either checking/updating all supported systems’ subsystems to use ‘correct’ HTML, DOM and CSS
or implementing much more radical doctype switching: when a ’strict’ or ’standard’ doctype is detected, use a new rendering engine – instead of merely modifying the behaviour of Trident.

When I look at the Wine project implementing a ‘bogus’ Trident with the help of Gecko, maybe looking at how Free developers develop an HTML engine for win32 may help.

Posted using Mozilla Firefox 2.0.0.4 on Linux.

You said “you must be willing to scrap and rewrite huge portions of platform code and really get the engine right at the base level.” If I felt that weren’t true, I would already have quit. Seriously. The challenge – what I’d hoped you would be there to help with, and what I hope you still will help with through the webdevout test pages – is in getting it “really really right”. That’s a standardization challenge in many cases as well.

-Chris

Posted using Internet Explorer (Windows) 7.0 on Windows.

Thank you very much for this post. You’re my hero for keeping your integrity in the face of temptation. I’ve been working at an extremely low-paying job for several years now, partly for convenience, partly for freedom. But I must admit money has been quite tight for a while.

To know that I’m not the only one in the world that is willing to trade large amounts of money and/or benefits for freedom gives me resolve to continue sticking with this ideal. Thank you.

Posted using Safari 419.3 on Macintosh.

Like I said in the main post, I agree that the “Internet Explorer is Dangerous” page needs to be rewritten. I plan to put more focus on the fact that IE’s platform development halted for five years and it simply isn’t realistic to catch up on all of that lost time overnight. I feel this is the single largest contributor to IE’s grumpiness toward web developers. Unfortunately, a lot of people I talk with regularly are under the delusion that if Microsoft really cared about standards it could pump enough money into IE to get it caught up to Firefox and Opera literally overnight. I know enough about product development to know that, in many ways, adding more and more and more full-time developers to a project can actually slow down development after some point, and there’s no magic bullet you can apply no matter how much money you have.

Although I would put less emphasis on the “fundamentally broken” bit, I still believe that there are some really serious fundamental problems with the IE engine, from HTML parsing and DOM building to the CSS implementation. When it’s possible for CSS to select comments and doctypes, you know there’s a problem, and I don’t think making special exceptions for those selectors is enough to fix the core issue. Raw CSS parsing and error handling in IE have lots of problems (many of which are beyond the current scope of my standards support tables) which could complicate future development of the CSS standard. The ever-so-prominent IE hasLayout model needs to be completely scrapped, and the slightly analogous CSS block formatting context model needs to be created in its place. The HTML object element is still chock full of problems in IE. So I still feel that there’s some truth to the “fundamentally broken” aspect of IE, but I’m going to elaborate a lot more on exactly what is broken and how these implementation problems cascade to other features.

As for security, I think the biggest issue was ActiveX, but now that that’s been more or less locked down (shoved into the closet, so to speak), I think the biggest issue now is the security response process. IE and Firefox have had roughly the same amount of vulnerabilities lately, with comparable severity, but the difference is that most of the IE vulnerabilities were publicly disclosed without a patch while most of the Firefox vulnerabilities were fixed preemptively. At a glance, it would seem that this could be attributed to IE’s higher usage and thus attractiveness to malicious users. But the problems aren’t really quite that shallow. Even if the two browsers have as many known vulnerabilities, Firefox vulnerabilities seem to be fixed more often and more quickly than IE vulnerabilities. There are tons of old IE vulnerabilities that were never fixed. Sure, most of them are spoofing vulnerabilities and other relatively minor issues, but Mozilla and Opera seem to consider all of those issues worth fixing. I’m disappointed that Mozilla has been seemingly letting more and more of these relatively minor issues slip by lately, but their consistency has still been better than IE’s. This is something Microsoft really could fix if it wanted to. In most cases, fixing some security bugs isn’t exactly as big a project as adding support for CSS generated content.

In regard to the “‘gotcha’ fine print”, I wasn’t talking specifically about employees, but also deals with other companies, organizations, and customers. From an outside perspective, Microsoft in general seems to convey an attitude of “whatever I can get away with is fine in my book”. Microsoft has partaken in competitive strategies which critics have named “embrace, extend, extinguish” which very much capture this impression. Microsoft appears to be attempting this strategy once again with Linux, by waving around a claim of patent infringements for which Microsoft has so far refused to present any evidence, in order to get buddy-buddy with major Linux distributors and ultimately put themselves in a power position. Yes, it’s business strategy, but it’s also sneaky and damaging to PR in the long run. It causes people like me to have doubts about signing any kinds of deals with Microsoft.

I know you’re the kind of person who believes in what you’re doing and has to believe in what you’re doing. I respect that. A web browser is a very complex product that takes a tremendous investment to develop to a usable state. Compare Internet Explorer (which I personally consider to have the most inferior of all major layout engines) to something like Amaya, and IE suddenly doesn’t look so bad. There’s plenty to be proud of in developing a product that complex to a really usable state. But this is a situation where we web developers have to develop our sites to work in all widely used browsers, and the fact remains that Internet Explorer is the main bottleneck of web standards usage. If Amaya had 80% usage share, I’d probably throw up my hands and apply for a job at McDonald’s. On the other hand, if IE were to suddenly disappear from existence, there wouldn’t be any clear “last place” engine among popular browsers. Despite what lots of web developers would say, the IE engine isn’t a total loss, but in order to move forward and try to catch up with the other browsers (that is, catch up on half of IE’s life of lost development time), you must be willing to scrap and rewrite huge portions of platform code and really get the engine right at the base level.

Gecko 1.9, which will be used in Firefox 3, underwent some major rewrites in the main document processing engine to fix the page reflow system. Gecko 1.9 completely overhauled its graphics platform, switching from Gecko’s old custom-built GFX system to Cairo. Gecko 1.9 completely changed the way CSS units are handled internally. Gecko 1.8 also had some big rewrites to fundamental areas of the CSS engine. Presto (Opera) and Gecko are where they are today because their developers were never afraid to tear down engines and frameworks and rework them from scratch in order to fix a few relatively picky issues. And they had five extra years to do so. In order for Internet Explorer to catch up to the competition, it must make the same commitment to perfection, and the work must be done faster than the competition. That’s a tall order for any development team, and I will have tremendous respect for you guys if you can thoroughly pull it off.

Going back to the second topic of this post, Mozilla finally got back to me and gave me a phone interview a few days ago. I thought it went terribly (I was ready to answer questions about front end web development work, and he started asking me questions about lower-level programming algorithms and stuff I couldn’t answer off the top of my head), but they’re apparently interested enough to have me come down to meet with them in person next week. The position would involve front end web development on the add-ons site and some interesting projects in the works. And Mozilla is about as transparent as a “corporation” can get. It isn’t as major of a position as Microsoft was looking for, but it’ll probably be a more comfortable transition for me. Let’s hope the on-site interview goes a bit better than the phone one.

Posted using Mozilla Firefox 2.0.0.4 on Linux.

Posted using Internet Explorer (Windows) 7.0 on Windows.