“But IE 9 Is Just Around The Corner…”

The single most frustrating thing for me as a web developer is the incredible disconnect between day-to-day development and the shiny, shiny stuff showing up in HTML5 and modern browsers. It’s made all the more frustrating by bold pronouncements from any (every?) vendor about how much more awesome the web will be thanks to the shiny stuff their upcoming release. The reality for web developers is that those features won’t matter on a relevant time-scale. Not your next project. Not your next 5 projects. No, the lag today between new features and when you can use them might as well be measured in geologic time.

How bad is it? The next time you read some tech journalist write about how some new browser version is just around the corner and how it’ll make everything better, remember that:

  • 50+% of Windows users are on XP a year after Windows 7 shipped and 3.5 years since Vista shipped
  • Windows XP will be supported until 2014, giving organizations on XP extra breathing room to limp along on IE 6-8
  • There will be no IE 9 for Windows XP
  • After something like 4 years of MSFT urging customers in the strongest possible language to get off of IE 6, it still has 16% of the market, and it’s not falling nearly fast enough. At the current 0.75%/month dropoff, we’re looking at 20+ more months of IE 6. At this rate, kids born today will be walking and maybe even talking by the time we can write IE 6 into the history books.

The goal here isn’t to drive you to drink, but to call out the dichotomy between when users get features and when developers can address features. For users, things tend to get better as soon as you pick up a new browser. Developers have to wait until every user makes that sort of choice. Said differently, users can benefit from the bleeding edge but developers are beholden to the late adopters.

We should cut tech journalists and users a little slack, though; in a world without adoption friction, the launch of a new feature would translate directly into the sort of “now you can build sites and apps with awesome thing X”. The instinct to believe that’s how it should be is spot on. But that’s observably not how the world works. Don’t believe the hype. New browsers alone haven’t fixed the problem so far, so what makes us think they will in future? No vendor wants to talk about their last version, but for web developers stuck in the trenches, that all there is to talk about. That’s where the pain is, after all, and counting on browser upgrades to fix the problem quickly isn’t working well enough. Chrome’s aggressive auto-update feature is changing the way things will work in future, but for now we’re still stuck in the slow-upgrade dynamic.

We need a Plan B.


  1. sethaurus
    Posted September 24, 2010 at 12:37 pm | Permalink

    The exception here is mobile. You can all but rely on a WebKit variant or a similarly advanced engine, and the market sector with the greatest volume of mobile web usage is also the sector with the lowest adoption friction, due to frequently-pushed OS updates.

  2. Posted September 24, 2010 at 12:40 pm | Permalink

    One of the biggest disappointment’s that I have with HTML5, is that for the things it brings which web-plugins already do ( in some cases better ) if I want to use HTML5, I’ll be doubling up my workload by having to implement a fallback for everyone still using IE 6/7/8, Firefox 2/3, Safari 3/4, and early versions of Chrome.

    Sure my HTML5 project will be cutting edge, and impress college students, but my clients who are predominantly corporate users unable to “simply update their browser” will be left the in cold unless I do twice the work by creating an HTML5 version, and a Flash fallback. So yeah, that also contributes to why I’ll be way more excited about HTML5 in a decade.

  3. Posted September 24, 2010 at 3:02 pm | Permalink


    You’re right about mobile. All technologies start out in “defensible spaces”, markets where new capabilties can be used without needing to worry about re-work. Mobile is that space for HTML5 right now. GCF is letting us break out into the desktop.


    First, there is no such thing as “early versions of Chrome”. Thanks to the aggressive auto-update, statistically speaking they don’t exist. Other browsers with auto-update systems enjoy similar (if more muted) behavior. Safari 2 and 3 are mostly gone, for instance, as is FF 2.

    That leaves us with IE. Chrome Frame is an answer that helps us in those corporate situations by giving us a way to have HTML5 co-exist with legacy apps in a way that doesn’t cost organizations more. They can make an orderly transition to HTML5, continue to support legacy apps until it makes economic sense not to, and not worry about the costs of browser upgrades for the sake of a single app. Win.

  4. Daniel K
    Posted September 24, 2010 at 7:31 pm | Permalink

    The problem is that for a vast majority of people, not necessarily in the corporate box, “upgrading” from Windows XP to Windows Vista or Windows 7 is tedious and unnecessary. People have better things to do than change what is working for them, with the likelihood they’ll eat up hours or days trying to learn a new environment, or getting applications installed and working as they used to or whatever.

