Engineering POV


To date, this blog has focused on the engineering specifics of what we’ve done with the IE product. From our point of view, it’s been a useful forum both for talking and listening. Looking at the comments, we can understand what makes sense to readers and where we need to be clearer.

At the same time, we’ve seen many questions about broader topics, like IE6, HTML5 and other standards, or benchmarking. With IE8’s release and Windows 7’s “sign-off,” now is a good time to add another kind of blog post. We want to use these posts to share our Engineering Point of View about broader topics and see feedback on them ahead of the next release.

Why? For many web technology questions, finding many passionate and often contradictory opinions is easy. For example, just on the topic of video codecs within HTML5 (much less the rest of the spec), finding strong language from smart people disagreeing with each other is easy. This blog is from the IE engineering team, and everything we write here continues to be from the “Engineering Point of View.” We simply want to be clearer about what we’re thinking and what we balance as we build and service IE.

Your comments are always welcome. We read all the comments on this blog (and many of the posts and comments on many other blogs). We’ll also keep posting and reading comments on specifics, like How to make IE open tabs faster and How to log into two webmail accounts at the same time. Comments about other posts you’d like to see are also always welcome.

Thanks –
Dean Hachamovitch
General Manager

Comments (24)

  1. Fred says:

    I am just curious why Microsoft is constantly changing, re-arranging, re-naming and basically totally re-working so many features of Internet Explorer? You take buttons which have always been on IE and remove them or relocate them, actually making it more inconvenient than it is convenient. You change the names of toolbars, make it even harder to find and make changes for customization. Improving the way IE functions is one thing, changing everything about it is another. Example: Why is the Links Bar now called the Favorites Bar? What exactly has that done to improve it? What exactly does all that do to make IE better?

  2. @ Dean Hachamovitch [MSFT]

    "

    Your comments are always welcome. We read all the comments on this blog (and many of the posts and comments on many other blogs). We’ll also keep posting and reading comments on specifics, like How to make IE open tabs faster

    "

    Really.. Is that so?

    Opening a new tab is too slow (bug 366238)

    Slow starting/Open new tab after immunized by Spybot – Search & Destroy

    (bug 424856)

    Opening a new page in a new tab is slow ("Connecting…") (bug 375006)

    IE8 RC1 – Slow New Tab opening (bug 411584

    These bug reports have all been closed by design (except for the Spybot S&D).

    Other bugs have been closed "by design" and they shouldn’t have been IMO: they should have just been postponed and they should have been reopened, reactivated. A few have not been reopened: no reason were provided as to why.

    I’ve asked how the IE team distinguish between by design and wont-fix and never got a response.

    Severity, gravity, importance, priority form fields for bug reports are what many other browser manufacturers or big corporations (NASA, W3C, etc) do, you know.. And they also use form fields like component, keywords (hasReduction, needsReduction, testcase), etc.. They don’t bug you with useless, pointless, annoying DHTML tooltips hovering over and again duplicating content, hogging cpu and memory, senselessly abusing user system resources, etc

    I am convinced IE 8 still has a lot of bugs (incorrect implementations of W3C technical recommendations: HTML 4, CSS 2.1, DOM 1 & 2) needing/deserving to be fixed. There are people (James Hopkins, Dan, "mvdleij", Hilbrand Edskes, Philip Taylor, Zoffix Znet, Marc Pacheco, Colin Snover, Garrett Smith, "the_dees", etc) who have so far convincingly substantiated this in their respective websites.

    I am convinced IE 8 has many accessibility failures (UAAG, WAI) or flaws; when I compare with Firefox 3.5 or Opera 9.64, this is utterly obvious to me. Accessibility is an area where IE 8 is far behind other browsers.

    Tuning, tweaking, improving connect’s IE beta feedback is an area where beta-testers have demanded consistently in many ways, many forms to be heard, listened. During months, years.

    HTML authoring tools (coming from Microsoft), softwares (coming from Microsoft) and MSDN documentation  for people who want to upgrade their websites are weak, incorrect, inappropriate, lagging, etc..

    MSDN web authoring webpages, articles, columns are full of validation markup errors, CSS errors, in the source code as well as in examples, code chunks. It’s been like that for the last 10 years and still today.

