Still broken

It’s been a few days now since the announcement on A List Apart. Since then there has been some good debate but a lot of noise and fury signifying nothing. Eric put it best when he said:

If you feel the need to impugn the integrity or intelligence of another person to oppose an idea, you’re undercutting yourself, not your target nor the thing you oppose.

I think that Jeffrey is right when he observes that there might have been less vitriol expended had this proposal come from anyone other than Microsoft. Still, it’s hardly surprising that the initial reaction to Microsoft’s version targeting is one of shock and rejection. After all, that was also the initial reaction of both Eric and Jeffrey.

Let me clarify what I think of the proposed meta element for version targeting, just in case it wasn’t clear from my previous post:

This proposal is fine. I don’t think it’s great but I think it’s a pretty good solution to a very real problem that Microsoft are facing. As Jeffrey said:

I don’t love version targeting but I see that it serves a need.

That need is ensuring that Microsoft’s customers won’t have to spend time and money jumping through hoops every time that a new version of Internet Explorer is released (as happened with IE7). Yes, I know that if these people—mostly in the Enterprise world of intranets—had just written to standards in the first place then they wouldn’t have had any problems. But that doesn’t change the reality of the situation that those people are in. Any proposed “solutions” that involve abandoning these customers—or, in the more extremist suggestions, abandoning Internet Explorer—simply aren’t facing up to reality.

The http-equiv="X-UA-Compatible" instruction is a solution that requires minimum effort from publishers (although it does require a lot of future effort from the browser manufacturer). The switchover from IE6 to IE7 cost a lot of people a lot of time and money as they scrambled to move from CSS hacks to conditional comments. It’s no secret that Microsoft view this as a failure, hence the new proposal which will ensure that sites can be frozen at the level of IE7 even after IE8 is released.

Microsoft can now provide their customers with a ludicrously simple answer to any future problems. All they have to do is add a meta element to their documents (or set up their server to automatically output that header).

But that isn’t what Microsoft is proposing. Instead, IE8 will behave as IE7 by default. The meta element is required if you want IE8 to behave as IE8. Mark Pilgrim encapsulates the madness of this approach:

  • If you give me non-standard markup, I will render it according to standards.
  • If you give me standard markup, I will not render it according to standards.

Let me make it perfectly clear: I understand the need for version targeting. But the onus must be on the publisher to enable it. The effort required is minimal—much less than any pain that was endured in the move from IE6 to IE7.

On the plus side, future versions of Internet Explorer might not cripple themselves on encountering HTML5 documents. That’s good news for the future. Sometime in the ten years, this will be a solved problem. In the meantime, there will be millions of documents written in HTML 4.01 and XHTML 1.0—both perfectly fine DOCTYPEs. There’s no reason why they should be treated as second-class citizens by default.

Chris Wilson read my misgivings and commented:

However, the problem is that we can’t tell the difference between your content (valid HTML with a correct HTML 4.01 DOCTYPE) and that of someone who hacked up their own content to work with IE6/7 when faced with an IE browser, and whose content will break entirely if we correctly implemented standards as specified.

It’s true that a browser can’t read minds but rather than ship a browser that’s crippled by design, it seems far, far better to me to allow people to tell the browser to regress. I think that asking people to add one meta element isn’t asking too much. As compromises go, it would be a great way for publishers and browsers to meet each other half way.

I believe that Microsoft want to help their customers. I believe they have found a reasonable solution. But in defaulting to crippled functionality in their next browser, they are hurting everybody else.

Jeffrey describes this as an implementation detail but it is far more than that: it is the implementation.

A few years ago, Microsoft introduced some proprietary CSS properties for coloured scrollbars. Fine. Anybody who wanted to make use of this “innovation” was free to do so. The rest of us could just ignore it. Now suppose that instead of making coloured scrollbars an option, they were the default… and if you wanted to revert to the standard way of displaying scrollbars, you had to use the proprietary code. See the problem?

Non-standard browser features are nothing new. Internet Explorer has already given us XMLHttpRequest and innerHTML. We are free to use these or to ignore them. Now, for the first time, a browser feature is mandatory: if you don’t want to use the new feature, you have to disable it… by using the new feature.

Jeffrey is in favour of this proposed default behaviour. He argues very persuasively that it will help the non-professionals who are building websites:

Teachers, pastors, coaches. Managers and owners of very small businesses that can’t afford to hire anyone reading this page to create a snazzy website for them.

