Skip to: Navigation | Content | Sidebar | Footer

Weblog Entry


July 27, 2006

Here we go. According to CNET, Internet Explorer 7 is going to be sent out automatically to XP users as a priority Windows Update upon its launch. We’ll see a massive migration, and IE6 numbers will take a sharp dive in the very near future. This is a good thing.

However, a sharp surge of immediate adoption could be a little painful for those not currently testing in IE7; now is most definitely the time to start, if that’s you. (Hey, don’t feel bad, I still only have IE7 beta 1 myself.)

Also, as of the launch (no set date yet, some time in Q4 is the current best guess), I have a strong feeling the case for testing in IE5.x will be unjustifiably hard to make in most cases. Enjoy your last few months of the box model hack.

Josh Delsman says:
July 27, 10h

Hooray! Thank god.. but..

What about Windows 2000 users? They will most likely still be behind on IE6, which isn’t exactly good news. I wonder what Microsoft will plan to do about the non-profit organizations, corporate users, etc.

Ismael says:
July 27, 10h

OMG These are happy, happy news!
Just this week I’ve been strugling to make IE6 support position:fixed and correct z-indexes.

Zach says:
July 27, 11h

I have this beautiful image of myself dropping IE6 support in my mind already.

Will IE7 work on Win2k, though?

Ingo says:
July 27, 11h

… but z-index is still broken, Ismael
Anyway. Hooray.

July 27, 11h

This is good news, but MS are also releasing a tool to block the update:

“Additionally, Microsoft on Wednesday plans to make available a special tool to block automatic delivery of the new browser version, Schare said. The tool is meant for business users who might not be ready for an IE update.”
ZDNet -

Lets hope not many businesses use it.

July 27, 11h

Zach - IE7 is a WinXP SP2 and above only update I’m afraid. So everyone who’s using something below that now will still be using what they’re currently using once IE7 comes out. :/

On the most non-techy site I run, people running IE6 in < XP accounts for 7% of visits. Not exactly insubstantial. Thankfully though, IE5.5 are virtually non-existent (for me).

web says:
July 27, 11h

It’s like the apocalypse for all poorly written CSS.

Except without the kick-ass fire and brimstone falling from the sky.

Mau says:
July 27, 11h

=) go to to get your browser. =)

I wonder if they’ll ever tempt Mac users to install IE7. I guess not since they completely abandoned IE builds for Mac.

Andy Budd says:
July 27, 11h

That is really great news. However I do worry that we’ll end up in a situation where all those on IE6 will upgrade to IE7, but all those still stuck in IE5 will still be stuck in IE5, thus widening the gap even more. I’ve been thinking about this for a while, but when IE7 launches, maybe it would be time for another browser upgrade campaign?

July 27, 12h

Ugh, I’m on a Mac and I use Win2k in Virtual PC for IE 6 testing. I guess this means that I have to upgrade to Windows Vista when it comes out just so that I can test in IE7…

Is IE7 set to launch before Vista, or are they planning to release them together?

Dan Rubin says:
July 27, 12h

A bunch of us were talking about this very issue last week at WebVisions, and the general wish was that MS would do what they’ve just announced.

We’re past the point of another official browser upgrade campaign: what we need to do now is stop supporting IE versions earlier than 6 (and by “stop supporting” I mean give them the content, but not the experience – just give pre-IE6 versions an unstyled page, or at most a minimal style that doesn’t require much effort to develop or maintain). I’ll even argue that it’s time to start thinking about scaling back support for IE6 (I’m giving that serious consideration already with some projects), and focusing on IE7 as the soon-to-be-current version.

Though a tiny little notice at the top of sites visited by IE6 users, encouraging them to either upgrade to IE7 or use Firefox, wouldn’t be a bad idea once IE7 ships… :P

Nick Husher says:
July 27, 12h

Dan, I think a better idea would be to serve a styled page, but with a (fairly) prominent option to turn off the stylesheeting. It might be a desirable option for standards-compliant browsers, but primarily because it offers older browsers or ones that that you don’t explicitly support a fair swing at rendering to the standard. If they can’t hack it, they have a usable fallback.

