Discuss: Version Targeting: Threat or Menace?

by Jeffrey Zeldman

21 I still don't buy it

I agree with Brian LePore above. The mechanism is fine, but the logic is flipped. These intranets that are so dependent on specific IE versions are also those that are in the most highly controlled and maintained environments. Getting them to add a header to their servers is an utterly trivial 100% fix, and something they can do right now rather than waiting until a new IE release blows it up. Let’s say there are 100,000 such sites worldwide, vs. 100,000,000 other sites. That means that 99.9% of sites are going to have to change all their pages unnecessarily in order to keep broken stuff working on networks that are invisible to the outside world. So are MS going to pay for the millions of wasted man-hours they are planning on having everyone spend needlessly? I see any argument in favour of this is just insane.
The “mom & pop” sites that may break in a minimal way don’t really have much bearing on the argument – they have already been broken for years in browsers that do things right.

posted at 06:44 pm on February 19, 2008 by MArcus Bointon

22

@David One said:

There’s reasons for that, and they were all pointed out to you, over and over again the last time we all had this conversation. You’ve not added anything to your earlier list of fallacious, emotive, and wrong-headed thinking from last time.

It’s hard to argue with an unsupported list of negative adjectives.

I’ll leave others to more fully deconstruct the attempted justification presented in this article.

You mean you’ll leave it to people who state opinions supported by evidence? That sounds like a good idea, since calling someone “fallacious, emotive, and wrong-headed” is not the same as engaging in a discussion.

Nor, for that matter, is quoting Håkon Lie the same as engaging in a discussion about the ideas presented in this article. Håkon and I are friends; I respect him greatly; any sentient web designer is in awe of Håkon’s contributions to this medium. But as CTO of Opera, he is hardly a disinterested party. (And even if he were disinterested, quoting him is not the same as advancing your own argument.)

Version targeting, as I explained, is the seat belt Microsoft must wear to continue improving its web standards support. If you dislike version targeting, don’t opt in — and by opting out, contribute, albeit slowly, to the decline of IE’s market share. If you’re unwilling to cheer the notion of IE’s improving its DOM support, or to read with an open mind an explanation of why the toggle (like DOCTYPE switching and Firefox’s JavaScript targeting) must be opt-in in order to work, that’s fine with me.

But give us something to work with beyond name-calling.

posted at 07:02 pm on February 19, 2008 by Jeffrey Zeldman

23 The real reason

Surely the real reason behind the decision is not to avoid “breaking the web”, but to avoid “breaking the revenue stream”. I would suggest that the real problem for MS is companies choosing not to upgrade primarily because of problems with web apps and intranets that would break in the next release of IE.

This problem could be solved simply by having a user-configurable whitelist of “legacy sites” and/or complete switch “Make IE render all sites in legacy mode” that can be applied by sysadmins. Legacy mode would behave like IE7 and respond to IE 7 hacks as appropriate. Legacy mode would also be used for sites that specifically request it via a header or meta tag. Similarly, sites could be able to request the full standards mode to override any legacy settings.

The big question is: do we want to have standards by default, or do we want non-savvy web designers continuing to code for the quirks of IE for years to come? Since the alternative solution above would satisfy MS’s problem, I think the answer is clear.

posted at 07:07 pm on February 19, 2008 by Robert Whittaker

24 It's Microsoft's problem

IE6 is the reason why some sites break in IE7. Microsoft needs to fix the problem, not put the onus on informed developers to work around it.

I think the adoption of competing browsers like Opera and Firefox has been severely hindered by poorly constructed sites built specifically for IE6. At least that was my impression when I first made the jump from IE to Opera 6.

Microsoft’s anticompetitive solution guarantees an infinite future of poorly constructed sites built solely for IE7.

posted at 07:23 pm on February 19, 2008 by Scott Lenger

25 Why not shift to user based control of rendering.

From a business point of view i fully agree with the Microsoft direction. It makes sense not to break the internet and the way they have suggested although rather hack like does have some merits.

But the view i would rather see Microsoft take is a button on the tool bar (or a system preference) which allows the user to switch to IE7 ‘IF’ a website does not render correctly in IE8 mode. This i admit would not fix the core issue – but it would allow the user choose their rendering engine and encourage the faster adoption of a more standards compliant mode by users and developers.

I wonder if its time for Microsoft to kill of IE and start a new browser development. I have no idea how possible that it is, but it seems to make sense to lay IE to rest and start a new browser which can be more agile and nimble in the modern marketplace.

posted at 07:36 pm on February 19, 2008 by chris bush

26

Craig Francis’ comment above about making the meta tag a date makes a lot of sense to me.

I think a lot of the negative reaction to this change is due to its Microsoft-centric nature. Which makes sense because, as Jeffrey points out, it’s a Microsoft-centric problem.

If the meta tag looked something like this:

the opt-in would be not about a company or its browser. Instead it’s a general purpose piece of information that a browser can choose to ignore or use. And if that browser chooses to use the information, by default it’s browser specific.