    Websites which are entirely under the control of Microsoft should show the examples regarding best, recommendable, forward-compatible coding practices (markup, CSS, semantic, accessibility, etc), interoperability and this is not the case at all.

    Every single post in IE blog stating that IE 8 uses by default the best web standards IE rendering engine has been forcing the compatibility view. A blatant incoherence, contradiction.

    Even posts, webpages about interoperability, web standards support have dozens and dozens of validation markup errors, CSS parsing errors, etc. Anyone can verify any/all of my claims here.

    In 2000, 2004 and 2006, we were told that valid compliant websites under the control of microsoft or created by its employees was a low priority. In mid-2009, it sure seems it’s still like that and that it will still be like that for the next 10 years.

    Are you still going to claim, just like that, in broad daylight, that IE Team listens, understands, cares, etc..?? And I’m not even talking about marketing campains based on greed ($10,000) or charity (hungry children) or over-excessively-stretched "facts" about IE 8 here… or invitation to strange, weird congratulations parties in Las Vegas (drink+eat as much as you can, pay for your travel expenses, spend as much as you can, etc.).

    Gérard Talbot

  3. Blake says:

    Gérard, these aren’t comments. They’re rants. My guess is that they’re reading and ignoring.

    What you document on your website matters much less than what you take to the CSS WG. That’s where there’s a test suite, and that’s where they’re taking feedback on their implementation.

  4. mogden says:

    The only thing I care about is more rapid adoption of the solid parts of newly emerging standards such as CSS3 and HTML5, and supporting ancient standards like SVG and MathML.

    If IE was up to date with Mozilla and Safari in terms of standards support, well, that would be the best thing in the world for a web developer.

  5. cpradio says:

    I have a question, hopefully it won’t offend you, but when do you see IE 8’s enterprise installer going to be ready and available for companies to push to their users?  We currently as a company refuse to support IE 8 100% until that becomes available to us to push to our users within the company.  From what I have been told from my supervisor, this has yet to be released and it gets getting pushed back month after month.  What exactly is the hold up?

  6. blogden says:

    @mogden: yeah, because so many people really use MathML and care. HTML5 is ridiculously under construction that supporting it is meaningless.

  7. Typhoon87 says:

    @ cpradio- Download IEAk 8 (Internet Explorer Administration Kit for IE 8) is already out.

    You can create an enterprise installer to distribute through your favoite automated deployment SMS, SCCM, Zen ect.

    I have created and am testing a deployment for my employer. You can preconfigure almost every option in IE if you wish.

  8. wai says:

    @Gérard Talbot

    for 424856, an update in few months ago fixed the problem in my pc.

  9. wai says:

    @Gérard Talbot

    > Every single post in IE blog stating that IE 8 uses by default the best web standards IE rendering engine has been forcing the compatibility view. A blatant incoherence, contradiction.

    EricLaw answered in previous post

    http://blogs.msdn.com/ie/archive/2009/07/28/internet-explorer-july-out-of-band-cumulative-security-update.aspx#9859283

  10. @wai,

    > EricLaw answered in previous post

    It’s still a blatant inconsequence/negligence/contradiction as far as IE 8 and Microsoft are involved that has been going on for months now.

    "compatibility issues in IE8 Standards Mode. Most of these occur when sites expect legacy behavior that no longer exists in IE8 Standards Mode. Upgrading your site to run in IE8 Standards Mode is the best option in the long run (…)

    SOLUTION: Ensure your markup is well-formed and valid.

    "

    2009/03/12/site-compatibility-and-ie8.aspx

    Microsoft can not seriously suggest/recommend to others to fix their markup code when microsoft has not been doing it in the last 12 years in all of its microsoft-controlled websites (including ones on web authoring, tutorials, etc like MSDN; newly created webpages as well) without running into serious credibility problem. This is a blatant wide-scale contradiction. There is (never was) no excuse for that.

    Gérard Talbot

  11. Gavin Greig says:

    I would like to second @mogden’s comment, particularly regarding support for SVG and MathML. Vector graphics have broad applicability that’s stifled by the lack of support for common standards; and mathematical notation, while admittedly a bit of a niche requirement from some points of view, is so fundamental to the communication of scientific knowledge that it deserves native support from mainstream browsers.