My biggest gripe at the moment is ‘default deny’ on rendering pages rather than ‘default permit’. For instance, Opera get’s a ‘ur browser is teh suck’ at without even a fair shot at using the page. I’m 99% sure that Opera can hack it, but because my favorite browser isn’t on their list, I’m not allowed to visit their site. I can’t say I’m too broken up, though.

There’s a big difference between ‘we dont’ support…’ and ‘we don’t allow…’. Be a ‘we don’t support’er rather than a ‘we don’t allow’er.

Nick Toye says:
July 27, 13h

I have just installed IE7 Beta 3, and it has caused me problems already. My site is showing errors similar to Opera and Safari. So does this mean Firefox is behind?

I have also noticed that Dan Cederholm’s SimpleBits has some graphical glitches in it too. This is not to say that Dan has suddenly lost the touch, but it sets bells ringing for me that IE7 Beta 3 is still not ideal.

Dan Rubin says:
July 27, 14h

@ Nick: I’m not advocating a ‘default deny’ approach, just giving older browsers a lower-end experience (similar to how standards-based developers approached NS4 for a long time). I’d rather display a clearly scaled-down version than something that almost works, or might even break. Again, IE versions prior to 6 are my target in this case, and unless a site’s audience uses IE5/5.5 in large numbers, it shouldn’t cause problems or discouragement.

I like your idea of allowing IE6 users (for instance) an easy way to turn off styles, so if something doesn’t quite work they can at least get rid of it. However, I still believe that as developers, we will be in a unique position to influence the browser market when IE7 ships – perhaps we won’t be able to get everyone running Firefox, but in most cases, IE7 will be a huge improvement over what we have to deal with today.

July 27, 15h

We’re still getting visitors with versions of IE that are less than 6. Here’s a month to date list of just Microsoft browsers:

Browsers = Visiters
Microsoft Internet Explorer 3.02 = 1

Microsoft Internet Explorer 4.01 = 29
Microsoft Internet Explorer 4.01 (AOL) = 17
Microsoft Internet Explorer 4.5 (Macintosh) = 1
Sub Total: 47

Microsoft Internet Explorer 5.0 = 1,210
Microsoft Internet Explorer 5.0 (AOL) = 117
Microsoft Internet Explorer 5.0 beta 2 = 1
Microsoft Internet Explorer 5.01 = 1,624
Microsoft Internet Explorer 5.01 (AOL) = 18
Microsoft Internet Explorer 5.22 (Macintosh) = 779
Microsoft Internet Explorer 5.23 (Macintosh) = 2,095
Microsoft Internet Explorer 5.5 = 8,811
Microsoft Internet Explorer 5.5 (AOL) = 1,624
Sub Total: 16,279 = 0.82%

Microsoft Internet Explorer 6.0 = 1,576,173
Microsoft Internet Explorer 6.0 (AOL) = 116,334
Microsoft Internet Explorer 6.0 beta = 59
Sub Total: 1,692,566 = 85.61%

Microsoft Internet Explorer 7.0 beta = 95

Microsoft MSN Explorer = 33,082
Microsoft MSN Explorer 2.5 = 988
Microsoft MSN Explorer 2.6 = 66
Microsoft MSN Explorer 6.1 = 5,502
Microsoft MSN Explorer 8.0 = 700
Microsoft MSN Internet Access = 1,110
Sub Total: 41,448 = 2.09%

I’m happy to say that Firefox has increased its share at 154,943 = 7.84%. We have more Safari visitors at 2% of visitors than IE 5.x at 0.82%, but is it worth turning away 16,000 plus potential customers?

I suppose that we could use conditional comments for IE 5.x browsers that would display a “Please upgrade” or “Site best viewed with browser x” message.

