Just what has Microsoft been doing for IE 7?

Jeff Reifman writes "Last week, Windows columnist Paul Thurrott ripped into Microsoft for ignoring CSS standards with its upcoming Internet Explorer 7.0. "Microsoft has set back Web development by an immeasurable amount of time. My advice is simple: Boycott IE. It's a cancer on the Web that must be stopped. IE isn't secure and isn't standards-compliant, which makes it unworkable both for end users and Web content creators." With the redesign of my own site last month, I discovered just how non-compliant IE is with basic CSS: IE 52% vs. Firefox 93%. Is Microsoft purely incompetent and tone-deaf to customers — or simply counting on IE's non-compliance remaining a de-facto standard?"

I vote de-facto standard

[edflyerssn007 (897318)]

I believe that they are just hoping that IE remains the standard as it will come pre-installed with Vista and will be going out on automatic update, so the vast majority of windows users are going to move over to IE7 with-in a year or two.

-Ed

Re:I vote de-facto standard

[fm6 (162816)]

"Hoping" is the wrong word. They know that they're guaranteed 85% of the user base, and don't see any reason they should care about any standards except their own.

And before somebody says, "OK, IE is the de-facto standard, we can all just code our pages to use it." Ask yourself this: when you write code in C++ or Java or Perl, do you blindly guess what might work? No, you look up the language features and APIs that are documented to do what you need done, and you use them. But when it comes to coding web pages there is no documentation. Yeah, there's the Microsoft documentation, but it's badly written, and it reflects an implementation that nobody outside of Microsoft really understands, and that could change at any time.

Standard compliance is important. Not to make your web pages work on everybody's browser. But to make them work at all.

Re:I vote de-facto standard

[Anonymous Brave Guy (457657)]

Ask yourself this: when you write code in C++ or Java or Perl, do you blindly guess what might work? No, you look up the language features and APIs that are documented to do what you need done, and you use them.

And then you discover that no compiler on the planet actually meets the C++ standard and silly little things don't work on someone's current compiler, Java is frequently a write-once, debug-everywhere platform, and many Perl modules in CPAN aren't nearly as platform-agnostic as they claimed.

If you want things to work, de facto standards you can actually test against are worth more than theoretical, formal specifications any day. But of course, both are merely a means to an end, and useful exactly as far as they help you to achieve your objective at the time.

Re:I vote de-facto standard

[fm6 (162816)]

No, no compiler actually meets the C++ standard, at least not completely. But, unlike IE and CSS, they make a decent effort. Just think about what life would be like if there were no basic rules in writing a C++ program. When you claim that defacto standards are worth more than formal specifications, you're ignoring how much of programming comes from specifications. You don't know them, because you've never read them — they're just something you take for granted. Imagine what would happen if compiler writers didn't read the spec before implemnting a for loop, or decided to get creative with the way output strings are formatted.

Rules are necessary for any game. The fact that you can't get 100% compliance is beside the point.

Last year, not last week

[jazir1979 (637570)]

This article is over a year old, the poster got it totally wrong. Does anybody have any info on whether the comments are still relevant in the latest IE7? I have no reason to expect that they aren't, but just checking.

Re:I vote de-facto standard

[Silas is back (765580)]

I hope you are right, since "the folk" is just too lazy (or call it dumb) to download a better browser.

I'm glad the IE-bashing gets popular even amongst Win-supporters, we Mac- and Linux-users have been alone on that trip for too long.

Auto-boycot

[soloport (312487)]

Simple way to boycot:
if IE --> Download Firefox Link [mozilla.com]
else --> Welcome visitor!

Re:Auto-boycot

[envelope (317893)]

Apparently Newscloud tried that [newscloud.com], and it just made a lot of people mad.

Re:Auto-boycot

[Joe U (443617)]

That's because it's a blitheringly stupid idea.

1. Don't block your target audience.
2. Don't force them to do something they don't want to.
3. Don't try to fragment the web, it won't work anyway.

If they want to use a broken browser, have a popup window say 'your browser is broken, use firefox', and that's it, end of story.

Your end users DO NOT CARE about your personal crusade to rid the Internet of poorly designed browsers. Really, they don't.

Re:I vote de-facto standard

[eln (21727)]

How about publishing Windows without a browser and allowing OEMs to choose what browser to bundle? Most people are going to be getting Vista bundled with their machines anyway, your average person doesn't upgrade their OS unless they're upgrading their computer anymore.

While some OEMs may choose to bundle IE7 anyway, I think that if Microsoft is barred from any reprisals, most OEMs are sick enough of Microsoft's pressure tactics over the years that they may choose to bundle something else, with the most obvious winner probably being Firefox (since it's the only other browser most people would have heard of).

Re:I vote de-facto standard

[Deathbane27 (884594)]

Bundling IE doesn't prevent OEMs from doing their customers a favor by installing Firefox and making it the default browser. There's no good reason not to bundle it.

Plus, I'd rather be able to download and install Firefox on a newly-built computer using IE, than have to download it from another computer and copy it across the network or burn it onto a CD. And what if I don't have access to another computer when build time arrives?

Not having a browser installed = pain in the ass to get one installed = bad idea.

Re:I vote de-facto standard

[99BottlesOfBeerInMyF (813746)]

Bundling IE doesn't prevent OEMs from doing their customers a favor by installing Firefox and making it the default browser. There's no good reason not to bundle it.

