements that don’t actually overflow (Note that this worked in IE6, IE7, FF etc). I then added the IE7 meta tag hoping I could temporarily fix the issue by making IE8 render as IE7 only to find that your ‘render as IE7’ is also broken and introduces new bugs of its own.


    It seems to me that you are no longer releasing one new browser with its own set of bugs but you are now releasing 3 browser with their own unique set of bugs, i.e IE8.8 IE8.7 and IE8.6.


    The sad thing is there is no way for me to use conditional comments to target a subset version of IE8 since if I use [IF IE 7] it will target IE7 and IE8.7 both which don’t consistently render my web page the same way.


    Unless you guys pull off a miracle and fix your three versions of IE8, you will need to introduce some new conditional comments to target each sub version of it.

  10. hAl says:

    About breaking addons…

    I wonder if you have contacted the MS Office team for their horrible research addon that uses ton of resources even when the offcie research pane is not activly displayed and used in IE.

    That addon should not use any significant resource unless the office research pane was activated in IE.

    I surely expect an update for this terrbile addon before IE8 launch as you would expect other addon creators to improve their addons as well I presume ?

  11. nelson says:

    IE8 RC1 is NOT ready.


    IE8 really badly breaks the scope of javascript variables and is doing value assignment when NOT asked to.

    IE8 RC2 ETA please.

  12. @boen_robot

    As far as I’m concerned, bug 339309 as filed, as described and as testcase-ed has been fixed. Text underlining still has at least 2 remaining issues.

    1 remaining issue is that IE8 RC1 still fails this testcase:


    which should be in another bug report, I’d say.

    I’ve filed bug 408071


    for another remaining issue.

    Regards, Gérard

  13. boen_robot says:

    @Gérard Talbot

    Hmm… now that I reread your bug description, I think I see what you mean. Though not passing your test case is exactly what I had in mind when I mentioned your report. The underline seems to only appear when the dummy text goes over into the monospaced text. And even then, the last line is not underlined.

    I think IE8 will be ready when all currently reported bugs in Connect get fixed, but not before that. I’m really starting to think the  MSIE team should release IE8 RC2. I’d be really awful if new regressions suddenly creep up with those fixes, but keeping those bugs (some of which are regressions!!!) doesn’t exactly sound well either.

  14. Ian Boyd says:

    Did you guys un-end the Report A Broken Web-Page program?

    Last time i tried to report a problem, it told me that the program has ended, and i can uninstall the Report A Web-Page Problem addon from the Control Panel.

    So i did.

  15. EricLaw [MSFT] says:

    @hAl: Yes, we’re talking to the Office team about the Research add-on.

    @Ian: The Beta version of the add-on expired in November.  

    Please see //go.microsoft.com/fwlink/?LinkId=110518 for the latest version of the add-on.


  16. Mike says:

    There are some simple layout issues with this page, especially lower down the page. Comes up Okay in compatibility mode.

  17. aldo says:

    Sometimes I can’t view a web page at all… The page just plain crashes IE8 RC1, and it tries again, crashes and tries again, and stops.

    Its quite annoying, and the only way to fix it is if you are fast enough to click the Compat View button before it crashes again.

    I don’t quite understand why there is not an option on the error page IE shows after it has crashed 2 times to open the site again but in IE Compat View mode.

    FaceBook’s website started making IE8 RC1 crash today, but then it stops and starts again. 😐

  18. Kristofer says:

    Page rendering is the least of your worries; how about re-enabling support for "window.open" and other really basic client scripting things in IE8?

    I kept getting "no such interface" whenever window.open was called in scripts after I installed IE8RC1. Uninstalled it after an hour.

  19. greg says:

    @Kristofer – window.open still works in IE8 RC1 (I’m not sure why you were having issues) The only think I can think of is if you were trying to call this from the about:blank window (e.g. if you created another popup or iframe with about:blank as the source, then calling ***ANY*** JavaScript on this is broken (since IE7) in the "ANOS" (Almighty Name Of Security) I’ve yet to see any conclusive evidence how this makes anyone more secure but that’s apparently history now.

    PS is that the real spelling of your name? 😉 its the coolest version of "Christopher" I’ve seen.

  20. todd says:

    Doh! its next week – my bad.

    The link to the IE Chats should be available on the main front page/every page of this blog.


  21. Mike says:

    I think this posted on ajaxian this morning really shows how much microsoft is holding back the web.


  22. Dan says:

    Mike– Cool page, but it really shows how EVERY browser is "holding back the web".  Or that maybe, just maybe, the web isn’t really being held back at all.

    Kristofer– Make yourself useful: post a repro URL.

    Aldo– Uninstall your broken add-on.