Ugh! Who are we kidding? When do you make a decision to not support a specific version of a browser?

manuel says:
July 27, 16h

I second Andy Budd. Saying that “the case for testing in IE5.x will be unjustifiably hard to make in most cases” is ridiculous.

Users are *stuck* on IE5, they won’t go from 5 to 7. Otherwise they would’ve switched to IE 6 already.

Mau says:
July 27, 16h

To John F: Microsoft first said that they were not going to release any upgrade from IE6 to IE7 before Vista was released… That they were going to release them together.

Since Vista is taking forever to be released, I guess they finally gave up and probably Firefox built more pressure by announcing their 2.0 version.

So, seems to be that they are going to release IE7 first, and some time in 2007… maybe 2008 they will release Vista… that probably gives them some time to patch the first security holes once it gets released. So Vista will sport an IE7.2 or something.

You know how that goes.

Andy Beeching says:
July 28, 02h

I also disagree with the statement, “the case for testing in IE5.x will be unjustifiably hard to make in most cases.”

Having done work in the voluntary sector, I noticed many of our clients there were charities using earlier versions of windows (with refurbished equipment that doesn’t support XP). The majority of employees were still using IE 5.x and like Andy Budd mentioned, will have no way to upgrade to the latest version. (I am under the impression IE7 will only be available as an upgrade to XP SP2 users?)

It wasn’t even an option to drop support for these browsers, as like any client they wanted the design they signed off on to work perfectly in the browser of their choice. So will web developers keep locking out these types of visitor to sites they create in the future? I think it probably comes down to the question of target audience and/or the content that will be featured on the site.

To be honest I’ve found that it really doesn’t take that much longer to produce a site for IE5.x if your already taking IE6 into account as well…

July 28, 03h

On one of my sites targeted at non-techies, IE6 on WinXP is just over 50% of visitors, but on pre-XP it adds up to about 9%, roughly the same as Firefox.

Considering all of those IE6 users aren’t going to upgrade to IE7 for various reasons (i.e. they’re not on SP2 as they have a bootleg, or they don’t have automatic updates turned on) then I still think there will be enough IE6 users around to have to take them in to account. I’d bet IE6’ll stay larger than Firefox’s market share for at least six months, and a heck of a lot larger than Safari.

Unfortunately for me, I can’t check IE7 at the moment as I’m on Win2000. Looks like I’ll have to get an upgrade, just so I can check a browser. Joy.

Kishore says:
July 28, 05h

The problem is that there are many computer users out there who have no clue about updates etc. These are the casual users and therefore I am not putting too much hope on IE 6 getting outdated so fast.

Yet, I think we should atleast appreciate what Microsoft is doing here. The intention is there, but the actual result we have to wait and see.

Dave S. says:
July 28, 08h

“I think it probably comes down to the question of target audience and/or the content that will be featured on the site.”

Right, exactly. I’m surprised that can’t go without saying by now. You need to measure your audience and make your own decisions.

Paul Solecki says:
July 29, 10h

I’ll be quite happy to continue supporting IE6 as there’s no reason people on earlier versions of Windows can’t use that. Anything before that is just not worthy of support.

At work we are in the process of finally stopping support for IE5.5 and earlier in any of the web and other work we do involving browsers.

Andrew says:
July 29, 22h

I cringe at the sites I’m going to have to fix with this update. I told them hacks were not a good idea, yet they wouldn’t listen to me. Now we’ll be paying for css ugliness…

Kev says:
July 30, 01h

But what do we do when IE6 still represents the bulk of what our users are using, and we need to keep it on our systems, and can’t afford to have it give way to IE7, and don’t have a spare/second system to have both running? Surely there’s a large number of developers in this boat who simply can’t replace IE6 with IE7 and move on? I know at least one (puts hand up…).

Ryan Barr says:
July 30, 10h

Kev: We technically can’t, and yes surely most can’t run IE6 and IE7. Other than the fact that IE6 doesn’t have that great of CSS2 support, I am sure that we won’t really need to worry all that much. As long as you have experience with IE6 and you know how to avoid problems (at least most of them) with the browser then the IE6 users should be fine.

