Overview of Platform Improvements in IE8 RC1


This is one of my favorite times in the product cycle. IE8 is platform complete and as we get closer to releasing the final product, more and more web developers and designers will take advantage of the browser’s features to enable scenarios we haven’t even imagined!

Since the release of IE8 Beta 2 we’ve listened to feedback from many channels including IE8 Beta Feedback, standards working groups and this blog. We’ve made thousands of platform improvements in response to both feedback, and from running the test cases that Jason Upton blogged about on Tuesday. The platform is ready to be built on. I want to give you an overview of web platform improvements since Beta 2, some of which we’ve already talked about and some of which will be covered in more detail in the coming weeks.

Compatibility

It takes time for web developers, designers, and IT professionals to migrate a site to an updated browser. Yet our mutual customers expect the web to look and feel the same after they install the latest IE version. We’ve built Compat View and IE7 Standards Mode to ease migration and allow web developers and designers to opt-in to IE7’s behavior while they upgrade their site. Since Beta 2 we’ve improved IE7 Standards Mode fit-and-finish such that RC1 is very close to the real IE7 web platform.

Interoperability and Standards

  • CSS 2.1. Layout interoperability with other browsers through the CSS2.1 standard has always been a top goal for IE8. Beta 2 supported all properties in the CSS 2.1 specification and passed over 3,200 test cases. We’ve made significant improvements since Beta 2 and this week’s RC1 passes well over twice as many test cases as Beta 2. For example, one of our favorite new features is IE8’s new support for High-Res layout, which we’ll blog about in more detail later. We expect very few changes between this RC and the full CSS2.1 support in the final product, which web developers and designers can use to write their pages once and have them rendered the same across browsers.
  • HTML, Document Object Model (DOM) and JavaScript. Throughout Beta 1 and Beta 2 we’ve talked about how IE8 is much more interoperable with other browsers in core areas including attribute handling and element lookups like those through getElementById(). To help ensure future interoperability with other browsers and standards compliance, the Release Candidate includes the following updates, which we recently blogged about:
    • Mutable DOM Prototype includes the new ECMAScript 3.1 conformant getter/setter syntax.
    • ARIA supports the dash syntax “aria-checked” across all IE8 document modes. This means web developers can write code once that works across IE8 modes and with other browsers.
    • Cross-Domain Requests (XDR) now checks Access-Control-Allow-Origin header for a match to the origin URL as well as wildcards. As a result, data is only shared with sites whose origin the server specifies.
  • Performance. Similar to interoperability, better performance helps improve developer productivity. To this end, we investigated core performance scenarios and focused on optimizing common AJAX design patterns. Web developers and end users alike will experience performance improvements since Beta2.

Development environment

  • Developer tools. Beta 2 introduced more power with the JavaScript profiler, save to file, and console.log support. RC1 has dramatically improved stability and a much more accurate view of the HTML tree and CSS tracing. It also offers more flexibility by adding a menu option for viewing source with Notepad, the built-in viewer, or any other choice of viewer.
  • Documentation. We think good documentation and communication is an important part of the development environment. To this end, we have updated our Internet Explorer Readiness Toolkit and MSDN IE Development Center for web developers and designers to use as references. Stay tuned to the blog for much more detail on improvements we’ve made since Beta 2 and tips for upgrading to IE8 Standards Mode.

We take your feedback on critical issues very seriously and we know that behavior changes to the platform have the potential for broad impact. We intend to make few platform changes between now and the final product. We’ll be very deliberate about what changes we make and diligent about documenting and communicating them.

Please download the RC1 and test with confidence if you haven’t already. We’re very excited about the improvements we’ve made to IE8’s web platform and developer tools and even more excited to see how web developers and designers build on them to enable new, incredible scenarios! If you’ve already enabled a new scenario using IE8 features, leave a comment and let us know about it.

Marc Silbey
Program Manager

Update 10:15 pm: fixing getElementById() reference.  Thanks Gérard!

Comments (135)

  1. Roman says:

    Now speaking of interoperability… RC 1 appears to score one less than Beta 2 on Acid3 (20 versus 21). What gives?

  2. Roman says:

    And by the way. I tried to post this in reply to "Upgrading to RC1" but then all comments were rejected and I don’t know if anybody looks there now.

    It appears that IE installer is incompatible with Adobe Reader 9; see the discussion at http://www.microsoft.com/communities/newsgroups/list/en-us/default.aspx?mid=5d8d0b83-d54b-4f9c-945e-66d83df66ffa&dg=microsoft.public.internetexplorer.beta.

    I can confirm this issue still persists for RC1. I think it’s probably an Adobe bug, but I’d expect a mention of it in the release notes.

  3. Roman says:

    And by the way. I tried to post this in reply to "Upgrading to RC1" but apparently there was a temporal issue with posting then, and I’m not sure if anybody looks there anymore.

    Apparently IE installer is incompatible with Adobe Reader 9. See the discussion at http://www.microsoft.com/communities/newsgroups/list/en-us/default.aspx?mid=5d8d0b83-d54b-4f9c-945e-66d83df66ffa&dg=microsoft.public.internetexplorer.beta.

    I can confirm this issue applies to RC1 as well. It’s probably an Adobe bug, but I’d expect a mention in the release notes.

  4. gabe says:

    is there any news on whats after ie8

    is microsoft working on ie9 or whatever post ie8 is called

    is there any idea on what it will include will

    it include any css3 or xhtml or anything

  5. gabe says:

    also from what ive heard microsoft has another layout engine

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

    why doesnt microsoft make ie9 use the expression web engine or a combination of it and trident

  6. @ Marc Silbey

    > element lookups like those through getElementByID().

    It is not getElementByID()

    It is getElementById()

    "Lower case ‘d’!!"

    http://www.quirksmode.org/dom/w3c_core.html#t10

    Regards, Gérard

  7. Mitch 74 says:

    Margins are not collapsed correctly in some cases:

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

    Test case 3

  8. orbanbalage says:

    Hi, it is definitely better and better.

    Ctrl+T however still opens a tab very slowly, which is very disturbing – I’m used to the instant Ctrl+T # Ctrl+V combo, but now I have to open the new tab, wait two secdonds, and paste whatever I want to open/search for.

    Thanks

  9. Poppyto says:

    Too slow too slow ! I don’t understand why IE became so slow…Chrome & Safari Rocks…

    You should use V8 for the javascript engine.

    I think a lot of users will use IE if it were faster than Firefox….you should considere this (2 sec for tab opening…you kidding?)

    I don’t use IE anymore since Chrome.

  10. I really hope that performance of IE8 will be much better since the only authorized browser at my work is IE.

  11. I installed the RC1 1 today, and now my entire GUI is "delayed" or slowed down.  If I drag any window in Windows, I see trails of slow GUI updates (redraws, whatever you want to call it) and explorer.exe sky-rockets in CPU cycles until all the GUI updates are done.

    All worked fine last night; RC1 installation, plus disabling the One Note, Research, Discussion (whatever that is), Send To Bluetooth (I never use it), and Windows Messenger (no publisher on that, surprisingly).

    I’ve been VERY excited about the advances in ie8 – it really does take Web Browsing to the next level and makes it more productive.  However, having to uninstall ie8 today was disappointing and I have high hopes Microsoft can get this cleaned up soon.

  12. RalphFM says:

    Installed IE8 RC1 and subsequent Live goodies…  all working well.  On our systems, it is faster than FireFox 3.1 b2.

    I have one request, we need a password revocery/editor for IE. Since its wise to change page passwords, a method of recovering/editing/exporting the passwards saved would be extremely useful and wonderful.

  13. RalphFM says:

    Installed IE8 RC1 and subsequent Live goodies…  all working well.  On our systems, it is faster than FireFox 3.1 b2.

    I have one request, we need a password recovery/editor for IE. Since its wise to change page passwords, a method of recovering/editing/exporting the passwards saved would be extremely useful and wonderful.

  14. Update: Uninstalled ie8 RC1 and now back at ie7 – all GUI delays gone.

    Fairwell, for now ie8; I REALLY look forward to your return as I’m already missing the address bar enhancements!