    I would welcome some explanation of Microsoft’s thoughts on implementing support for SVG and MathML; even if the indication is negative. That would at least save us from hoping only to have those hopes dashed, again.

  12. Stefan Wenig says:

    "We’ll also keep posting and reading comments on specifics, like How to make IE open tabs faster and How to log into two webmail accounts at the same time."

    Maybe I got that wrong, but comments for both posts are closed. (That said, I could post a suggestion via the Email link up there, and got an answer.)

  13. Chris says:

    I think MathML is a niche, and if it is "fundamental to the communication of scientific knowledge", I’m sure those who need it will simply grab themselves Firefox.

    I don’t think it’s worth implementing in IE 9, when other things like SVG, HTML5 and CSS 3 are clearly in more demand.

  14. Mitch 74 says:

    While Gérard’s tone is acidic, one has to admit he has a point.

    Or several, as it were.

    If something as ‘basic’ as list loading optimization escaped bug testing (it’s not as if the Spybot inoculation feature isn’t well-known, and saying that using an IE security feature present and working ever since IE 5 came out is a big no-no is, I think, quite rude), then it would stand to reason that others did, too – and, through my own tests and stuff, it seems that standards support in IE 8 is quite shaky: it works on simple test cases, but in complex situations, the layout engine goes ga-ga – or not.

    About MSDN code samples, then I wholeheartedly agree with Gérard:

    – the website is a DOG to browse: eventhough it mostly contains text, interaction is sluggish; maybe the amount of "warning"s appearing in my Javascript debugger is a hint…? Don’t mention the ‘low bandwidth’ version, please.

    – examples often contain gross mistakes or suboptimal code, when it’s not pure proprietary rubbish (calling for document.all in Javascript SCRIPT node in an HTML document) or don’t mention erroneous behaviour (getElementById returns a collection when a document contains – erroneously – several identical IDs, or returns a matching name attribute instead of returning false when no matching ID is found)

    And then, about Microsoft websites…

    – loading and logging into the Mozilla-optimized version of Live Mail (Hotmail) triggers a couple hundred warnings, and bing.com (hardly a complex page) triggers 22; performing an (X)HTML validation on it triggers 12 errors – a lot, if you consider that it’s supposed to be XHTML, and that if it were served with its proper MIMEtype, most XHTML-capable browser would clam up and refuse to load it because it’s invalid!

    And, although I understand that this doesn’t depend on the IE dev team, I do also find the marketing campaign around IE 8 downright _despicable_ and _revolting_, which makes this rather good release (it IS miles above IE 6 and a great improvement over IE 7) quite sour-tasting.

    Mitch 74

  15. domenic says:

    My comment is this:

    I would like IE9 to have 100% full support for setting/getting the .innerHTML of EVERY SINGLE HTML element.

    For the getting part we want VALID markup returned. No UPPERCASE tags, No DBl-QUOTE-less attributes, No MISSING closing tags or closing slashes.

    For the setting part it should work on EVERY element. This includes the Head, Tables, Selects, Divs, Pres, Objects, etc.

    Slapping "readonly" on certain HTML elements because your implementation failed to address them is NOT a resolution.  Furthermore the Wizard of Oz hiding behind the curtain is fully exposed on this because we all know full well these elements are NOT readonly, we can set all of their content using W3C DOM methods.

    Note using the DOM Methods isn’t a fully acceptable workaround as it takes longer to process that setting the .innerHTML and as I’m sure you are fully aware IE’s support of .setAttribute() in the past (pre IE8 Standards Mode) is notoriously very badly broken.

    Sorry if this comes off like a rant, it isn’t meant to be.  However I think it is high time we start addressing some of the legacy bugs in IE that really honestly need to be fixed if IE plans to be taken seriously as a Web Browser of the future.

  16. Anand Kumria says:

    Hi,

    What is the engineering point of view when it comes to Video codes?

    Is one, or the other, technically harder to implement?

    What’s your take on existing ActiveX controls that simulate the <video> tag?

    Would you actually implement this within the browser, or using COM or deferring to Windows Media Player?