  23. Chas says:


    is not the same

    "# re: I found a web page that doesn’t work, now what?

    Wednesday, February 11, 2009 8:16 AM by boen_robot


    If it worked in IE7, it’s the same button, though it didn’t, so no, there isn’t.

    (I know you were sarcastic… I hope we can see some native SVG in IE9)

    For SVG, there IS a magic *link* that you can click though. See:


    (on Windows Vista, make sure you run it as an administrator)"

    That play is not supported anymore

    Will the IE team make a post for developers to start writing in web standards code?

  24. nate says:

    Off topic, but i can’t find anywhere to report ie8 bugs.  Why is IE8 so abysmally slow when viewing very large forms (hundreds of elements)?  it’s unusable to the point that entering a single value into a form can cause the app to lag for several seconds.  scrolling?  you might as well forget it.

    sorry this is out of place.

  25. hAl says:


    You do understand that it is actually forms with hundreds of elements that are unusable for internetusers ?

  26. Tino Zijdel says:

    I would like to repeat my request for a clear indication in IE8 which renderingmode is used for a given page (and possible also why in case of IE7 rendermode). Now when doing tests I find myself keeping on hitting the compatibility button until I get the balloon saying the rendering is IE7 mode to know for sure which mode is being used…


    nelson: https://connect.microsoft.com/IE/feedback/ViewFeedback.aspx?FeedbackID=413508

    That one really shocked me; imo that’s really a blocking issue :X

  27. EricLaw [MSFT] says:

    @Tino: The compatibility view button looks different when it’s enabled: it shows in a pushed state.  Alternatively, you could use the ShowDocMode bookmarklet; please see http://www.enhanceie.com/ie/ie8.asp#ShowDocMode

    The Javascript bug you’re referring to does not repro in current builds.  

  28. vasko_dinkov says:

    OK, here’s another bug I stumbled upon that I consider pretty crucial – I still haven’t found a way to remove completely the "clip" CSS property once that it is applied to an element. IE8RC1 seems to remove it for the original element but any child nodes in it are still clipped to the element’s box. Example:

    <div id="test" style="position:absolute;top:10px;left:10px;background:red;width:100px;height:70px;clip:rect(0 100px 50px 0);">

     <div style="width:50px;height:100px;background:green;"></div>


    Now I have tried the following:

    document.getElementById(‘test’).style.clip = ‘rect(auto auto auto auto)’; // that’s what works in IE < 8.0RC1

    document.getElementById(‘test’).style.clip = ‘rect(auto,auto,auto,auto)’;

    document.getElementById(‘test’).style.clip = ‘auto’;

    document.getElementById(‘test’).style.clip = ‘none’;

    document.getElementById(‘test’).style.clip = ”;

    IE8RC1 also doesn’t support removeProperty() so it can’t be used either.

    Let me know if you have an idea here. I believe this issue may cause trouble to a lot of people.

  29. Brian LePore says:


    I had the window.open bug when I was using the beta versions, but by it was fixed for me when I upgraded to the release candidate.

    The RC actually fixed a lot of things for me, but the debugger seems to have lost some functionality. The ‘click this element to select’ feature no longer works.


    Actually there would be no IE8.6 because IE8 does not contain an IE6 rendering mode. Would be nice if it did.

  30. Catto says:

    Hey Now,

    Using the Compatibility View is good.

    Thx 4 the info,


  31. It is important to note that compatibility mode can also be triggered programmatically and I believe this is what every developer should do.  The reason is simple, you want to give your customer the best possible experience on your site.  In fact, when the compatibility meta tag is explicitly declared in your code, the compatibility button disappears.  You don’t need to give your customer unnecessary choices.

    There are several scenarios:

    – If you have already tested your web site in IE7 and you are not yet prepared to make the transition you can just add the compatibility meta tag to all  your pages and they will automatically render in IE7 mode.

    – If you are gradually transitioning to more standard code, you can selectively add the compatibility meta tag to the pages that you have not fixed and maybe never will (outdated content, support pages, etc…).

    – If your code works well with IE8 then you should add the IE8 met tag!  This will ensure that your code will continue to render correctly even after the next browser transition.

    In summary, the compatibility meta tag is the perfect way to ensure that your customers will enjoy your pages the way they were intended.

    At ArtinSoft we have created the IE8 Compatibility Wizard that allows you to automate the meta tagging process.  You can check it out at: http://www.aggiorno.com/aggiornoexpress.aspx

  32. @ vasko_dinkov

    Yes, I think you have indeed stumbled on a bug affecting clip.


    … but I don’t believe this issue may cause trouble to a lot of people.