  15. Jace says:

    For those of you seeing slow tab open times, click Tools, Manage Add-ons.

    If you don’t use an Add-on, disable it.

    The far right column of the Add-on manager will show load times.

    The worst offenders I’ve seen so far are the following:

    Research – unknown, but slows new tabs down for sure

    Java – about 1 second

    Norton NCO BHO – .77 seconds

    Norton Intrusion Prevention – .38 seconds

  16. Jace says:

    Also,

    IE8 RC1 completely breaks the following website:

    http://www.byu.tv/

  17. Arjon says:

    Nice that there is a possibility to DISABLE add-ons. But how can I DELETE them..

    Maybe something for the next version IE8.5. It would be great anyhow to release a IE8.5, only to update functionality, not related to the trident engine.. like UI changes and things like removing add-ons..

  18. Jason says:

    Why does IE not respect the "Open links from other programs:  open new tab in current window" when clicking the email button in Windows Live Messenger?

  19. Jeff says:

    I have to access some sites that do not support anything beyond IE6.0  For those I use the User Agent String Utility v2.0.  Unfortunatly, that utility does not support IE8.  Will there be new version of the User Agent String Utility?

  20. EricLaw [MSFT] says:

    @Jason: There are some cases (specifically, cases where IE is instantiated via COM) where opening new windows in tabs is not possible.

    @Jeff: The UAPick plugin here (http://www.enhanceie.com/ietoys/uapick.asp) can be used in IE8 RC1 to emulate any other browser or version.  I’m curious: are these sites on the public internet?

  21. Arieta says:

    For everybody who is having a slow new tab opening problem, disable the Java SSV Helper under add-ons. The Java console can stay, but the Java SSV helper is known to make tab opening extremely laggy.

    If you don’t have Java installed, try running IE in no-addons mode, and see if it works faster – if it does, you’ll just have to find out which addon is causing the problem.

  22. EricLaw [MSFT] says:

    @Mitch 74: At 100% zoom, the "margin collapse" test passes in the current internal builds.

    Are you using a non-100% zoom?

  23. Arieta says:

    EricLaw: On the public IE8 RC1 build, test case 3 on that page will not pass on the first opening. You have to resize the browser window for it to pass. Tested in fullscreen mode, no zooming.

    Are you saying that this been fixed in later internal builds?

  24. Jim says:

    I would definitly focus on tab opening speed. Beta 2 sometimes took 10+ seconds to open a new tab. RC 1 seems to take up to 5 seconds which is better but still irritates me to wait just to open a blank tab (and considering other browsers do it in 1 second or less).

  25. Daniel says:

    @EricLaw [MSFT]:

    I looked into Mitch’s testcase and filed: https://connect.microsoft.com/IE/feedback/ViewFeedback.aspx?FeedbackID=408828

    After unhovering the green box, the element with negative margin is incorrectly repositioned (in this testcase it’s generated content, but any element would expose the bug as well).

  26. EricLaw [MSFT] says:

    @Jim: You almost certainly have one or more buggy addons installed.  How long do new tabs take if you start IE without addons?  www.enhanceie.com/ie/troubleshoot.asp#crash

    On my computers, new tabs open in well under 1 second.  

    @Arieta: yes, the case passes on current internal builds at 100% zoom.

  27. Teuvo says:

    Webslices stopped working when I upgraded beta2 to rc1. Connection problem: Internet Explorer cannot display the webpage.

    Not a firewall problem, cause I tried that already.

    Any ideas?

  28. Esben Sundgaard says:

    Hi

    I think IE8 is really great and I don’t have any problems with it. It is definitely the best IE ever! Off course it would be nice if you would add CSS3 in it but still it is much better than IE7.

    I know that you don’t plan a RC1 for Windows7 but I really hope that you would reconsider that decision because RC1 is MUCH better than Beta2 or whatever is included in Windows7 :-)

  29. Esben Sundgaard says:

    Another thing: The WYSIWYG Html editor in IE (also in IE8) is producing very bad HTML. Wouldn’t it be a good idea to borrow some from one of your other tools like Visual Studio or Microsoft Expression Web which makes really nice HTML? :-)

  30. Brian LePore says:

    Not that I’m complaining, but why are all browsers adding support for ARIA at this time? I’m aware of what it is and how helpful it is to those that need accessibility tools, but is the assistive technologies lobby (best term I can think for them) stronger with browsers than say, those pushing for SVG? Is it very little work on your part to add support for something like this (relatively speaking), compared to adding W3C DOM event handling?

    Again, I’m not digging Microsoft on their choice here because it is an admirable one, it just seems highly an odd one to me.

    Shameless request for IE9: W3C Range/Selection support! The differences between the W3C Range and the MS TextRange objects are much more difficult to work around compared to the difference in the event models (which are well documented and solutions for most cases already found).

  31. Rocky Moore says:

    I have a site:

    http://www.SilverlightCity.com

    That comes up with compatiblities issues and is forced to IE7.  The site validates on HTML, CSS and the Feed, so I have no clue what IE8 is having a problem with.

    It there any tool coming or perhaps a validation script that can give us developers feedback as to "what" is incompatible with IE8 mode?

    Rocky

  32. Sondre says:

    Thanks for a nice upgrade to the browser. I have one feature request for you: Allow the user to decide which side the scrollbar should appear on. With more and more touch and tablet devices and the growing number of left-handed individuals, it only makes sense to support those of us who prefer using the left hand. This is a major annoyance when surfing the web with a slate tablet PC.

  33. Jamie Lin says:

    Thanks for working so hard on making a good browser.

    Can I recommend increasing the timeout before IE complains abouot slow javascript?  Right now when we try to have it run JS that is a bit more complicated, say drag-and-drop photo management interface on Sosauce, it will start complaining very quickly.  We’re constantly testing with all browsers and the other 3 don’t have such behavior.  Whereas in IE series, the timeout is the same from IE6 to IE8.  We believe people’s client computers are much better these days in handling complex JSs and therefore should be allowed to run more complicated JS for longer before throwing a warning at the user.

    If you need a sample page and how IE6-8 is seriously over-cautious on slow script than other 3 competition, please feel free to drop me a comment on Sosauce.  Thanks.

  34. stability says:

    IE8 RC1 crashes on these simple little lines:

    http://thinkweb2.com/projects/prototype/misbehaving-ie8/

    and IE8 RC1 is not ready for primetime yet according to Ajaxian (and anyone I’ve talked to):

    http://ajaxian.com/archives/push-back-digital-tv-or-ie-8

    so the question is, when will RC2 be out?

  35. notfor therest ofus says:

    Re: "On my computers, new tabs open in well under 1 second. "

    Well you must be pretty special! I don’t have _any_ of my IE8 RC1’s opening a new tab in less than a second.

    I’ve disabled (and or removed) every addon except Adobe PDF, but I have yet to see a new tab open in a respectable time frame.

    According to this blog (the Internet at large) and IE Feedback on Connect – I am not alone – In fact every user *except* EricLaw[MSFT] experiences painfully slow tab loading in IE regardless of how many addons (buggy or not) are loaded.

    From the Connect bug report (and chats with the IE team) it has been discussed, and disclosed that there is a MAJOR architectural flaw with the tab implementation/addon implementation that is guilty of the slow new tab rendering.

    Please stop blaming buggy IE addons (unless you are going to point DIRECT EXPLICIT fingers at the guilty ones).

    Running a browser without addons is *NOT* an acceptable workaround.

    http://www.digg.com/about:Tabs”>http://www.digg.com/about:Tabs is not a valid url

    nor is about:Tabshttp://www.digg.com/

    but that is what we all get when trying to paste a url in the location bar after opening a new tab.

  36. JP says:

    The slow tab opening (whether you claim it is a bug or by design) is unacceptable. The problem is not with add-ons either.  In every other browser, opening a tab is immediate.  In IE8 there is *always* a lag.  This better be fixed by the final release, otherwise you will continue losing market share because it makes IE8 feel extremely slow and clunky.  That is the reality of the situation.  Opening a new tab should be immediate.  Any lag at all is unacceptable.  