I passionately believe that anyone should be able to publish online. That is the vision of Tim Berners-Lee’s beautiful Web. But I don’t think those people will be affected by this decision one way or the other; chances are they are using old-fashioned table-based markup. Besides, they are not the people that Microsoft are trying to help. Microsoft are trying to help the people with money (which is fair enough: Microsoft is, after all, a profit-making business).

I’ve listened to the arguments in favour of IE8 behaving as IE7 as default. I still believe it is the wrong behaviour. Now what can I do about it?

Jeffrey counsels a polite approach:

And, presumably, if enough developers believe the default should be “current version” instead of “IE7” — and if they argue their case coherently and dispassionately — Chris Wilson and his colleagues might come to agree with you, and the version tracking default might end up the way you want it instead of the way it was originally proposed.

If that’s the part that bothers some folks here (like it bothers Jeremy Keith), the thing to do is politely take it up with Microsoft’s Chris Wilson. That’s how stuff changes: by politely talking to the people who make our browsers.

I concur. I hope that the decision regarding default behaviour lies in the hands of people like Chris Wilson and Marcus Mielke. I know that Chris is passionate about standards. I know that if he believes that changing the default behaviour is the right thing to do, he will do it.

Eric agrees with me. Many of the comments on Jeffrey’s blog also agree:

  • I support version targeting but the part that confuses me is that if the tag is excluded, it defaults to IE7.

  • I’m with Jeremy — it SHOULD default to IE8.

  • Many of the comments against the A List Apart articles seem to more or less echo what Jeremy Keith is saying.

  • I think that in IE8 the default should be IE8; but render as IE7 when asked (with version targeting).

  • Make IE8 work like IE8 by default. If your site breaks, then - and only then - add the meta tag to beg for IE7. Adding the meta tag will be easy for the small number of sites that require IE7 (ie. those sites that still ignore the 10-20% of non-IE browsers such as FF, Safari and Opera).

  • After a lot of reading and some hard thinking I think Jeremy is right. The meta could be a useful tool if the default is IElatest, not IE7 8 or any specific version.

  • If you’re going to have a fallback, you may as well use the gold standard in incorrect rendering; the one that the entire web was coded to for 5 years: IE6. If you’re going to implement something as logically backward but seemingly necessary as this version targeting, you may as well do it in the most backward way imaginable.

  • Far better is to force those who made poor choices in the past by working to IE7 implement a simple 1-line fix to get their legacy site to render under the IE7 engine. I’m sure the MS folks can get the message across: “Broken site in IE8? Add XXX to fix it.”

  • I’m also with Jeremy Keith (and others on this).

  • I’m going to go with Jeremy’s request for sites to use the latest rendering mode unless the meta tag say otherwise.

  • The most logical execution for this tag is that when it is not there, it defaults to the latest engine.

  • I really cant agree with the attitude that in order to do things right in a certain browser, you have to go out of your way and add a tag.

  • IE8 rendering *must* be the default in IE8 to further promote standards-based, interoperable websites.

  • IMO, I think they should do it the other way around. Make a flag available to mark ie8 sites into ie7 mode.

  • I am still not clear on why any site I make that conforms to standards has to include a new line of code to make it display according to standards.

  • If you don’t specify this meta tag, you will be served an IE7 rendering engine. Having opt out by default is one of the key arguments for most of the people saying that the suggested implementation of version targeting is wrong.

  • I am in favor of the version targeting, because it seems a very good way to untie the hands of IE’s design team. I am still not convinced by your argument for defaulting to IE7.

  • Making the default version IE7 rather than “edge” seems like a bad idea. If it’s as easy as making a one-line change to your web server configuration (to add the HTTP header rather than the meta tag), then corporations will have plenty of time to future-proof their intranet apps against IE8.

    Adding the versioning capability, but having a default of “edge” gives corporate sites/intranets a way to protect themselves. This allows Microsoft to stop worrying about backwards compatibility and lets them focus solely on implementing the standards correctly.

  • I don’t agree with the direction of the switch; it really should default to IE8 and give people the meta to invoke Trident.

  • The switch is all cool, but it’s opposite of what it should have been. Instead of forcing developers that are following standards to insert this meta element to actually be able to use these standards, Microsoft should reverse the switch so those targeting a rendering engine has to insert it instead.

  • PLEASE, reconsider that default behavior, it is NOT a “detail.”