    Regards, Gérard

  33. Tino Zijdel says:

    EricLaw: "The compatibility view button looks different when it’s enabled: it shows in a pushed state."

    Well, then the pushed state is almost not distinguisable from the normal state since I never noticed any differences (which makes this more of a general GUI design issue). The ShowDocMode bookmarklet is nice but still doesn’t show why a page may be rendered in IE7 emulation mode.

    Good to hear that the JS scope bug is fixed, I take it this still has to be confirmed on the Feedback report?

  34. EricLaw [MSFT] says:

    @Tino: If a page opts-in to a particular rendering mode (or if it’s on the MS-supplied Compatibility list) the Compatibility View button is not shown.  It should be pretty easy to tweak the ShowDocMode bookmarklet to see if the page contains the META tag.

  35. Alan Gresley says:

    Shouldn’t that be thank you to Internet Explorer for destroying the Internet / Web Standards.

    First it was the DOCTYPE since Microsoft couldn’t develop a web compliant browser.

    Second it was IE6.

    Now sites get blacked listed to render in IE7 mode forever without some META switch.

    I guess for Microsoft it was just Line-of-Business, Line-of-Business, Line-of-Business, Line-of-Business.

    Well these Lines of Business are welcome to IE6 in QUIRKS forever.

    Looking forward to the EU decision!

  36. Pete says:

    Alan, I’m hoping that maybe you’re simply too young to understand what you’re talking about.  Firstly, the fact that Microsoft and Netscape both focused on "web compliant" browsers is how the current mess came about, since both wanted to be compliant with the web of the era.  

    Only later did the notion of "standards-first" development arise. Of course, "standards-compliant" and "web-compliant" meant opposite things until only very recently, and the IE team now has to find some way to turn the ship when half of the web wants nothing to change.

    You could make the same complaints against Netscape’s lack of standards except they failed to build a compelling enough product to ensure their long-term survival.

    You’re also jumping to absurd conclusions.  What do you suppose the odds are that ANY site will be on the Compatibility View list that does not WANT to be there?  (Hint: 0)

  37. @ Pete

    > Alan, I’m hoping that maybe you’re simply too young to understand what you’re talking about.

    Alan Gresley understands well what he’s talking about. In a serious debate, personal remarks and sarcastic comments never promote a person’s point of view as sound, serious or as acceptable.

    > Microsoft and Netscape both focused on "web compliant" browsers is how the current mess came about,

    There is no such thing as "web compliant": it was at first a competition of proprietary DOM (document.all vs document.layers) and proprietary extensions of W3C specifications, where each opponent tried to impose its technology on the market… just like VHS vs betamax and competition of new technology (there are lots of recent examples available here: blue-ray, DVD, etc)

    > Only later did the notion of "standards-first" development arise. Of course, "standards-compliant" and "web-compliant" meant opposite things until only very recently,

    As soon as march 1998, AOL-Netscape stated officially and formally that they would adopt open (W3C web) standards and their 2000 Gecko engine proved they were serious, commited, dedicated. In late october 2001, it was already obvious that NS 6.2 was more, much more W3C-web-standards compliant than IE 6: the 3 problems with NS 6.2 was application stability, rendering/layout speed and a declining market presence… but it was not W3C web standards adoption and compliance.

    Netscape soon (1998) embrased W3C web standards; Microsoft, in the mean time, had been double-talking during years and years (from 1999 to 2004) about W3C web standards adoption and there’s plenty of proof available which can document this overwhelmingly.

    Even Chris Wilson (May 11th 2006) in his personal blog clearly suggested he would resign, quit, leave, "bye-bye" if Microsoft did not resume for real the widely accepted, established W3C web standards (HTML 4, CSS) in IE.

    > IE team now has to find some way to turn the ship when half of the web wants nothing to change.

    Let’s give an example of what should have been done by Microsoft leaders and/or the IE team over 1 decade (10 years!) ago. Let’s start with a juicy quote from Eric Meyer:


    (…) when faced with bugs in their implementations: fix it or preserve it? The classic example of this was the original implementation of height and width in Internet Explorer, which was wrong per the CSS specifications. The IE team at the time became aware of this fairly soon after they shipped it in IE3 … and yet the problem wasn’t fixed until IE6, a delay that slowed the adoption of CSS and gave rise to a whole family of JavaScript sniffers and CSS hacks.



    Let me underline this: instead of truly fixing that CSS 1 box model bug in IE 4, they just let that major incorrect implementation linger on and on, rot during years and then they created in 2001

    a) a new rendering mode and

    b) a new rendering engine and

    c) a new way to trigger that rendering mode in IE 6… instead of fixing the bug as soon as a next IE release to be done.