  37. Pete says:

    Please consider adding extensive tests for your VML support. It’s a shame (an embarrassing) to see that VML is still broken in RC1, and only works for the very simplest of cases with VML. Especially working with VML programmatically is buggy to say the least.

    Also annoying is that if you enter a URL in the address field right when IE has started up (but still not fully loaded because of the long startup time), IE will tell you that it can’t complete the request.

  38. Dan says:

    JP: my tabs open almost faster than I can click the stopwatch (200 milliseconds?).  How long do yours take with no plugins?

    As for specific plugins that are slow the following three [Java SSV, Windows Live Toolbar, Office Research] have all been mentioned as culprits.  IE8 has a cool "Load time" column in the Addon manager that tells you how long each takes.

  39. Arieta says:

    >In fact every user *except* EricLaw[MSFT] experiences painfully slow tab loading in IE

    New tabs open instantly on my end.

    Like I said, the only thing affecting them is the Java SSV Helper, which I’ve disabled.

    Have you just disabled all plugins, or did you run IE in it’s no-addons mode? It’s not the same thing.

  40. Orl says:

    I took the advice Arieta gave about disabling that Java SSV Helper add on, and sure enough IE loads much faster and tabs open much quicker, less than a second.

    I just only wish I had known about that sooner.

  41. WindowsFanboy says:

    Please make it so you can slipstream the final version of IE 8 with nlite.

    http://www.msfn.org/board/index.php?s=b4c79a08571a25e7c89f9765b7c92b5c&showtopic=127583

  42. Grd says:

    Please make the browser opening time faster and the beta version on windows 7 causes a flicker on the taskbar when chaning windows.

    While competing browsers can work fast why not IE.

  43. gabe says:

    @ windowsfanboy

    microsoft does not support nlite

  44. ajaxwiz says:

    Installed IE8 on Vista and it required a restart to use. I don’t understand why this is necessary. It’s just a browser. No reason for it to be so integrated into the OS that it requires a restart. Makes no sense to me!

    There’s STILL no console object with a logging function. When scripts are loaded and executed on-demand and there’s an error, the line # for the error is very high, fluctuates around 80 million or so. That’s definitely a bug. And there’s still no reference to the JavaScript file that is producing the error. I do a lot of JavaScript development and IE is still not even close to being a development platform.

    And while I’m ranting, why on earth did the IE team not use an already proven safe, fast, and standards-driven web engine like Gecko or Webkit? Why do you guys keep insisting on re-inventing the wheel? (although your wheel isn’t quite round and breaks often) Why are you guys not focusing on JavaScript speeds and development? Do you guys not realize the future of the web? I mean, come on, you guys are Microsoft! Where’s the much requested border-radius implementation? I guess you guys still haven’t learned from the past. Maybe you’ll start to understand when IE’s browser share drops below 50%. I’m sure the programmers are trying their best but they have to answer to someone who just doesn’t get it. Oh well. </rant>

  45. Glen Fingerholz says:

    "There’s STILL no console object with a logging function"

    Yes there is. http://blogs.msdn.com/jscript/archive/2008/09/17/jscript-debugger-in-internet-explorer-8-beta-2.aspx

    I’m not at a machine where I have IE 8 RC1, but I noticed that /w IE 8 BETA 2, in the developer toolbar at least, console.info(window.external) prints garbage, like it’s dereferencing an invalid pointer.

  46. sokzzuka says:

    element.addEventListener() stil not implemented :/

  47. Mike says:

    Since the release of IE8 Beta 2 we’ve listened to feedback from many channels

    Apparently heard nothing about users wanting svg, canvas, css opacity, MathML, proper event handling.

    Need I go on, IE8 is a step in the right direction and people like me would probably stop moaning if the team could announce that they will continue to work at a similar pace on IE9 start adding requested features.

    This browser is not going to stop the continued decline of IE market share because it is still so far behind the rest.

  48. Arieta says:

    Orl: it has been mentioned on this blog for many months, I think I first heard about it back in october.

  49. Arieta says:

    Orl: it has been mentioned on this blog for many months, I think I first heard about it back in october.

  50. hAl says:

    Those slow (I would even call it buggy) addons is a big nuisance.

    Especially the MS Office research addon. This is an almost unused plugin that should not interfere with new tab behaviour unless research plugin was activly used by browsersusers (which is extremely rare.

    You should seriously be spanking your Office team for introducing such a horrible addon. I suggest to have it updated by the Office team before IE8 release.

    Add that Java helper is terrible too of course

  51. davidyannick says:

    when a 64 bits version will be on line for french version ?

    thx

  52. tyler says:

    IE8 RC1 performance / application locking.

    When IE8 RC1 is executing a long running Javascript loop there is a complete lock up of IE8.  You can’t view the IE console/dev tool window if it is under the current IE window (which is always if it is detached for readability).

    In addition the statusbar says: "waiting for about:blank" which is obviously incorrect.

    Performance goes in the toilet with logging to the console in IE8 RC1… setting innerHTML in the current web page is much, much faster!

  53. tyler says:

    oh one last thing… scrolling performance in the console log is dreadful.

    I’m not sure how it was developed but it needs to have some better off screen buffering/loading of content.

  54. Typhoon87 says:

    I have noteced that with the latest version of the Skype beta having the IE add on turned add makes tabs open slowly. If I diasable the skype ie add on tabs open extremly fast.

  55. Klimax says:

    I have one simple wish.IE8 RC1 for Windows 7.(included beta has a lot of regressions and I cannot install obviously RC1 for older systems)

  56. ajaxwiz says:

    @Glen Fingerholz:

    Invoking console.log() generates a JS error in IE8 RC1. I read that link, but it’s simply not true, unless of course I need to enable the console object. But, I’ve found no reference to that.

    Also, that link refers to JScript, which is NOT JavaScript, but rather Microsoft’s implementation of JavaScript. I can only hope that’s just an old reference and IE8 is not actually using JScript.

    IE8 is ramping up to be an EPIC FAIL.

  57. ajaxwiz says:

    Apparently the console object only works after you open up the development tools and click on the Script tab, which doesn’t make sense to me. What if a developer left some console commands in the code for the live site on accident? It’ll generate an error, when there really isn’t one.

  58. Glen "Double Dutch" Fingerholz says:

    @ajaxwiz

    You are correct in that it does not work like the Firebug one. For some reason you have to open the developer toolbar or you get errors. It probably doesn’t create the console object until the dev. toolbar is loaded.

    As for JScript, I can assure you that IE 8 uses JScript. They have not changed their JavaScript engine out /w JScript.NET or anything like that.

    IE 8 is a mixed bag for developers. We have better tools, and a better rendering engine /w IE 8 (and Fiddler was great /w IE 6 and IE 7 – much more powerful than what Firebug comes with), new DOM APIs (and corrections for old ones). We still don’t have XHTML, the W3C JavaScript event model, a more compliant JavaScript engine (and "for each" would really be nice), SVG and/or Canvas. That said, fairly compliant rendering should be a huge time-saver.

  59. gerald says:

    console.info(), console.log(), console.error(), and console.warn() all throw errors if the IE Development Tool Window isn’t open yet.

    speaking of logging, is there a way to have the logging appear on the left rather than that small window on the right?  The HTML tree is fine and all, but its not even close to Firebug in terms of usefulness.

    Its a minor thing, but why can’t IE look sexy? the browser, the dev tools, they all look like my kid brothers VB6 applications – lame and boring.

  60. ajaxwiz says:

    @Glen Fingerholz

    I agree. The debugger is a step in the right direction, but there is still a HUGE gap in standards/speed/features between IE8 and Gecko/Webkit. It’s simply too little, too late.

    Microsoft is not catering to the developers, but rather the average internet user who can’t tell the difference, users who will just use what Windows comes with and not really care. This leaves the developers out in the cold, as is usual for Microsoft. Although I’m sure all .NET programmers will be fine as usual.

    With the way things are going now, IE’s market share will continue to decrease in favor of better products. If only the US would impose some of the EU restrictions on Microsoft so that IE becomes completely independent of Windows. But, I’m pretty sure Microsoft dumps a lot of money into preventing that from happening.