Web savvy developers would know that they need to give IE a production date to trigger the behavior. And if there’s a new version of their website, they bump the production date after testing (production date can also be thought of as “the last date this site was tested to ensure compatibility.”)

Other browser makers have indicated that they don’t intend to use the X-UA-Compatible meta tag. Fair enough.

But could they use a piece of meta data like a site’s production date? Predicting the future is messy business: I wouldn’t rule it out by any means.

-ch

posted at 08:17 pm on February 19, 2008 by Craig Hockenberry

27 I really look forward to IE47...

I really look forward to IE47 that you are talking about, which at least going to have 43 different render modes and probably going to be a couple of gigabytes in size just to fit all those different render engines that has been produced over the years ;-)

To be honest I do not even think IE will ever reach version 47. Somewhere at version 20 or something (or probably much sooner) the IE-team will be so busy fixing old security bugs in all different render engines that they probably going to decide to drop the whole project, just after they have realized that every new version of their browser gives them exponentially more work every time they found a security flaw they have to fix ;-)

posted at 08:42 pm on February 19, 2008 by Martin Odhelius

28 My experience follows Zeldman's narrative

I create user interfaces for IE-based reporting applications. Good money, glamorous work. I also support standards. As a matter of unremunerated pride, the layout structure, design templates, CSS, and JavaScript I turn over to development teams and documenters comply with modern coding standards. The screens hum Firefox under their breath.

Often within minutes or hours after turnover, neither IE nor Firefox can handle the screens as designed.

Firefox compatibility is most often broken by programmers’ scripting efforts. It is amazing to see what damage trained and experienced object-oriented programmers can do with JavaScript. “Document.all” and other unnecessary code continues zombie-like. Structure is mixed promiscuously with behavior in one big mish-mash.

On the other hand, IE is most often broken when some tool decides to add inappropriate doctypes or reformat syntax.

So I go to the source of the fire and stamp it out. Part of my job. That is why the “opting in” part of Microsoft’s latest hack is attractive. The new tag is one more tool to keep the people most knowledgeable of Web pitfalls in control of the process.

No silver bullet. Nothing will prevent Adobe Dreamweaver from automatically adding this meta along with inappropriate doctypes for instance. Just one more tool of arcane Web lore that will allow us humble cubicle dwellers to keep our bosses impressed.

posted at 10:21 pm on February 19, 2008 by Brett Merkey

29 More to work with

@Jeffrey Zeldman:

It’s hard to argue with an unsupported list of negative adjectives.

My post, if not apparent, was not an invitation for dialog – I was simply voicing passionate disapproval for the proposal, and referencing articles and comments which are available to anyone with a few minutes searching. It wouldn’t benefit this forum for me to fill it with copy and paste.

quoting Håkon Lie the same as engaging in a discussion about the ideas presented in this article.

Again, the purpose of that post was not to engage anyone in direct discussion. It was a relevant article that would be of interest to anyone involved in, what I understand is, a debate.

Håkon … as CTO of Opera, he is hardly a disinterested party.

Even if Håkon has sold his soul to Opera’s plan for world domination, he’ still advocating the best route for open web standards. Opera has been arguably the most standard’s compliant browser since launch, and the fact that Håkon is defending web standards now says more about his passion for standards than it does a profit margin at Opera Inc. Whereas, Microsoft will have to buy a bigger piggy bank if they manage to create a false standard that favours their browser.

Version targeting, as I explained, is the seat belt Microsoft must wear to continue improving its web standards support.

No matter what the benefit to some corporation, we cannot accept a standard that benefits that corporation at the expense of its competitors. The default rendering of each browser must be as standards-compliant as possible, not a rendering mode used by only one company on the planet.

If you’re unwilling to cheer the notion of IE’s improving its DOM support, …

Classic strawman. Just because I’m against the default rendering proposal, does not mean I’m suddenly against anything and everything.

or to read with an open mind …

I’ve been reading with an open mind, thanks, and I still don’t agree. Do I need to open it more? Remember, it’s possible to be so open-minded your brains fall out. ;)

But give us something to work with beyond name-calling.

Hope that helped.

posted at 11:02 pm on February 19, 2008 by David One

30 I have an idea...

Still not swallowing the blue pill. It’s very obviously wrong regardless of all the repeated attempts to sell the idea to us.

Here is a better idea:

Release IE8 as a completely new browser (new name, etc) that can run independently of IE7. The new browser would have a different UA string and would have no need for an IE7 emulation because IE7 would still be available.

Users can then do their day-to-day browsing in a shiny new MS web browser that’s standards enabled by default.

Found a crappy old site that won’t work in modern browsers? No problem – just open the blue “e” and view it in good ol’ Internet Exploder.

I think this approach solves all the problems without causing new ones:

Businesses can rest assured that IE7 will always be IE7
Microsoft gets a fresh start to make a super new browser (IE8 under a different name)
The entire planet knows that the future of the web is in standard compliant sites

Buggy old sites can still be viewed with IE7

In effect, you cut the cancer of old buggy sites out of the internet. People can still view those sites, but everything going forward will need the standards compliant browser.

posted at 11:25 pm on February 19, 2008 by Guy Fraser