    If that makes developers suffer, well boohoo. Shiny may be what drives some people, but a greater number aren’t so caught up in such things.

    People still use IE 6 because IE 7 and 8 didn’t make a compelling case to upgrade to. Seriously. I find that despite our bellyaching, IE 6 works well enough and does so without wasting any time installing 7 or 8.

    Windows 7 and Office 2007 and 2010 don’t make a person’s life better if they are using XP and Office 2002 or 2003. Seriously. All they do is shuffle the deck so you’re searching around for stuff because Microsoft in their infinite wisdom thought they knew better where to stash things such as menu options and toolbar buttons.

    That’s the thing about “progress” – it doesn’t always benefit everyone. That’s reality, and if that means that HTML 5 will roll out slower, so be it. They’ll still be reasons and environments to deploy on and to, and if someone is using XP and IE 6, then don’t worry about them because they’re probably doing quite fine without all the new shiny stuff coming out.

  5. Posted September 25, 2010 at 8:07 am | Permalink

    –“For users, things tend to get better as soon as you pick up a new browser. Developers have to wait until every user makes that sort of choice.”

    I couldn’t disagree more. We can build things today that users don’t all get to “experience”. I don’t mean that we exclude users on IE6 or 7 or 8 from accessing the content. They just get the less shiny, less new, less cool version.

    I’m currently experiencing this kind of situation outside of the www realm. TV stations are broadcasting shows that make use of wide screen TV resolution, while I still have a TV that is pretty much square shaped. I haven’t upgraded because my current TV still works. The plan is, when my TV ceases to work, I’ll get a new one, but for now I just have to go on missing what’s on the sides that everyone else gets to see. :)

    Treat the web the same way. Just because users aren’t upgrading, doesn’t mean developers can’t. We just need to be realistic that if we are excluding people from accessing content that is necessary on the site or app, we aren’t using the new, shiny stuff responsibly.

  6. Posted September 25, 2010 at 10:43 am | Permalink

    Hey Bridget,

    I didn’t mean to suggest that motivated developers can’t or don’t create fallbacks or intentionally exclude some users. I was instead arguing that the economics of the web make that rare. It’s not usually what happens, and if you didn’t have to make a hard choice about supporting down rev browsers, I’m guessing you wouldn’t.

    The same dynamic is actually in play in TV-land: the rework and extra constraints imposed by doing two versions of things is causing a lot of lowest-common-denominator use of that extra screen real estate since producers understand that some large fraction of their audience won’t see that content. Put another way, that part of the screen isn’t getting the same kind of investment as the center is, and it shows. It won’t be until the wide-screen format (and analagously, HTML5) can be counted on that we should expect those new features to really open up the creative and user-experience benefits that they should be delivering.

    Some developers will get on the HTML5 thing faster than others without worrying about down-rev users and will deliver better experiences because of it. Others will constrain themselves to whatever subset is portable or will continue to patch over it with JS libraries and the like. In any case, everyone wants the new thing tomorrow and if they’re using the new features they’re paying some extra cost to do so, either in smaller addressable market or in reduced performance/richness.

    You’re making an explicit economic choice which I applaud. My interest here is in giving you back some large part of that market so that those who follow you don’t have to make a hard choice: the web should be awesome for everyone.


  7. Laura-
    Posted September 25, 2010 at 3:32 pm | Permalink

    I know you think Chrome Frame is the solution, and I admire your passion.

    But at what point do you realize nobody’s using Chrome Frame, IT organizations (and many corporate users) don’t care about cutting edge features… and that it’s time to move on to something else?

    I wish this was a solution, but it’s clearly not. There are many more corners of the web that could use your brainpower.

  8. Bob Haugen
    Posted September 25, 2010 at 5:30 pm | Permalink

    “At this rate, kids born today will be walking and maybe even talking by the time we can write IE 6 into the history books.”

    Wow! So in a year or two, we can kiss IE6 goodbye? That’s good news! ;)

    (Or maybe that was not what you meant?…)

  9. Posted September 25, 2010 at 5:48 pm | Permalink

    Hi Laura,

    Do you have a better idea? You haven’t refuted the analysis in any way, so I’m sort of unmoved. To answer the question: I’ll give up when someone convinces me there’s a better plan that’ll help us move the needle faster. So far nobody has.