  61. JP says:

    About the speed of opening a new tab.  In my opinion, it should be instantaneous (like it is in firefox, opera, and chrome).  Even 0.2sec is too slow because it is a noticable lag and makes the browser feel unresponsive.  

    When I hit ctrl-T a new tab should come to the front immediately, and my cursor should immediately move to the address bar.  When I say immediately, I don’t mean in under one second, I mean with no perceptible time lag.  

    In my opinion, this is the worst flaw in the user experience in ie8.  

  62. steve says:

    Why does clicking "Delete Browser History" (select all) take Sooooooooooooo long to delete everything?

    I can do a full delete of my history, etc. in Firefox in less than 2 seconds.

    steve

  63. Kin says:

    Oh no, you guys just made IE8 RC1 behave like Chrome which annoys the hell out of me. In Chrome or FireFox, when clicking the Back button to return to the prior page, the old page is first displayed from the top AND THEN it jumps to the old scroll position.  Now IE8 RC1 is doing the same, I really have no reason to use IE8 anymore.  PLEASE PLEASE I don’t mind the slowness but don’t FLICKER the user like crazy.  If you want to find out the effect, go to ebay and search something, then scroll to the middle of the resulting search page, then click an item, the click the BACK button and you will see the "Top then old spot" page flicker. Basically all "list based" pages will have this problem.  IE7 does not have this problem.

  64. Creating new tab (Ctrl+T): too slow

    ———————————–

    I’d like to add my voice to all those who said that creating new tab (Ctrl+T) is too slow. It is too slow. No matter which or how many add-ons are enabled.

    CSS bugs

    ——–

    There are still many, many CSS bugs remaining to be fixed. A good bunch of them.

    Developer tools

    —————

    Developer tools are a weak, very weak area in my opinion.

    - I can not see/get a list of CSS parsing errors: so, realistically speaking, how is a web developer going to upgrade webpages, to correct errors, etc with IE 8 dev. tools? No error console like Firefox 2+ has, like Opera 9+ has, like Safari 3+ has, like Amaya 11+ has, Icab 3+ has etc

    - I can not see a list of validation markup errors: so, again, the developers tool have very limited usefulness, helpfulness, relevance. I can do that instantly – offline or online – with a Firefox add-on. Already, people have been filing and reporting webpage layout issues in bug reports at IE beta feedback and they have no clues, no hints, no ideas whatsoever that their webpage is full/has lots of markup and CSS errors. How are they, how can they upgrade their webpages and websites without those?

    - no possibility of examining computed CSS property values, except the layout box model values (offset, margin, border, padding, content) and even there, "auto", "medium", etc still do not refer to computed values. This is one area where all other browsers (Firebug, DOM Inspector, DragonFly, Safari Web Inspector, etc) are clearly ahead of IE 8.

    - no built-in HTML Tidy, no HTML Tidy add-on: such would be useful for old FrontPage generated webpage or MS-Word exported webpage, you know…

    - view source: it’s not as developer-friendly as it could and should be. I shouldn’t have to set its text size all the time. Also, linefeeds, CR should be saved too (there is a bug filed on this)

    Documentation

    ————-

    This is the weakest area, weakest element by far of the IE 8 package. Everywhere I read, everything turns around the same repeated ad nauseam issue: how to switch to compatibility view, how to use the metatag to opt-in into IE 7 rendering, to click that broken-document-switch-to-compatibility-view button, to download a compatibility list, etc.. Nowhere at MSDN can I read anything instructive, helpful, useful about how to upgrade webpages, to upgrade "tag soup" webpages, how to correct (validation, parsing) errors, what to look for (typical errors, misnesting, doctype decl., recommendations, tips, web standards reference), how to create a lean, performant and efficient stylesheet, how to properly structure a webpage, what is "classitis", "divitis" and why they’re bad, etc, etc. None of this type of documentation exist at MSDN. Zero.

    User agent string detection as a way to sniff, to detect browser is refocused again in several documents at MSDN when lots of expert web designers on the subject of cross-browser development have said that object/method support detection is more reliable, more manageable, more forward-compatible.

    Gérard Talbot

  65. Dustin says:

    @Those who complained about the developer tools:

    First of all, at least they are there to help a little bit.  It’s a new feature, so it probably will have bugs.  In addition, the looks don’t matter.  It seems to fit well with the browser’s default theme, so does it really matter what it looks like?  The IE team is restricted in what they can do with it anyway because of Windows itself, unless you want them to use a bunch of custom images and everything to slow the browser down more (yes, I’m aiming this at those that complained about new tabs being too slow despite the fact that they are slow here, too).

    @Gérard Talbot:

    The documentation isn’t there because it isn’t needed for authors.  Authors should know how to fix these things.  After all, stylesheets, doctypes, etc. aren’t a part of IE’s problems.  They’re the authors’ problems.  The only standard that IE seems to want to follow is ECMAScript, aside from some RFCs regarding internationalization and the Unicode standard.  The W3C follows standards, and they also make use of their own recommendations.  Notice that I said "recommendations".  That’s all they are; they are not standards.  None of them actually need to be followed.  IE implements some of them because they’re useful (like CSS for example) and because they’d be dead already if they didn’t.

    @IE Team:

    Great job with the new IE thus far.  Keep up the good work, and you might regain the share you lost to alternative browsers like Mozilla Firefox.

  66. Leslie Glen (Les) says:

    Gotta concur with the other people complaining about tab performance.

    Browsers that are virtually instant in creating new tabs:

    Mozilla Firefox 3.0.5 (latest, /w around 18 extensions enabled), Safari for Windows 3.2.1 (latest).

    Still faster than IE 8:

    Google Chrome 1.0.154.46.

    2 year old laptop with XP MCE, SP3. Intel Core Duo T2050 (1.60 GHz, 533 MHz FSB), 504 MB of usable RAM. No known malware infection. Not ideal (more RAM would be desirable), but fast enough (except for Visual Studio 2008, which is a DOG even on the fairly new, powerful machines we have at work – an unrelated rant).

    I would be more forgiving if I had the security enhancements that Vista and Windows 7 people get (and that it appears Chrome was able to implement in XP).

  67. hAl says:

    <blockquote>No matter which or how many add-ons are enabled.</blockquote>

    New tab opening seems fast enough. Less than or close to a second and I have an AVG plugin enabled that takes about 0.5 of a second to start.

  68. Marco says:

    What a lame post. Why not simply start to adhere to international standards, once for all? Pick WebKit. It’s already out there and it works. There would be nothing wrong, MS already picked up code from mosaic a long time ago. It’s never too late to fix things.

  69. Mitch 74 says:

    @EricLaw: Daniel’s report does the trick. I can’t test under Vista, so it may be a platform-specific (XP-only) bug.

  70. hAl says:

    @Mitch,

    I can see Daniels testcase failing in IE8 RC1 on Vista.

    Nut sure if the testcase is correct but it does fail on Vista as well

  71. hAl says:

    @Mitch,

    I can see Daniels testcase failing in IE8 RC1 on Vista.

    Nut sure if the testcase is correct but it does fail on Vista as well

  72. norman says:

    Well according to Web bug track – Microsoft has acknowledged the bug with slow tab opening but has indicated they have no intention of fixing it before IE8 ships.

    bug reference: (search Web Bug Track) I tried to put a link in here but the stupid antispammer bot for this sites sucks so bad that any post with a link is considered spam!)

    This is too bad! Wish this was one of the first things addressed in IE8 development since the issue was only introduced in IE7.

  73. EricLaw [MSFT] says:

    @Norman: I’m not sure what "acknowledgement" you’re referring to exactly, but I can tell you two things: 1> slowness in opening tabs is demonstrably related primarily to slow browser add-ons, and 2> we continue to work hard on improving the performance of the browser in targeted scenarios, including startup.

  74. DT says:

    @ajaxwiz

    > Also, that link refers to JScript, which is

    > NOT JavaScript, but rather Microsoft’s

    > implementation of JavaScript. I can only

    > hope that’s just an old reference and IE8 is

    > not actually using JScript.