    Thanks,

    Anand

  17. @cpradio: in addition to the availability of the IEAK as Tom mentioned, we will also make Internet Explorer 8 available via WSUS (Windows Server Update Services) starting on August 25th.  See my post from 6/29 for some additional details:

    http://blogs.msdn.com/ie/archive/2009/06/29/it-professionals-prepare-for-internet-explorer-8-availability-via-windows-server-update-services-wsus-in-august-2009.aspx

  18. broader topics says:

    why yes, broader topics for IE engineering like:

    Why can’t I drag links in IE to all the places I want/used to be able to in IE6?

    Why does bookmarking a bookmarklet fail silently in IE?

    Why is organizing my favorites in IE so mind bogglingly complicated? Better yet why wasn’t this fixed in the last 2 versions of IE?

    Why is the IE7/IE8 UI so horribly un-customizable?  Why can’t I hide the tab bar if only 1 tab is open? Why does the command bar flow backwards? Why does dragging the favorites bar to the tab bar throw everything out of place? Why can’t I close the last tab in IE? Why are bookmarklets severely truncated to the point that they are just not worth making for IE? Why did it take a whole version of IE to get an option to move the reload button back to the left of the address bar?  Why do I have no control over the unreadable address bar "In the almighty name of security"? Why can’t I resize the addressbar/search box? Why does the addressbar dropdown portion suck so horribly? Where did my favicons go? Why does the find as you type searching not work like users expect it to?

    Where is the built-in spell checking in IE7 or IE8?

    Where is the download manager?

    When will the JScript^H^H^H^H^H^H^HJavaScript error console provide something more meaningful than "Unknown Thing doesn’t support Unnamed Property or Method" – I’m sure there is a possibility that there could be a more useless message here but I have to admit you would need to try pretty hard!

    When will the JScript^H^H^H^H^H^H^HJavaScript error console provide a clear button?

    etc, etc, etc.

    The IE team is well aware that the community is not happy with the IE UI and the IE usability.

    What ticks us off is the complete lack of acknowledgment that a complete re-vamp has been needed for 8 years… and has yet to be announced.

  19. Jatinder Mann [MSFT] says:

    @cpradio: For more information on the IEAK8 and also a case study of customizing, managing and deploying IE8, please refer to the following blog posts:

    http://blogs.msdn.com/ie/archive/2008/09/22/the-ie8-ieak.aspx

    http://blogs.msdn.com/ie/archive/2009/03/18/improving-the-education-experience-with-internet-explorer-8.aspx

    Thanks,

    Jatinder

  20. @ broader topics

    > Why is the IE7/IE8 UI so horribly un-customizable?  Why can’t I hide the tab bar if only 1 tab is open? Why does the command bar flow backwards? Why does dragging the favorites bar to the tab bar throw everything out of place? Why can’t I close the last tab in IE?

    (…)…lots of questions…(…)

    > The IE team is well aware that the

    > community is not happy with the IE

    > UI and the IE usability.

    Broader topics,

    IE Team closed bug 363573: Poor Interface

    connect.microsoft.com/IE/feedback/ViewFeedback.aspx?FeedbackID=363573

    and closed bug 331901: User interface = poorly designed

    connect.microsoft.com/IE/feedback/ViewFeedback.aspx?FeedbackID=331901

    Ability to customize the UI has been covered by some parts of UAAG technical recommendation. In any case, it must be said that

    bug 460283 : Reorganize IE’s user interface layout

    connect.microsoft.com/IE/feedback/ViewFeedback.aspx?FeedbackID=460283

    is still active.

    regards, Gérard

  21. Geld lenen says:

    That post from jatinder was very helpfull.

  22. Anonymous says:

    I think there’s some irony in that Broader Topics and Fred are complaining about opposite things (one wants a complete UI rehaul, the other wants no UI changes at all).

  23. James Chen says:

    IE 8 doesn’t work on my XP computer, and now I can’t reinstall an earlier version of IE.  Why do you keep making poorly-performing software?

  24. Matt says:

    James, rather than whining helplessly, why not explain specifically how "IE8 doesn’t work" on your computer?  Since it clearly works for 10s of millions of people, I’m sure there’s a way to make it work for you (you probably have a buggy browser addon, easily fixed).

    When you uninstall IE8 using the control panel, your old IE is automatically restored.