  • I’m not seeing how the current form of the tag is better for standards over making the tag an opt-in for IE7 rendering.

  • If the default was reversed I think no one would raise so much as an eyebrow for this proposal. In fact most people would positively support it. As it stands this problem is just crazy.

  • Even if IE8 can display everything my style sheet declares, if I omit the META declaration, IE users will get a sub par experience.

  • *”Edge” should be the default, not IE7!* (That’s my new sig ;)

  • If the default behaviour of any future IE version is to render as ITSELF, then surely all lousy developers need to do to fix their sites is put in the required meta tag.

And here’s a cherry-picked sampling from the IE8 blog:

  • I’d prefer having the superstandards mode on by default.

  • SuperStandards should be ON by default. If IE7/quirks mode is needed, the meta tag should be used.

  • Standard should be on by default, or just call your browser ‘IE 7 extended’.

  • I’d prefer having the superstandards mode on by default.

  • I agree, super-standard mode should be the default mode.

  • Let’s have the IE8 Standards mode on by default.

  • Make it an OPTION to work in IE7 or IE6 or whatever mode. But make it DEFAULT to run IE8 “super standards” (this should be called standards mode) mode.

  • Standards mode should be the default.

  • What i think would be a good idea is to turn this exactly around: add a meta tag or http header to sites that need to be rendered by IE8 in ‘crappy-mode’, ‘ie6 mode’ or ‘ie7 mode’ and kick it into standards compliant mode by default.

  • There SHOULD be an option to trigger IE 8 to use old rendering engines but it should be disabled by DEFAULT

  • A site that wants IE8 to behave like IE7 should have to specify the meta tag, not the other way around.

  • Use standards compatibility mode by default.

  • I personally think that you should really consider making the real IE8 engine (“super standards” as some have called it) the default option.

  • So let me get this straight… I have to add an additional meta tag to my page in order for IE8 to support standards?!?!?

  • Standards mode should be the default.

  • Standards mode should be the default if the doctype is declared.

  • Super-standard should be the default, period.

  • I’m coming round to the idea that this is workable *if* the default is changed to IE8 - standards compliance mode - the meta then provides a fix for any site that breaks.

  • As some have previously mentioned, the Standards compliant mode should be the default.

  • All browsers must be in Standards Mode by Default.

  • Please make fully standards compliant mode the default.

  • In case it’s not clear: Make the standards mode default.

  • It’s great that IE 8 will have a much better standards support. It’s bad that it won’t be on by default, requiring a switch.

  • The idea is good. The default is bad.

  • Standards Mode by Default. Standards Mode by Default. Standards Mode by Default. And now repeat!

  • I too think that this is a great idea, but the default behavior is evil.

  • It is fundamentally wrong to stick us with IE7’s broken implementation as a default.

  • Standards should certainly be default, not an opt-in.

  • I’m no web developer, and I never, and I mean never say anything anti MS,but why can’t standards mode be enabled by default?

  • Make IE8 standards compliant by default.

  • Fellas, I really appreciate what you’re trying to do. Improving the renderer was the right thing to do, and you’ve done that. Bravo! The wrong thing to do is to have the correct behaviour require yet another tag.

  • The default when processing a correctly specified doctype should be standards mode.

  • I have to agree that full standards compliant mode should be on by default.

  • I agree that super compliant mode should be enabled by default and the meta tag should be used for Quirks mode.

  • I have no problem with the concept of a meta tag that allows developers to target specific browsers, but the default should ALWAYS be the latest and greatest.

  • After reading the various articles (and comments) on this approach (i.e. the meta tag “switch”) I really think that IE 8 should default to the latest and greatest rendering first - not the IE7 rendering engine.

  • If you do go ahead with this: make super standard mode default please.

I kind of feel like John Turturro to Chris Wilson’s Gabriel Byrne:

I’m praying to you! Look in your heart!

Have you published a response to this? :

Previously on this day

15 years ago I wrote No comment

Paul Haine got in touch with me and asked:

18 years ago I wrote The horseradish challenge

Roll up, roll up, ladies and gentleman. Watch a grown man attempt to eat an entire jar of horseradish in less than 10 minutes.

18 years ago I wrote Space News

Could it be that black holes don’t exist after all? Stephen Hawking may have to rewrite his books.

18 years ago I wrote ICQing Argentina

I just had a nice chat via ICQ with a gentleman from Argentina.