    JScript is IE8’s implementation of ECMAScript, not JavaScript. I can only hope that you also take issue with all of Mozilla’s non-standard extensions to JavaScript, their implementation of ECMAScript.

  75. ajaxwiz says:

    @DT

    Yes, you’re right. I only take issue with Microsoft’s implementation because it’s so slow.

  76. UloPe says:

    No CSS3 (whatever that will eventually exactly mean) is just SO pathetic.

    Take a look around guys. You high five yourselves with CSS 2.1 compatibility while your competition has nearly arrived at CSS 3 (Webkit, Gecko, Presto).

  77. Leslie Glen (Les) says:

    After disabling Research, Java (sorry Applet developers), and Acrobat-related plug-ins (leaving only "Diagnose Connection Problems" and "Report a Webpage Problem" enabled), creatings tabs was much faster.

    Most users aren’t going to have that setup. In fact, most users are probably going to have more toolbars and BHOs installed (Google, Yahoo!, MSN, Alexa, Ask!, McAfee, Symantec, MyWebSearch, etc) and enabled.

    Why doesn’t Firefox have this issue? Or really… any other "major" browser?

  78. jesse says:

    @EricLaw [MSFT]

    Re: 1>

    Either list the "naughty" addons that cause IE7 and moreso IE8 to open new tabs slowly, or refrain from trying to blame someone else s software.

    We’ve all tried IE7 and IE8 with EVERY addon disabled and the performance of IE opening a new tab vs. ANY other browser PROVES that IE is significantly slower.

    Then when you do enable the addons that you NEED to use on a daily basis to make IE a usable browser, the performance gets worse.

    Re: 2> If Microsoft *really* is working on fixing this issue with IE, then (*cough*, *cough* you’ve just acknowledged this issue *again*) please post an article here on the blog, indicating that it is slow and that you are working on it, and that it will be fixed before IE8 RTM ships.  If not this banter about rouge addons is just hearsay and a bunch of FUD to try and pull the wool over on us before the IE8 RTM launch.

    Only hAl, Dan & Ted the EPIC FAIL have stated that they think new tab opening speed is fine.  Considering their reputation (ms fanboys) I don’t think the evidence from everyone else that there is a serious performance issue can be dismissed by blaming 3rd party software.

  79. DT says:

    @jesse

    > Only hAl, Dan & Ted the EPIC FAIL have

    > stated that they think new tab opening speed

    > is fine.

    What about Arieta? And tabs open fine on my end as well.

  80. John Hrvatin [MSFT] says:

    Hi everyone,

    Regarding the console object…

    The console object is only added when the tools are opened.  The best practice here is to check for existence of console and then provide a custom object (possibly just a no-op) if it doesn’t exist.  This also accounts for any browser that doesn’t support console.  I believe some libraries already do this.

    And hopefully the script errors generated will prevent unwanted console statements from accidentally ending up in live code.

    John [MSFT]

  81. Mitch 74 says:

    @hAL: Daniel basically took the test I created, isolated it on a single page and made it simpler (replaced unordedered list with multiple list elements with a single paragraph, replaced generated content with a thick border).

    His test is correct, and if anything "more correct" than mine because it’s simpler and doesn’t rely upon features added (generated content) or fixed (ghost spacing between unordered list elements) in Trident in this major revision. If it fails on your Vista machine too, then it’s not platform-specific.

    I should mention that my IE 8 install is raw: there’s no Java, no .Net, no toolbars, nothing installed on this machine apart from stock Windows XP Pro SP3, security updates and stock IE. Heck, there’s not even third-party hardware drivers! It runs in a VM emulating either a Cirrus 5446 video card, or a plain VESA-compatible card (using Bosch’s open VGA BIOS), and all drivers are Microsoft’s.

    I’ll try and create a new, fresh, no options install, and see if I can still reproduce this bug.

  82. vasko_dinkov says:

    Not sure whether this would fall under "critical issues" (probably not :)) but here’s a very simple bug:

    ——

    <span id="s1" style="border:10px solid red;padding:5px;background:#bbb;">Incorrect offsetWidth returned for inline elements with border/padding set</span>

    alert(document.getElementById(‘s1′).offsetWidth)

    ——

    IE8 will not take into account the border/padding.

    Sorry for posting this here but, unfortunately, as far as I know, you still don’t have a public bug tracking system ala Bugzilla..

  83. Mitch 74 says:

    Addendum: just tested on a clean XP sp3 install (fresh off an Windows XP Pro+sp3 CD) + IE8 RC1, with no accelerator, no filters, no nothing (in fact, all UI improvements added over IE6/7 and asked for in the startup wizard have been deactivated).

  84. steve_web says:

    @EricLaw[MSFT]

    Re: [I’m not sure what "acknowledgment" you’re referring to exactly] – Regarding slow opening and responsiveness of new tabs.

    Quite simple really, the first major acknowledgment was on September 22, 2008.

    View this bug report in Connect:

    https://connect.microsoft.com/IE/feedback/ViewFeedback.aspx?FeedbackID=366238

    And note the comments (by Microsoft) in this "Closed (By Design)" bug report.

    [QUOTE]

    This is a long known issue that opening a new tab in IE will be slower than other browsers.

    The main reasoning behind this is due to the current architectural design in IE

    and maintaining backward compatibility issue.

    [/QUOTE]

    Well hello there! Honesty! Welcome to the discussion! Microsoft states:

    "This is a long known issue that opening a new tab in IE will be slower than other browsers."

    [QUOTE]

    Other browsers do not build on the same architecture therefore do not suffer this problem.

    [/QUOTE]

    What is that you say? IE has a problem, that other browsers do not suffer from?

    [QUOTE]

    We are actively looking for ways to improve performance in IE8, including this known limitation.

    [/QUOTE]

    Known limitation? Well apparently only the IE users are aware of this because EricLaw[MSFT] is not sure.

    [SKIP]

    link to docs on other unrelated performance tweaks in IE

    [/SKIP]

    [QUOTE]

    Best regards,

    The IE Team

    Posted by Microsoft on 9/22/2008 at 9:40 AM

    [/QUOTE]

    It should also be noted that this bug has:

    a 4.7/5 rating in terms of importance

    a 93% validation level

    and 0 workarounds

    Followed by the most recent IE chat (re: slow tab opening)…

    [QUOTE]

    JohnHrv [MSFT] (Expert)[13:27]:

    "there may be ways to make the process more efficient and we’ve been working on that in IE8."…

    …"so it’s definitely an area that could use optimization"

    [/QUOTE]

    I’m not trying to start a fight here, but lets put the "blame the user’s addons" to rest, admit that the issue is definately there, and that something needs to be done about it before IE8 goes RTM.

  85. @ vasko_dinkov

    Bug 389825: Inconsistent offsetHeight for inline elements

    connect.microsoft.com/IE/feedback/ViewFeedback.aspx?FeedbackID=389825

    Regards, Gérard

  86. EricLaw [MSFT] says:

    @steve_web: You’re pretty much missing the point, but I can’t blame you when the response was so unclear.

    "The main reasoning behind this is due to the current architectural design in IE and maintaining backward compatibility issue."

    That is referring (in a painfully obtuse way) to the fact that browser add-ons are loaded on a per-tab basis, and thus if you have poorly-performing add-ons installed, they will cause slow tab startup.  That architectural design maintains backward compatibility with existing extensions (which are coded on a per-tab basis).

    The response was entirely misleading in suggesting that the issue was not conditional upon having poorly-performing add-ons.  When only high-performing add-ons are running, new tabs appear in well under 1 second.  

  87. Seth says:

    << You high five yourselves with CSS 2.1 compatibility while your competition has nearly arrived at CSS 3>>

    Ummm…methinks you missed the part where no other browser currently gets CSS2.1 right.  When I went to school, you didn’t get to move on to the next assignment until you correctly finished the current one.

  88. steve_web says:

    @Eric Law[MSFT] As I and many others can assure you, the performance of opening a new tab in IE8 is slow regardless of how many addons you have installed, or if you have _any_ addons installed.

    However, if you insist that it is buggy addons that slow things down, then for pete’s sake point the finger so we know which addons to remove.