Phil Nash says:
July 31, 10h

Let’s not get too excited. I think that if you need to support IE5 users then that will not change and IE6 users will remain significant for a long time. See my reasoning here:

August 05, 11h

It would be interesting to guesstimate the collective amount of profit lost in production time, hacking and reworking the CSS on thousands upon thousands of sites because of all the browser problems and inconsistencies.

Me, I run a business, I need to eat. I haven’t got hours to go back and rework sites I’ve already finished and billed.

Seems like a lot of folks here don’t mind the idea of putting in a lot of free hours.

George says:
August 07, 08h

This has been coming for a while and there have been plenty of warnings about IE hacks not working or breaking in IE7.

You can get a standalone IE7 for XP and using Evolt’s browser archive you can test on IE5 - 7. If you take a few simple steps now you can limit the impact on your code after the upgrade.

As far back as October Microsoft encouraged developers to stop using hacks.

Although in IE7 the code should be interpretted closer to the spec I’m trying to avoid the following:

- html > body
- html
- head:first-child + body
- head + body
- body > element

Sometimes though I feel it is like trying to build a house without bricks…

Jive says:
August 07, 09h

Then Greg, you are probably not running your business as effectively as you can. The fact that your clients site does not work correctly in ie7 is not your issue, it is the clients issue. If they would like their site to be compliant with ie7 then they should have to pay for the updates necessary to make it that way, just like any other maintenance update. It only means new billable business.

August 08, 18h

That is such great news about IE7, should speed the transition significantly.

But, did I read somewhere that IE7 will only run on XPSP2? If so, that’ll leave a lot of people out of the upgrade loop, as I’ve been surprised how many windows pc’s I come across that do not have SP2 installed!

August 11, 09h

If they would like their site to be compliant with ie7 then they should have to pay for the updates necessary to make it that way, just like any other maintenance update. It only means new billable business.

Jive, if only it were this simple. Clients will not view this as their problem. They will view it as me coming back to nickel and dime them about a technical issue that they don’t understand. My clients will not be happy about this, and if you’ve been in business for any length of time you will understand that unhappy clients mean less referral work, and more time needed on sales work to make up that loss of referrals.

Sure, I could probably get most of my clients to pay the ‘IE7 fee’, but they’re not going to be happy about it. And that’s probably a higher long term cost than just eating the IE7 work.

Cecil Ward says:
August 24, 13h

A couple of points.

IE7 RC1 is out today.

Dave should get a copy of IE7 RC1, as beta 1 was completely worthless in terms of site testing, in that it wasn’t even close to being representative of the level of CSS support that the future shipping IE7 offers.

About IE5; you don’t have to “drop support” for it, or “turn customers” away. Simply deliver it unstyled, by using conditional comments. Remember, delivering a standards-compliant website unstyled means it is perfectly _usable_, just not as pretty. Andy Clarke and others made the point “why should websites look the same in all browsers?” Isn’t it reasonable for a website to look prettier if viewed with more advanced software?

The point is that putting in a lot of effort to support obsolete software is a distraction, and designing down to the lowest common denominator is selling yourself short.

So “to hell with IE5 and IE/Mac”. Don’t worry. You’re not making anything inaccessible to anyone.

Cecil Ward says:
August 24, 14h

Accesskeys and IE7: I notice that CSSZenGarden’s use of accesskey attributes means that I lose keyboard access to basic menu functionality in my IE7 RC1.

I have reported this to MS as a critical “bug”, or more correctly a design defect.

The madness of accesskeys issue has been covered in a persuasive earlier article. It might be a pain to get it done, but CSSZenGarden would do well to strip out all these discredited attributes soon as a service to all of us keyboard users.

Browser manufacturers should take accesskeys far less literally, there’s nothing stopping them from ignoring the idea of literal single keystrokes completely and using them to build an “accesskey list menu” or other UI object.