The Problem with Internet Explorer

T

Internet ExplorerAfter watching the second keynote from PDC last week, I’m now puzzled by Microsoft. In the first keynote, they seemed to get everything right. Embrace open source. Embrace cloud. Embrace what their customers want and need. Embrace strategic change

In the second keynote they seemed to do the opposite. Steven Sinofsky, in a somewhat self-congratulatory ‘review’ of the ‘lessons learned’ developing Windows 7  bored us to tears with big number telemetry stats and other ‘look at how responsive we are’ reminders, before moving on to Internet Explorer.

And whilst I was anticipating an exciting announcement about Microsoft finally getting Internet Explorer right (especially after reading reviews here, here and here), instead I was left shaking my head about the priorities that have been set.

This post from the IE blog contains most of what Steven presented. Again web standards seemed to be a focus. Guys, get over web standards please. They should just be a given. They aren’t a feature any more. And then rendering performance improvements. Sure, some animated fonts are smoother, but seriously who cares? I haven’t heard anyone ever complain about how they wished that their browser would render some special font effect better. Admittedly though, better panning of maps would be a nice improvement.  And then on to other ‘features’ of extremely limited appeal like rounded corners. Sure, there’s a very vocal (but tiny) group of people who want this, but it isn’t what the majority wants (or is even aware of!). Nor is JavaScript performance. Whilst I’m sure graphs like the following (from Sinofsky’s talk) are based on some legitimate tests being conducted, in reality I’ve never noticed much difference between IE and FireFox:

JavaScript performance in browsers

Instead I’ve noticed that Chrome seems to be a lot faster, but most of that perception is simply because it starts faster when I click it on my start bar. I’m up and browsing in an instant, not waiting a few seconds like I do with Firefox and IE.

But for argument’s sake let’s assume that JavaScript performance is a key differentiator in the browsers. In that case the improvements in IE9 should just be released as a simple patch for IE8. They shouldn’t be heralded as a reason to look forward to IE9. It’s really lame when an 8th version product tells you that a feature of the 9th version is that it is going to catch up with its competitors. Especially when there’s no release date for IE9 even being mentioned yet. Sinofsky mentioned in his keynote (around the 44:09 mark) that the above performance improvements were a result of development progress just 3 weeks into the project, so I’m assuming it isn’t that big a change…

IE9 has almost no appeal to me based on Steven’s demo. Instead, what was I hoping to see?
It’s simple: Extensibility.

The Internet Explorer team have wasted the last 2 versions (IE7 and IE8) building in accelerators, web slices and other useless guff that hardly anyone uses, and then capped it off with a ridiculous web compatibility mode that sent browser innovation back a decade.

Instead they should have learned from Firefox. Make extensibility easy. Mozilla’s available Add-ons for Firefox now run into the tens of thousands. And downloads in the billions. Yes, billions. It’s the reason why Firefox continues to grow in popularity, especially amongst developers (the very crowd that Sinofsky was presenting to at PDC). Microsoft should be providing (or at least promoting) a powerful extensibility initiative for IE.

Sadly though Microsoft seems to have its head in the sand when it comes to IE. Instead of being an IE centrepiece, the IE Add-ons site feels like a we-ran-out-of-budget after-thought. It’s a ghost town in there… It’s pretty telling when locally the most popular Add-ons for IE are based on 2 ratings, and are mostly simple web slices. And perhaps even a little tragic when one of the most popular Add-ons is an accelerator to enable basic Google Australia searches in IE.

One of the reasons IE has such a big market share is because in the enterprise space it is unbeatable in terms of deployment flexibility. IT Managers can control almost every aspect of its install, functionality and branding. IT Pros have been well catered for by IE. Now it’s time to cater to the developers.

Apps are the key with Mobile (in spite of what Ray Ozzie might or might not have said), and they’re the key with browsers. If Microsoft wants IE to be a cornerstone of their ‘3 Screens and a Cloud’ strategy (unveiled by Ray in the first keynote), then they need to build a vibrant developer ecosystem around IE, and not mistakenly think that Silverlight controls will be a viable substitute. The App model is important. Thankfully Microsoft has realised it with SharePoint, and they’ve even realised it with data. So, why can’t they realise it with IE?

20 comments

  • Strongly agree with you. IE has been on a wrong approach for years. I gave up using IE since 1997 (IE4), and I have never regretted about it.

    Just a few additional comments to yours.

    1. The biggest security deficiency, ActiveX, is still with IE and treated as a major approach to develop add-ons. It is a burden for MS now. You can’t throw it away as it has a large user base. You can’t say No to it as it is by design. You can’t upgrade it as it would cause more weakness. You can’t stop upgrading it as its vulnerabilities are critical and disclosed. Manually disabling ActiveX has been proven an efficient way to reduce the surface of attack for IE.

    2. ActiveX is also the biggest bottleneck to cross-platform compatibility, which makes ActiveX components totally useless when users are working on a non-Windows platform, because ActiveX controls are natively supported by Windows only and binary-compatible on x86 hardware only.

    3. Security Zone (Web Content Zone), a key control protecting IE from malicious attack, is actually based on a kind of word games (determining security zone according to predefined string patterns), which can be easily compromised by a malicious DNS server. This unintelligent mechanism seems to be a permanent functionality in IE, though MS has added much more complex rules to make it look better. It now becomes a place to tweak IE if someone wants to look professional. :-))

    4. Some non-standard extensions or improvements (in MS wording) in HTML and user experience, including font rending as you mentioned, are really ridiculous. A serious developer won’t use or depend on such features as it would limit the scope of their audience and cause compatibility issues. The same thing from the point of view of security.