    Likewise, when running IE8 in no addons mode, the link in the Tools menu to go and find addons that are causing trouble (e.g. to go delete them) is grayed out. You need to go into Tools > Options > Addons > Manage Addons in order to see them.

    Finally, please fix whatever is BLOCKING the DELETE button from being displayed beside ***ANY*** addon.

    I don’t want to "disable" the "Research" addon, I want to throw it off a volcano into a pit of boiling lava.  It serves me no purpose other than to clutter my system and slow down my IE browser.

    Believe it or not, but in terms of IE development priorities I would put the IE Tab opening speed WELL above SVG, W3C event listeners and CSS rounded corner support… Well above.

    This is a performance issue that EVERY SINGLE IE User will encounter almost every day and in many cases several times an hour.

  89. yoshi says:

    can you add a maximum box to the print preview window? The border of it can resize, but we cannot max it ?

  90. bart says:

    from the help tips at:

    http://www.enhanceie.com/ie/troubleshoot.asp

    #5 Once you’ve found a broken extension, contact the manufacturer and ask for an update.

    Dear Microsoft, please make haste with Windows Update to completely remove the *Research* addon from Internet Explorer.  Feel free to do so with our without informing the MS Office team that you are fixing their code by removing it completely.

    Thankyou

  91. steve_web says:

    Well it seems I’ve uncovered a rather large failure in IE8 RC1.

    The "Delete" button that was grayed out for extensions was actually on an IE7 install.

    In IE8, NO SUCH BUTTON EVEN EXISTS! so there is NO WAY TO REMOVE an extension (Good or Bad)

    ===========================================

    I see this as a monumentally major issue.

    ===========================================

    How do I ever remove a spyware toolbar if it gets installed?

    Even if I open IE with extensions turned off, I have no way to delete them.

    What about extensions that I need one month, but not the next? (Citrix anyone?)

    Adding and removing addons should be JUST as easy as it is in Firefox.

    At any time I can pick an extension and disable it, customize it, or uninstall it.

    steve

  92. Daniel says:

    Steve, if you spent more time looking into things and less time ranting, you might get more respect.

    The "Delete" button was renamed "Remove" and it’s now behind the "More Information" link for each addon inside Manage Addons.  It’s not always enabled because sometimes you must run an uninstall program.

    A spyware toolbar isn’t going to support standard unregistration with the Delete command anyway.  You use anti-spyware programs to remove malicious software.  Of course, you should probably try to avoid installing malware in the first place.

    You can almost always uninstall legitimate IE plugins using the Add/Remove Programs control panel.

    It’s true that Firefox has a much different, less powerful, non-COM extensibility model.  They have to, because they run on platforms other than Windows.

  93. steve_web says:

    @Daniel – a few notes:

    1.) I’ve submitted a bug in connect for this:

    https://connect.microsoft.com/IE/feedback/ViewFeedback.aspx?FeedbackID=410121

    2.) Yes you are right, there is a "Remove" button burried deep inside the the addons dialog interface.  However there is no non-grayed "Remove" button available for _any_ IE addon in my list (maybe yours is different?)

    3.) Case in point, there is no item in the Add/Remove programs list for "Research" nor "Microsoft Research"… I have to "know" that it is burried deep into the Microsoft Office Tools application and that I need to "Change" my MS Office install to find and remove the "Microsoft Office Research toolbar" in order to fix this.

    4.) I don’t expect a spyware toolbar to provide a "delete" option, but I DO expect IE to provide one for me. IMHO the install should unconditionally fail, if an un-install is not provided.

    5.) I don’t believe that anyone is claiming that Firefox’s extensions are "less powerful" (nice try).  I have many extensions for Firefox that IMHO provide more power to me as an end user than any addons I have for IE.

    6.) I’m pretty tech-savvy compared to most windows users… if *I* had a hard time trying to find out where to delete extensions… how will the *average joe* user find it?

    steve

  94. fixperfo says:

    Please fix scrolling performance on pages with Flash. For example, IE8 scrolling performance is pathetic on youtube. Firefox and Safari have no problems.

  95. ike says:

    you are so out of business. IE is just ridiculous  performance sucks compare  with the other browsers, IE user experience is the worst and should be really embarrassing the fact that google’s chrome is way better that all the other 7  version that IE.

    MS

  96. yoshi says:

    sometimes i got a dialog box warning me for an application error, no IE crashed actually as i observed.. any way to generate some log files for your team to inspect?

    The dialog box detail:

    Title:

    iexplore.exe – Application Error

    Content:

    The instruction at "0x00d5ea13" referenced memory at "0x6363d854". The memory could not be "read".

    Click on OK to terminate the program

  97. JP says:

    @EricLaw [MSFT]: I agree with steve_web, IE tab opening speed should be top priority.  

    Tabs opening in under 1 second is not good enough.  They should open *instantaneously* with or without add-ons (just like every other browser).  

  98. SylvainG [MSFT] says:

    vasko_dinkov,

    You’re correct, our offsetWidth is right for blocks but wrong for inline elements. Good catch, thank you !

  99. Klimax says:

    For people with slow tabs:

    You have something wrong on your end,because tabs slowly opening are only in two discreet cases:

    1)computre is overloaded with applications.Not MS fault.

    2)there is big bad javascript running in multiple tabs.

    I have never seen otherwise problem you describe.Tabs are loading (in numbers when instructed to) in a matter of miliseconds,there is no room for improvement.(Yes the only addons are orbit downloader,Sun Java and Adobe Acrobat reader plugin)

    First look for problems on your side,not in IE.Maybe you managed to damage XP or Vista.(Like manual fixing of problem going wrong)

    But I need IE8 RC1 for Win7.Thats the only thing missing…

  100. Alahmnat says:

    CSS-only scrolling tables STILL don’t work in IE8 RC1, even those explicitly called out as working in IE6 and other major browsers (Firefox, Safari, Chrome). Examples:

    http://www.imaputz.com/cssStuff/bigFourVersion.html

    http://www.imaputz.com/cssStuff/bulletVersion.html

    In both of these examples, Gecko and WebKit browsers display a table with a scroll bar to display additional content. IE8 RC1 displays a table without the scrollbar, making it impossible to see anything beyond the bottom extent of the table "frame". In IE7 compatibility mode, the first row fills the entire height of the parent table, with no scroll bar to see additional rows, which is arguably even less useful.

    I really hate that I have to resort to JavaScript libraries or goofy CSS hacks that then break table scrolling in other browsers to make IE behave correctly. It’s a waste of my time and the computer’s resources (especially since JS speed in IE8 is still woefully behind the times).

  101. Pigglety says:

    Thank you for posting information about the platform improvements. However, I was hoping to find some information about about any changes to HTML+TIME (XHTML+SMIL) in IE8. Perhaps somebody from Microsoft could blog about this topic. I have not seen the animateMotion or transitionFilter elements work in IE8 RC1 regardless of which compatibility mode I try. I’d like to see some assurance from Microsoft that these will be working when IE8 is released.

    I’d also like to know whether IE8 will still support SMIL 2.0 or a more recent SMIL specification.

    After reading the Improved Namespace Support document, it sounds like developers will not have to use <html xmlns="http://www.w3.org/1999/xhtml&quot; xmlns:t="urn:schemas-microsoft-com:time"> and <?import namespace="t" implementation="#default#time2"?> in web pages … they should be able to replace these with <html xmlns:t="http://www.w3.org/2001/SMIL20/"&gt; and IE8 would then use the preinstalled time2 behavior. It would be helpful if somebody could confirm whether that will be the case.

    Thank you for your time.

  102. SylvainG [MSFT] says:

    @Gerard, re: bug 389825.

    You are right, this bug was not fixed in the RC build. The issue will be fixed in the final release. Our apologies.

  103. JP says:

    About tabs opening slowly, it is not something wrong on our end.  It seems to be the implementation in IE.  

    When I hit ctrl-T, I first see a tab opened in the background that says "connecting".  Only after time lag (sometimes slow and sometimes fast, but never instantaneous as in other browsers) does that tab come to the front and the cursor move to the address bar.  

