Quiz Microsoft's IE Team Leader

About as timely an interview as you can get: Microsoft released Internet Explorer 7 last week, and today we're gathering questions for IE team general manager Dean Hachamovitch. As usual, please follow Slashdot interview rules when posting or moderating questions. We'll publish Dean's answers verbatim as soon as he replies.

Comments

Re:Prediction:

[Roblimo]

I'll add another prediction: That lots of people won't read and follow this note in the Slashdot Interview FAQ:

You can ask as many questions as you'd like!

But please, only ask one question per submitted comment.

You can ask a compound (multi-part) question, but if you make your question so complicated that no one's sure what you're asking, it's less likely to be moderated up. If you have several burning questions, take a minute to organize your thoughts and separate them into multiple comments.

- Robin

Re:How much of SpyGlass is left in IE7?

[linuxci]

IE7 is the first release that does not mention Spyglass Mosaic in the about box.

Re:CSS

[AKAImBatman]

Are you aware that no matter what answers you give here, they will never satisfy the anti-Microsoft Slashdot crowd?

It's difficult to speak for everyone, but I can give my own opinion on your question.

I used to like IE5. Whether I disliked Microsoft or not, it was a superior browser in its day. The problem is, that the standards that Microsoft helped create all those years ago are not actually supported by Microsoft today. To use the example I pointed out in my own question, IE's lack of DOM 2 Events support means that there is absolutely no way to write DHTML code that works in both IE and Firefox. Yet, I can easily write code that works in Firefox, Safari, and Opera.

Is there any reason for this dichotomy? Yes and no. Code can be made to work across the major non-IE browsers, because they all provide at least basic support for the W3C standards. IE has its own attachEvent() model that is (obviously) incompatible at a code level, and subtly incompatible at the behavior level. All that microsoft needs to do is to lay a parallel API that supports the W3C standard, and I would be a happy fellow. Yet they haven't done that, won't do that, and I have NO IDEA WHY.

So I continue to write code that works in Firefox, Safari, and Opera, then special patches to make it work in IE. From where I'm sitting, I just want the problem to go away. If Microsoft fixes their browser, then I'll be happy. If Microsoft can't do that, then I will carry the "Down With IE!" torch until their browser is irrelevant in the market. Then I'll also be happy.

Basically, my web browser opinion is not one based on my feelings about Microsoft. I just want a market were I can target a single standard is all. If Microsoft abuses their Web Browser monopoly to stand in the way of that, then it is my duty as a web developer* to help smash that monopoly.

* What happened to the "Developers, developers, developers" jingle, hmm? Are we important, or aren't we?

Re:client-side persistent storage

[gradix]

For Ajax apps that you want to run offline (and synchronize when online), it can be useful...

Furthermore...

[Anonymous Coward]

Myself and many other web developers have no intention of ever running Windows. We have even less intention of working around flaky standards support in the Windows default web browser. Is there an end-user support address or bug tracking system where we can send users that wrongly blame us for problems with your software?

Re:Interface

[jmyers]

Agreed, the UI for IE7 is strange. If this is a hint of Vista I expect Mac and Linux will pick up a few desktop users afer the release.

FYI this reg setting will move the menu bar to the top.

[HKEY_CURRENT_USER\Software\Microsoft\Internet Explorer\Toolbar\WebBrowser]
"ITBar7Position"=dword:00000001

Re:How about this...

[Ed Avis]

Yes... IE7 does run under Wine - so it wouldn't be too hard for Microsoft to make an official Linux/i386 version.

Re:How about this...

[Anonymous Coward]

At the time, I found it more useful and reliable to build Mozilla nightlies.

Unless you needed NTLM authentication, which Mozilla couldn't do back then.

I asked Hakon about CSS and now I ask you:

[Chabil Ha']

This past summer Håkon Wium Lie was interviewed on /. and my question was selected concerning IE7's glaring lack of full CSS support. Why is it that MS has avoided meeting at least the ACID2 spec for CSS in order to bring some semblance of comformity for developers?

Håkon Wium Lie's response to these questions is boiled down to the fact that you do have the talent and resources to fix these issues and he says that "the fundamental reason, I believe, is that standards don't benefit monopolists" like MS.

How do you respond to his comments (the author of the CSS spec) and does MS have any near future plans to adhere to the existing CSS standard? If not, what would it take for MS to take a more proactive role in supporting it?

Re:Interface

[AdamKG]

Thirded. I used to use IE7 at work as the HTML editor for our ticketing system was IE only, and *I*, a computer power-user who has been using MS products since I could type, worked for a week before finding the menubar, quite by accident. When this gets pushed out via WU, millions of people will have no idea how to find the menubar, and I can't blame them.

I ended up using Portable FF with the IEtab extension enabled for the one site I need IE for. Much nicer. I get to maximize screen real estate, keep most sites rendering correctly (IE7's rendering of the school's homepage is messed up- something to do with CSS, ironically enough), and have the extensions I live off of. And I have a menubar.

Re:Not just the Events module...

[Snover]

While I agree that missing the Events module is definitely the BIGGEST problem, it's certainly not the ONLY problem, which is mostly what I was trying to get out. Also, I'm a fucktard and didn't preview my comment before posting (it's the only time I've ever not previewed a comment on Slashdot, I swear ;)) and Slashcode wiped out a bunch of what I was saying.

What I MEANT to write was:
* Tables that are created dynamically will not appear unless elements are added to <tbody> -- in other words, using DOM to write <table><tr><td>content</td><tr><table> does not display anything, even though <tbody> is optional.
* getElementsByTagName('object').item(0).getElements ByTagName('param') returns EVERY <param> IN THE ENTIRE DOCUMENT, not just ones under the requested object. Yikes!

As far as the navigator.plugins is concerned, it's definitely possible that it's a fragment of the browser wars, but that would be rather weird (except as an intentional attempt to sabotage Netscape Plugins API in order to push ActiveX -- something I find much more plausible than "fixing Netscape-only sites"). An empty nagivator.plugins would be just as easy to detect as one that didn't exist, and if something depended on it being there so that it could find a plugin, the site still wouldn't work in IE.

As an aside, between my last post and now, I ran some JavaScript unit tests from the W3C Compatibility Test Suite.

http://www.w3.org/2003/02/06-dom-support.html

DOM Conformance Test Suites
==========
DOM Core 1
----------
Internet Explorer 7 (37.95% failure)
Tests: 224
Errors: 4
Failures: 85

Opera 9.02 (11.34% failure)
Tests: 238
Errors: 12
Failures: 27

Firefox 1.5.0.7 (5.88% failure)
Tests: 238
Errors: 15
Failures: 14
==========
DOM Core 2
----------
Internet Explorer 7 (45.45% failure)
Tests: 11
Errors: 2
Failures: 5

Opera 9.02 (36.36% failure)
Tests: 11
Errors: 0
Failures: 4

Firefox 1.5.0.7 (18.18% failure)
Tests: 11
Errors: 0
Failures: 2
==========
DOM HTML 2
----------
Internet Explorer 7 (8.47% failure)
Tests: 685
Errors: 16
Failures: 58

Firefox 1.5.0.7 (1.31% failure)
Tests: 685
Errors: 0
Failures: 9

Opera 9.02
I tried about 20 times to get Opera to
complete this test, and it would constantly
fail to load files and Retry did nothing.
Oh, well. I'd guesstimate the failures are
probably around 2-3%.

I'd say there's a pretty serious fucking problem when nearly 38% of DOM Core 1 functionality is BROKEN.