    However, sometimes I have to use IE for a little while, just for updating Windows or accessing MS sites when the pages are not friendly to FF (MS’s fault in compatibility and user experience).

    Cheers,
    Bing

    • Thanks for your comments Bing.
      The frustrating thing is that extensibility should be really easy – there’s no need for it to be developed around ActiveX. Firefox, for example, is much easier. You need to understand a little Javascript and XML, and can basically build an extension after a short learning curve.
      Microsoft have been trying to simplify things a little, and also enabling Silverlight based toolbars etc, but the learning curve is much higher. They really need to provide clear, simple advice about how to extend IE and make it a priority to promote it. For example, the new Developer Tools in IE8 (F12) should have been built as an extension, open sourced, and allowed to be further extended by developers.
      And as you say, the cross-platform support isn’t there either – but I suspect this isn’t a priority for the IE team (and I’m fine with that actually – focus on your biggest market, the Windows users).
      Like you I still use IE from time to time (eg on MSDN). In fact, I actively try to return to it (given that I’m a MS fan), but each time I miss all my Firefox plugins too much and switch back to FF.

      • Yup, indeed.

        Basically, IE should be ideally for people living in the heaven at where people are all nice, no one does evil, and they have plenty of time to kill. People in the heaven all enjoy any AUTORUN design, and they do any thing that Microsoft wants them to do, from answering Yes for activating something to installing a 300MB runtime package just for a 50KB utility based on .NET, and to reading through the bulky MSDN for hours before writing a piece of code to patch up IE a bit.

        In the hell, where no one trusts each other, people only exchange messages in plain text. No codes. Nothing dynamic. Every thing is manual. They don’t use browsers, they do gossip.

        On the earth, people better use FF or something like that for living in a diverse, connected but conditionally trusted world. It is not perfect. It just works, across various platforms, and as well as its add-ons. The key is not what FF can do for you; it is about how it is built. A simple, efficient and open architecture defines why and how it grows up. Additionally, for IE, Microsoft is the world; for FF, the community is the world.

        Anyway, I am talking about IE and its host only, not the whole Microsoft and its product lines. I am also a MS fan. ;-)

  • Strongly agree with you. IE has been on a wrong approach for years. I gave up using IE since 1997 (IE4), and I have never regretted about it.

    Just a few additional comments to yours.

    1. The biggest security deficiency, ActiveX, is still with IE and treated as a major approach to develop add-ons. It is a burden for MS now. You can’t throw it away as it has a large user base. You can’t say No to it as it is by design. You can’t upgrade it as it would cause more weakness. You can’t stop upgrading it as its vulnerabilities are critical and disclosed. Manually disabling ActiveX has been proven an efficient way to reduce the surface of attack for IE.

    2. ActiveX is also the biggest bottleneck to cross-platform compatibility, which makes ActiveX components totally useless when users are working on a non-Windows platform, because ActiveX controls are natively supported by Windows only and binary-compatible on x86 hardware only.

    3. Security Zone (Web Content Zone), a key control protecting IE from malicious attack, is actually based on a kind of word games (determining security zone according to predefined string patterns), which can be easily compromised by a malicious DNS server. This unintelligent mechanism seems to be a permanent functionality in IE, though MS has added much more complex rules to make it look better. It now becomes a place to tweak IE if someone wants to look professional. :-))

    4. Some non-standard extensions or improvements (in MS wording) in HTML and user experience, including font rending as you mentioned, are really ridiculous. A serious developer won’t use or depend on such features as it would limit the scope of their audience and cause compatibility issues. The same thing from the point of view of security.

    However, sometimes I have to use IE for a little while, just for updating Windows or accessing MS sites when the pages are not friendly to FF (MS’s fault in compatibility and user experience).

    Cheers,
    Bing

    • Thanks for your comments Bing.
      The frustrating thing is that extensibility should be really easy – there’s no need for it to be developed around ActiveX. Firefox, for example, is much easier. You need to understand a little Javascript and XML, and can basically build an extension after a short learning curve.
      Microsoft have been trying to simplify things a little, and also enabling Silverlight based toolbars etc, but the learning curve is much higher. They really need to provide clear, simple advice about how to extend IE and make it a priority to promote it. For example, the new Developer Tools in IE8 (F12) should have been built as an extension, open sourced, and allowed to be further extended by developers.
      And as you say, the cross-platform support isn’t there either – but I suspect this isn’t a priority for the IE team (and I’m fine with that actually – focus on your biggest market, the Windows users).
      Like you I still use IE from time to time (eg on MSDN). In fact, I actively try to return to it (given that I’m a MS fan), but each time I miss all my Firefox plugins too much and switch back to FF.

      • Yup, indeed.

        Basically, IE should be ideally for people living in the heaven at where people are all nice, no one does evil, and they have plenty of time to kill. People in the heaven all enjoy any AUTORUN design, and they do any thing that Microsoft wants them to do, from answering Yes for activating something to installing a 300MB runtime package just for a 50KB utility based on .NET, and to reading through the bulky MSDN for hours before writing a piece of code to patch up IE a bit.

        In the hell, where no one trusts each other, people only exchange messages in plain text. No codes. Nothing dynamic. Every thing is manual. They don’t use browsers, they do gossip.

        On the earth, people better use FF or something like that for living in a diverse, connected but conditionally trusted world. It is not perfect. It just works, across various platforms, and as well as its add-ons. The key is not what FF can do for you; it is about how it is built. A simple, efficient and open architecture defines why and how it grows up. Additionally, for IE, Microsoft is the world; for FF, the community is the world.

        Anyway, I am talking about IE and its host only, not the whole Microsoft and its product lines. I am also a MS fan. ;-)

  • 1
  • 2
By Craig Bailey

Archives