    I would be perfectly happy with the following solution.  When I hit ctrl-T, IE should instantaneously bring the new tab to the front and move my cursor to the address bar so that I can type.  I don’t care if something is connecting in the background, I just want to see the new tab immediately and I want to be able to type immediately.  

    Showing a tab "connecting" in the background, as IE does now, is pointless and annoying (yet that seems to be able to appear almost instantaneously).  That visual step should be removed.  All I want is for the new tab to be the focus and my cursor to appear in the address bar and I want that to happen as fast the background tab "connecting" appears now.  I don’t understand why that is not possible and why it hasn’t been implemented that way.

  104. Dan says:

    JP, I don’t know why you have such a hard time understanding this.  

    Poorly-performing addons get loaded on every new tab and result in the problem you’re seeing.  When you don’t run the bad addons, you don’t have the performance problem.

    The IE guys made a special mode that fixes this problem, called No Addons mode.  Run IE with -extoff, and watch your problem go away.

    Of course, you could just disable or uninstall the bad addons and get the same effect, but for some reason you don’t seem to be getting that message.

  105. Morton says:

    Can Dan and Klimax clean out their ears!!!

    The rest of us find tab opening to be the slowest feature in IE and far slower than any other browser out there (hands down, no exceptions)

    This includes:

    * Running in No Addons mode

    * Having no other tabs beyond about:tabs or about:blank open

    * Running on a new, powerful PC

    WE ARE SICK AND TIRED OF BEING TOLD IT IS OUR FAULT!

    We are end users giving an honest opinion about a product.

    Outside of a few fanboys on this Blog, I have found (NOT A SINGLE LIVING SOUL!) that feels that the IE new tab opening speed is even ACCEPTABLE.  We are not talking about fast vs. super fast.  We are talking so gosh darn pathetically slow that we feel the browser is shown in a bad light by this flaw.

    I don’t care what the story or PR angle is – all I can tell you is that users ARE GOING TO HAVE SEVERAL addons RUNNING all the time…

    Acrobat, Java, Flash, SVG, Live Toolbar, Google Toolbar, Research, etc.

    The load time without these addons is slow, and it is frustratingly painful with the addons loaded.

    I don’t care who claims responsibility, who fixes it, or what. It just needs to be fixed.

    Morton

  106. fixperf says:

    Too many fanboy can’t see the truth. Run Firefox and Safari and you’ll see that IE runs pathetically slow. Safari opens tabs instantly but there’s always a delay with IE.

    IE is so slow that opening a new window in Safari is faster than opening a tab in IE.

    NO excuses. All other browsers do not have this problem. Blaming add-ons is just smokescreen.

    Unfortunately, I wouldn’t be surprised if this never gets fixed. Microsoft has a habit of keeping minor, but annoying bugs unfixed for a very long time. But don’t lose hope. Maybe this will get fixed in IE9.

  107. Max says:

    Tabs in IE are opened in separate processes as a security measure. It might explain why it takes longer than in other browsers.

  108. @Marc Silbey, EricLaw [MSFT] and others

    > We take your feedback on critical issues very seriously

    Here’s one which is rather serious, I’d say.

    Load

    http://www.gtalbot.org/BrowserBugsSection/MSIE8Bugs/col-can-make-webpage-blank.html

    and test thorough the page: move the mouse in and out of <body> height (or from chrome, say, status bar to window viewport and vice versa) after zooming in or out, after clicking Back or Forward buttons. Depending on your setting "Automatically recover page layout errors with Compatibility View", RC1 build 18372 will indicate that it is struggling with the page… or the whole page will become blank.

    Developer tools will also show some troubling issues with original code:

    <col></col>

    becomes in dev. tool

    <col/></col/>

    which is unusual and unexpected.

    More on all this in bug 409478

    Regards, Gérard

  109. steve_web says:

    I agree with Gérard bug 409478

    https://connect.microsoft.com/IE/feedback/ViewFeedback.aspx?FeedbackID=409478

    is a fairly major rendering glitch.

    That and the bugs with offsetHeight, offsetWidth, iframes not firing resize events for Vertical size changes, new tab opening speed, the ability to delete addons and various chrome and zoom issues.

    Has there been confirmation on an RC2 release yet? or is Microsoft just going to "hope" that they’ve fixed everything and go RTM to let developers start testing/porting their code to work in IE8.

    Yes, ***start***. There are too many bugs in RC1 that us developers are expecting to be fixed for us to start coding workarounds for. (where even possible)

  110. Alahmnat says:

    @MAx

    Chrome launches tabs in separate processes as well, and it does so instantaneously. IE is the only browser that can’t create tabs with consistent speed.

  111. Dan says:

    Alahmnat, funny, Chrome is the only browser that doesn’t support extensions.  Coincidence?  No.

  112. Klimax says:

    i tested RC1 on XP.Tabs (multiple at once) loaded very quickly(miliseconds).

    I t is problem in your setup.When I see no problems whatsoever on balasted and damaged XP which are running BOINC projects at full with all cores maxed out,with memory already filled by projects and still manages to beat your system,then sorry.

    And Firefox is the  worst.When only one tab present,the tab-bat is hidden and subsequently it will reappear upon second tab opning.THIS IS BAD!!! And its performance is not good.It is same and UI is unusable(too many bad decisions).Yes I am 20 hours a day on net,opening a hell lot of pages (4 windows,each 5-70 tabs) and I need reliable browser…(and FF is missing full recovery after crash)

    I hear you,but the problem is on your side.I suspect problem is in registry.(Some sort of corruption – I had it as well,but as a result of five crashes in a row)

  113. james says:

    @Klimax – get down from your high horse for a second.

    - Your IE tab opening may be very fast, congratulations

    - I’m sorry to hear your Firefox is causing issues, the rest of us have no problem

    -Klimax stop reading here.

    Now for the rest of the planet that finds IE new tab speed to be:

    a.) very slow

    b.) very slow in no-addons mode

    c.) slower than Firefox

    d.) slower than Safari

    e.) slower than Opera

    f.) slower than Chrome

    g.) unacceptably slow

    h.) ALL OF THE ABOVE

    Feel some comfort. You are not alone. In fact you are in a special little group we call "The Majority"

    I think it is high time this blog had a running poll, so that when questions like this come up, your end-users (not IE Dev Team users) can report real-world findings.

    e.g.

    Today’s Poll.

    Do you find the speed of opening a new tab in IE8 RC1 to be:

    (_) Very Fast

    (_) Pretty Fast

    (_) Average

    (_) Slow

    (_) Very Slow

    I would expect the results would favor the "slow" options by a 9:1 ratio or better.

    For those few that still think IE8’s tab opening speed is awesome, please post a video on YouTube (or similar) showing that you have some extensions enabled (yes, this is a real-world test, not a mythical fantasy candy land test) showing how fast your new tabs open up, and post us a link back here on the IE blog.

  114. Vinifera says:

    i realy see no reason for being excited about IE8, fact is that dev team or doesn’t get enough resources (coders and/or time) or MS itself doesn’t realy put much attention to its browser since users will get it no matter what kind of garbage it is…

    since IE 5.5 i dont see any improvement put in IE and worst thing is, you guys put new numbers on it, change the GUI and some little things under the hood and VOILA its NEW and BETTER… no it is not…

    you "solved" security in IE7 (and updated 6) on easiest and lazziest way possible: just block any possible script…

    im just saddened that users will NOT see or "experience" good IE untill maybe version 15, but by then i guess it will again be passed by other browsers…

  115. Steps:

    Load

    http://www.gtalbot.org/BugzillaSection/DocumentAllDHTMLproperties.html

    1- scrollWidth must always be greater or equal to clientWidth: that’s by definition (MSDN and CSSOM: http://www.w3.org/TR/cssom-view/#scroll-attributes ). So here, RC1 build 18372 has an implementation bug and a regression in comparison to IE 7 and IE 6

    2- Clicking the [right to left] radio button makes scrollLeft becomes -28 . That was not so in IE 7 and IE 6. Regression.

    3- Clicking the right to left radio button should increase clientLeft of the width of the vertical scrollbar (eg from 24 to 41; if I have Display/Appareance tab/Font size: Large Fonts, then from 24 to 43)

    That was not so in IE 7 and IE 6. Regression.

