credit where it’s due

There’s going to be a lot of talk this week coming out of MIX, about IE8. Early reports are interesting, if still often hidden behind “NOT PUBLISHED” links on MSDN, but the most interesting thing I’ve seen yet is their change in IE8 rendering mode default. The specific change is a good one, but even more than that I think it’s very promising as a matter of process.

The original decision to make IE8 default to matching IE7′s legacy rendering mode was made in secret, secret even from many of the web experts in the organizations linked to from Dean’s post. Once the conversation was opened to input from the rest of the world’s experts on web content compatibility, they were able to get to a much better decision, and happily that’s reflected in the updated plans for IE8. I hope that this case helps Microsoft understand more generally that significant web-compatibility decisions are too important to be left to closed groups: the web is simply too big, with too many stakeholders, for that to be a workable path to success. (Not that simply operating under “a standards body” is sure to avoid such secrecy, as they can also have related damage, but they at least tend to have more diverse viewpoints, which is a half-measure of some value.)

I was also heartened that they were able to make such a change after announcing it. We’ve all heard before that the IE team wasn’t able to talk about their plans until they were very certain, because people build businesses on those plans and will be harmed are made after an announcement. The conspicuous lack of bankruptcies attributable to WinFS being dropped from Vista aside, that they were able to listen to “global” feedback and make a significant change based on that feedback gives me hope that a new, more open process may be beginning here. Bravo and thanks to Microsoft for listening genuinely and making a change that I think will have a very positive effect on standards-based content on the web.

Some of the items listed in the “IE8 Readiness Toolkit” look pretty interesting, and in at least one case (the “XDomainRequest” API) seem pretty close to the subjects of some recent discussions in standards groups and various open projects about solving similar problems. I’m not sure if Microsoft proposed their API or semantics to that group, or shared the design thinking that went into their choices, but I’m permitting myself to hope anew!

(In light of my previous post about compatibility liability, I was also very interested to see this titbit:

[...]we do not believe any current legal requirements would dictate which rendering mode a browser must use[...]

It’s much more likely that they’re referencing the Opera suit than that they’re talking about Microsoft representatives’ previous claims of lawsuit risk stemming from changes in new product versions, but it made me smile nonetheless.)

6 comments to “credit where it’s due”

  1. Kim Sullivan
    entered 5 March 2008 @ 4:00 pm

    I wouldn’t talk about Microsoft in general, it’s probably just the IE team… It just takes one look at the recent OOXML standardization mess (covered e.g. on to see that it’s still the same old Microsoft it always was, the good decision on IE8 notwithstanding.

    Sorry if it’s a bit OT, but I’ve just read some of the more gory details of the voting on OOXML, and I can’t but think “dear god, tell me this isn’t really happening, there’s has to be some kind of misunderstanding, this can’t be real…”.

  2. entered 5 March 2008 @ 4:26 pm

    Indeed, Microsoft is a big company, and can act in multifarious and conflicting ways, but since I was willing to blame “Microsoft” for bad behaviour by the IE team, it only seems fair to credit “Microsoft” for the good behaviour. You’re probably right that the IE team is more enlightened here than some other parts of the company; working on the web has that effect on people sometimes. :)

  3. entered 7 March 2008 @ 5:48 am

    Re: XDomainRequest, see a blog post by the editor of the competing W3C spec:

  4. Mitchel Tyrell
    entered 11 March 2008 @ 12:52 am

    “It just takes one look at the recent OOXML standardization mess (covered e.g. on”

    That’s kind of sound like, “It just takes one look at the Democratic presidential compaign mess (covered e.g. on Fox News)”

  5. entered 11 March 2008 @ 2:31 pm

    Hey Mike, It’s not actually crrect to say that the decision to default to compatibility was made “in secret”. I’d been saying for over a year, in every talk I’ve done, that we would be needing web developers to opt in to standards compliance.

  6. entered 12 March 2008 @ 9:50 pm

    That’s true, you had said so, though the decision being announced isn’t quite the same as the decision being made in the open, or in a way that permitted broad feedback. It’s more that the mechanism — theoretically one that’s of cross-browser utility, given its naming and the examples on the site — was designed in secret.

    There’s more of this in the IE8 unveil — for example, XDomainRequest is just sort of out of nowhere, though MSFT ostensibly participated in the W3C’s efforts to find a solution to that problem for XMLHttpRequest. No proposal I can find anywhere about the scriptable connection-limit stuff, nor the timeout extension to XHR itself.