    Anyhow, just wait a little. Chrome Frame JUST went to stable. Now the real fun begins. Even if you’ve preemptively given up, I hope you’ll forgive my lack of despair. The web is about to get better at a rate we haven’t seen in most of a decade. Remember: the dynamics of this plan are completely different than for any other folks have tried so far. We don’t have to “win” by being installed everywhere, we just have to give developers an excuse not to develop sites for bad browsers.

    The logjam today looks like this: web developers see a big IE population and go “hrm, I want to do something better, but I can’t”. Companies look at the apps they can buy and go “well, there’s nothing that’s so awesome and cost effective that it’s worth a browser upgrade”. The key here is in changing the expectations of either party. How do we do that? By giving webdevs a way to give themselves permission to build the awesome stuff that’s going to cause companies to pony up for upgrades. How? By getting GCF installed in a lot of places…something that Google has a strong interest in doing. Just think about how much money gets spent to make things like GMail work on IE6. It’s mind boggling.

    This is a strategy that speaks directly to the “but I can’t use HTML5 yet” conundrum that most webdevs (myself included) sense palpably. This isn’t about any one browser, this is about affecting rates of change. The next 12 months are gonna be awesome. I can’t wait.


  10. Posted September 25, 2010 at 6:31 pm | Permalink

    2 years is waaaay too long to wait. And after IE6, we’ll be stuck with IE8 for another half decade. The problem here is structural, and we need a structural solution. Luckily we have one on the horizon…so *that’s* the encouraging news from my perspective.

  11. Justin
    Posted September 25, 2010 at 10:14 pm | Permalink

    I could google this if I weren’t lazy, but I wonder if people will be more likely to upgrade ie 7 & 8 than they were 6. My hope is that people are developing their new apps with less ie specific code. I’ve seen this happen around a fee fortune 500 it shops, but there’s a long tail.

  12. Posted September 26, 2010 at 4:37 pm | Permalink

    I disagree with you partially.

    We’ve been hearing about the “browser acting as the OS” for years, and now its becoming reality. SQL-like client side data storage, graphics and effects rendered through the hardware. If any of this gains traction the web browser will be just as powerful as the OS and its a brave new world. This will mean a new sort of developer–programming advanced graphics is generally needs a different sort of skill set than setting up a decent CMS.

    But that’s not to say traditional web development will go away, I don’t see Amazon or Ebay deciding to completely throw out their current site designs. A major chunk of the IE6 users tend to be from corporate or government agencies who dumped a ton of money into installing and customizing IE6 years ago, and have no appetite for making a similar investment. Therefore, I leave it up to the clients whether they’re willing to support these features.

  13. Posted September 26, 2010 at 7:17 pm | Permalink

    Alex, you know I mostly agree with you on the upgrade cycle problem – although I’d point out on the desktop, the upgrade cycle slowness is on IE, not on the other browsers. Although Chrome is most aggressive, Safari and Firefox users migrate to new versions quite quickly; IE users don’t, by comparison.

    However, “some variant of WebKit” on mobile isn’t really good enough – very quickly, a year-and-a-half-old branch build of WebKit won’t be a great experience either. Although, of course, my past experience leaves me with some empathy with the compatibility problems auto-updating causes, moving the web forward is going to require being aggressive here.

  14. Posted September 26, 2010 at 11:59 pm | Permalink

    Hey Chris,

    I’m curious to know if you think the IE upgrade slowness is due to scale (60% of everyone is a lot of people) and how much is related to the upgrade dynamics of IE specifically (tied to OS revs, not on aggressive schedule, etc.). I can easily see it going both ways.

    Anyhow, I’ll be first in line to kvetch about how slow mobile phones upgrade once they’re the ones holding us back =)

  15. Sean
    Posted September 27, 2010 at 7:35 am | Permalink

    You’re assuming that the IE6 dropoff rate is linear. It’s actually exponential. I took a random sample of all my client’s sites and IE6 usage is around 5.5%. That’s good enough for me to stop supporting it.

  16. Posted September 28, 2010 at 7:22 am | Permalink

    hey Sean,

    Out of curiosity: do you support Safari?

    If you look at the global rates for the last couple of years, it’s SUB-linear. That is to say, the rate of decrease is slowing. I was being generous in my 20 month estimate since I know that it’s highly geography and site dependent.

    Anyhow, lets say IE 6 were gone tomorrow…that leaves you with IE 7. I don’t know about you, but being able to drop just 6 and not 7 wouldn’t change anything I do (save, perhaps, the use of PNGs).