    Regards, Gérard

  116. Mitch 74 says:

    @EricLaw: a complete answer to your inquiry about the regression I mentioned:

    – it is present in Partners Build and in RC1. I have no idea how it behaves in internal build, as only you (and your team) have access to them. If I could try and see if it reproduces, I could confirm it solved. As it stands, the latest public build fails the test and has a regression over Beta 2. If you want that to change, release an RC2.

    – I installed Windows XP SP3 off a retail CD, applied all core system updates (no additions like .Net), rebooted twice, then installed IE8RC1. I didn’t tinker with settings such as zoom, developer tools, nothing: I typed the test URL, hovered the area with the mouse cursor, then left it, and IE failed.

    – considering it is a box sizing issue, it could have been fixed along with bug 389825, which was supposedly fixed in RC1 but wasn’t. Next time you release a build, please make sure you don’t forget to integrate patch sets – or you’d better release public builds more often. For me RC1 isn’t an RC, it’s beta 3.

    @Those who complain about speed: without add-ons installed, IE8 RC1 is about as fast in day-to-day operation, on the same machine, than a Firefox/Minefield 3.2 daily build with JIT Javascript enabled for page (which is a fair speed gain over IE7’s ten-times-slower than Firefox 3.0); that it fails tests such as Acid3 makes it fall in the "don’t make creative new websites" category, but there certainly is notable improvements.

  117. Stuart says:

    Jeffrey I too get the maxed out CPU when dragging windows even if IE8 isn’t running at the time. This suggests that some core GUI-related Windows core files do get updated with IE8 which explains the reboot. Interestingly, examining Task Manager, the extra CPU cycles are being used by:

    a: The application whose window is being dragged

    b: Any applications that the dragged window is being dragged over

    This suggests to me that video hardware acceleration has been turned off much like in Safe Mode and the applications are having to continually report to the GUI what it should be displaying during each rendering frame of the desktop.

  118. Stuart says:

    Re earlier comments I notice a pile of dlls which have been updated in the System32 directory with the IE8 install , including some related to DirectX eg dxtrans.dll.

    I also noticed while in system32 that scrolling through the directory maxes out CPU and the scrolling freezes. Going to roll back and see if it fixes

  119. Sundar says:

    IE8 incredibly slow executing Javascript.

  120. Stuart says:

    Rolling back had no effect so I’m guessing it is standard behaviour, it is the same on another PC running IE7 which has never been upgraded to IE8 RC1.

    Apart from the slow tabs (probably due to all the toolbars loading in a new process, I added up the time of all of mine in the manager and it was more than 10s) the main annoyances are slow screen-rendering when scrolling and occasional "Diagnose connection" pages coming up when IE7 or Firefox have no problems. Also it kills off my auto-updating Active Desktop.

  121. Klimax says:

    James,you really seem to discount possiblity I mentioned.I do not see that you have really investigated it.(I did,since I saw performance before and after and behaviour after was same as you and others described)

    I suggest to use Proccess Monitor or similar tool to capture activity in IE processes.

    And btw. there is some probability that some update or change to affected part of registry(or dll) will solve it.Since all dlls are loaded on per-tab basis such corruption or badly behaved dll can cause it.It solved it self before I was able to do full investigation,but PM showed some problems,start by those.And that is indicated by IE NoAddons behaviour(same).

    And last be sure to disable ANY process/services which are interacting with IE.(AV,monitoring tools when not needed and such including services!)

    -Thats why I say IE team cannot do with it anything.

  122. steve_web says:

    @Klimax – As several people have noted IE8 is slow opening tabs with or without extensions, under condition x or y, etc.

    I’ve started a bug report in connect with some real data from environments I have with and without extensions loaded etc. (once I find somewhere to host them, I’ll share the tools I used to track each browsers performance also)

    I’ve shared these with some developers I associate with, and they all return similar results.

    See the bug report for details. 411584

    https://connect.microsoft.com/IE/feedback/ViewFeedback.aspx?FeedbackID=411584

    As for the "better" performance with no addons loaded – that may (or may not be true) but the Reality is that we are all going to have them loaded, and/or need them, so a no-addons mode for daily browsing just simply is not an option.

  123. Dking says:

    I’d like to express my thanks to the team for the CSS improvements, particularly the page-break code. It has been a life saver!

  124. Chris says:

    With add-ons (Flash, SkyDrive uploaded, Silverlight, Java, Spybot, PDF):

    (_) Very Fast

    (_) Pretty Fast

    (*) Average

    (_) Slow

    (_) Very Slow

    Without:

    (_) Very Fast

    (*) Pretty Fast

    (_) Average

    (_) Slow

    (_) Very Slow

  125. Chris says:

    Time taken for blog comment to be submitted:

    (_) Very Fast

    (_) Pretty Fast

    (_) Average

    (*) Slow

    (_) Very Slow

  126. Rimbaud says:

    Please, please, please ditch your terrible trident rendering engine and just use Gecko or Webkit.  You have made web developers’ lives a nightmare for so long.

  127. Klimax says:

    @steve_web:

    OK.Only for comparsion(I forgot to include last info) it was measured on compaq nx7300.(Core2 duo with 1,5GB RAM)

    Your report has interesting results.Few details are however missing.What addons in each cases were,if any other software likely to interact with any browser was running(AV,Spybot and such) and specs of computers.

    I will test out my main PC with your EXE.(IE7,Core2 E4600?,1GB RAM,XP SP3,PATA HDD 250GB)

    And BTW this was missing in other reports pon this blogs.

    Thank you.

  128. Klimax says:

    Tool referenced in the link by steve_web is done against english version and does not work on IE with other languages.(regardless of version.)

    Will try to correct at least for czech version.

  129. Klimax says:

    Tool referenced in the link by steve_web is done against english version and does not work on IE with other languages.(regardless of version.)

    Will try to correct at least for czech version.

    (So far max was 269ms and median 109ms in row of 10 tabs,in already loaded browser on Win7,Lenovo R500 – core2)

  130. hAl says:

    Without addons

    (X) Very Fast

    With addons

    (X) Average

  131. steve_web says:

    @Klimax – hey, sorry about the language thing (one forgets sometimes about the other lang installs)

    In particular I noticed that the VERY FIRST tab created after the browser is opened takes quite a while. (e.g. not the homepage that is loaded, but the new tab created right after that)

    I’m not at my normal office so I can’t post any hardware specs ATM.

  132. Michael L. Parmley says:

    When I enable the acellerators, I get an error on every website of "The instruction at "0x…" referenced memory at "0x…". The memory could not be written.

    Click on OK to terminate the program.

    Click on Cancel to Debug the program.

    No matter what you do, you end up in a circle that requires "ctrl alt delete" "terminate program"

  133. Mr Squirrel says:

    If your tabs open slow, the blame goes to Java?

    Great. Now it’s all Java’s fault that IE can’t figure out how to play nice with the rest of the world. Again.

    As one of those rare folks who programs entire website backends in Java, and even writes game applets, I am tired of hearing Java blamed for your problems.

    I’ve occasionally had to uninstall Microsoft updates from my Windows machine, because they cause my Java applications to fail. As a result, I am biting the bullet and porting all my remaining Windows boxes to Linux. I am tired of Windows breaking my Java.

    It is as if Java is always the implicit evil here at MicrosoftCloudCuckooLand. I am dismayed to find out by reading these forums that Java has once again become the de facto bugbear for Microsoft’s development woes. It is amusing, in a sad way, but mostly just depressing.

    Anyway, good luck on achieving that day somewhere way off in the distant future, when your browser learns to play with the rest of the kids.

  134. Phil says:

    Squirreldude, Sun’s Java Virtual Machine and ie plugin is a performance, security, and reliability nightmare.  Of course, why would Sun bother to make it any good for ie/windows, since they have every incentive to make windows (which competes with their platform) look bad?

    Stuart, dxtrans is a part of IE’s display filters, and isn’t used by other programs.  it’s not a part of directx– it *uses* directx.

    Vinifera, IE "blocks any possible script"?  WTH are you talking about?