    Same thing with hasLayout: instead of addressing the CSS bugs and incorrect CSS implementations (in particular float, clearance, abs. and rel. positioning, overflow, etc), they introduced an additional layer of complexity wrapping around all of these faulty implementations. This is exactly how the original author, creator of hasLayout introduced such feature in an article at MSDN.


    August 31, 2005. What is "HasLayout" and why is it important? There are several bugs in Internet Explorer that can be worked around by forcing "a layout" (an IE internal data structure) on an element…



    AOL-Netscape never created as much damage with their browser as IE 5+ has.

    > (…) Netscape’s lack of standards except they failed to build a compelling enough product to ensure their long-term survival.

    AOL won against Microsoft for anti-competition practices and settled for an out-of-court agreement $750 millions compensation. Being able to embed, to bundle a web browser into an operating system is a huge, considerable advantage. Even the DoJ agrees with such understanding.

    Netscape 6.2 was better, more compliant than IE 6; Netscape 7.2 (august 2004) was better, more compliant than IE 7 (release in october 2006).

    > What do you suppose the odds are that ANY site will be on the Compatibility View list that does not WANT to be there?  (Hint: 0)

    Much more than 0. If a web author/webmaster fixes his website to be W3C-web-standards compliant, then he/she still will have to notify Microsoft to be removed from that compatibility list.

    Oh.. and by the way, *.microsoft.com is on that compatibility list. Isn’t that strange? Upgrade your IE browser but continue to view sites like your still using our IE 7 browser.

    Gérard Talbot

    "The browser wars are over (…)"



    Web Standards Project Applauds Netscape 6.0 as Milestone Toward a Standards-Based Web.

    Publication: Business Wire

    Date: Tuesday, December 5 2000

    Netscape Standards Challenge


  38. sroussey@network54.com says:

    BTW: the webpage that crashes IE8RC1 on demand has a bug report on connect:



  39. Chris says:

    of course there is no ‘web compliant’… no one should make webbpages, only standards!!  why would anyone want to do anything with the internet other than run standards validators?  everyone else can just watch tv!

    since it is much easier to write new standards than impliment them, we can write new standards fast enuff, no one will notice how bad they are because we’ll tell them ”no dummy, that’s version X and we’re already working on ver Y of the standard which is incompatible but less broken.”  

    down with the web… up with standards!!

  40. Ben Amada says:

    I have a page that doesn’t display right in IE8 (even in compatibility view).


    In IE8, the gold background doesn’t stretch from the top of the page to the bottom.  It works right in IE7, Firefox, Chrome, Safari, etc.  This page passes the W3 validator.

  41. Ben Amada says:

    It’s a little strange, that page I posted a link to sometimes works when the page first comes up, but if you hit Refresh or click somewhere on the page, it goes haywire again.  It’s a very simple HTML page.  I’m viewing it in the IE8 VPC download.


  42. PatriotB says:


    Funny how the "When can I use" page (http://a.deveria.com/caniuse/) conveniently fails to list the important features that IE has led the way in that most other browsers still don’t support after nearly a decade in IE — namely, writing-mode support for vertical text, and ruby.  These are in CSS3 modules as well.

    Yet the page lists things like CSS Transitions and Animations which were developed as part of one implementation (WebKit), and even 3D Transforms which is only in Mobile Safari!  These aren’t even available in draft form on W3.org yet, though the WebKit team wants the CSS working group to adopt them directly from them.

    If we’re going to include lists of things developed as part of a single implementation that were submitted to W3C, then how about HTML+TIME, VML, HTCs?  Why do the "little guys"’ proprietary ideas get listed but not Microsoft’s?

    @Gerard — Markus Mielke is surely not the "creator of hasLayout."  He just wrote the article.

    From what I gather about IE’s history, the concept of "layouts" existed long before CSS — thats why things like buttons, images, textareas, iframes, etc are in the list of elements that always have layout regardless of any specified CSS properties.

    My understanding of what "hasLayout" was, was a way to give normal "simple" flow elements like divs advanced capabilities, like sizing and positioning, that were already available to the aforementioned elements.  In other words, they had an HTML-based engine, and then CSS came along with a lot of requirements that were at odds with their existing engine.

  43. @ vasko_dinkov

    I filed bug 415011



    Regards, Gérard

  44. vasko_dinkov says:

    @ Gérard Talbot

    Thanks! BTW, you’ve compiled a great list and instead of sending cakes to other destinations


    they’d better pack one for you (or at least some bear). =)

  45. archie lukas says:

    I assume that Firefox will work regardless though/

    Good enough for me then

  46. Arne Garder says:

    the webpage that crashes IE8RC1

Comments are closed.

Skip to main content