The following are incentives not to bundle Firefox:

• It costs money and effort to add another browser and keep it up to date.
• It increases support costs to support multiple browsers and IE can't really be removed.
• Not working with IE only Websites increases support costs.
• Kickbacks from IE toolbar spyware would no longer provide extra money.
• Installing Firefox as default may anger MS, who has them by the short hairs with differential pricing of Windows.

Plus, I'd rather be able to download and install Firefox on a newly-built computer using IE, than have to download it from another computer and copy it across the network or burn it onto a CD.

To be in technical compliance with the law, MS would have allow OEMs to place Firefox or another browser on the install disk. Even if they don't OEMs can include an install disk for the browser. You can use the old version of the browser to download a newer copy of whatever you want.

Re:I vote de-facto standard

[Adam9 (93947)]

Considering the support costs that go into identifying and removing spyware, Firefox may be a less costly alternative.

Re:I vote de-facto standard

[NoMoreNicksLeft (516230)]

Yes, because M$ staffs a free tollfree IE hotline to support this stuff. *lmao*

There may or may not be good reasons to OEM firefox onto the machines, but this isn't one of them.

Re:I vote de-facto standard

[Ucklak (755284)]

I am really unaware of any 'support' issues that Firefox has. Notepad works and works well. Wordpad works almost as good. Firefox works way better than IE and I've never had an issue with notepad or Firefox.

IE developers use Firefox themselves

[theGeekDude (905574)]

Well, the IE developers use Firefox themselves anyway, so didnt bother putting in full support for CSS. After all it wont make any changes to their 'default' browsing experience....

Re:IE developers use Firefox themselves

[WaZiX (766733)]

I use Firefox to design websites and imagine a web with a multitude of CSS and PNG transparency... Then of course comes the time when I open it in IE 6... And I start hating making webpages again. Thank you Microsoft, really.

Re:IE developers use Firefox themselves

[loquacious d (635611)]

The "IE7" Javascript library [edwards.name] by Web guru Dean Edwards has helped me a lot with the IE6 blues.

It allows IE6 to render transparent PNGs (using ActiveX[?] hooks built-in to IE that allow it to render 8-bit transparency, but is mysteriously not enabled for PNGs by default) and programmatically alters the DOM and parsed CSS to enable complex subselectors and a smattering of CSS2/3 selectors as well (including fixed background positioning! [edwards.name]). It adds ~20K to pages using it, but it's a one-time cost as IE caches Javascript.

It's not a magic bullet, and sometimes causes issues itself, but definitely worth a look. Cause no one likes hacking their carefully-constructed divs back to tables, just to support a broken POS browser. (I also enjoy the irony that third-party Javscript hackers seem to be able to make more progress with IE's CSS compatibility than the actual IE team.)

The Percentages

[neonprimetime (528653)]

CSS 2.1 standard support:
IE 6: 52%
IE 7: 54%
Firefox 1.5: 93%
Opera 8.5: 93%
Opera 9: 96%

Ok, so I agree that the numbers seem to be good estimates, about right. But how on earth do they actually come up with these percentages? Is is a simple cumulative count of all css tags and attributes that work vs. don't work? Or do some have more weight than others? Seriously, they seem like fabricated numbers ... just like the /. article earlier today about how wide the universe is.

Re:The Percentages

[Reverend528 (585549)]

These numbers are based on web developer usage.

52% use IE 6
54% use IE 7
93% use firefox 1.5
93% use opera 8.5
96% use opera 9

As a result, most web sites (96%) look good in opera 9, making it the most compliant browser. Unfortunately, the other 292% of web sites look pretty bad in it.

Re:The Percentages

[RonnyJ (651856)]

Here's the link that goes over all the figures: http://www.webdevout.net/browser_support_summary.p hp?uas=IE6-IE7-FX1_5-OP8-OP9 [webdevout.net]

Re:The Percentages

[neonprimetime (528653)]

Thank you for your link. I found another link [webdevout.net] from your link which answers my question from above.

Re:The Percentages

[neonprimetime (528653)]

91.42% of all /. postings are not worth replying to

Boycott

[kevin_conaway (585204)]

Boycott I.E.? How are people supposed to do that? Just code to the standards and screw the users?

Most users don't care about your ideology or standards. Some of them aren't even aware that there are other browsers, much less why they would want one. If your site doesn't work, they'll just move on to one that does, not complain to Microsoft that xyz.com doesn't render properly.

Re:Boycott

[LWATCDR (28044)]

Sad but true.
I would vote for people recommending FireFox or Opera on every website. Maybe adding functionality for standards compliant browser that IE lacks.
The main thing is NO IE ONLY WEBSITES.
Don't make them and don't use them.
Yes sites need to support IE but they better support browsers that support standards just as well if not better.

Google is your friend (maybe)

[Jetson (176002)]

When Google announced that they were going to start offering an alternative search for blind people that rates sites based on how well they comply to the W3C usability standards, I really thought they might follow up with a search engine that rates the results according to general standards compliance. I'd love to see "works in any browser" sites on the first page and "IE-only" sites on page 10.... Suddenly all of those commercial sites would have an incentive to make their sites work instead of just making them flash-y.

Re:Google is your friend (maybe)

[doodlebumm (915920)]

If both Google and Yahoo did this, there would be massive web development work going on, and there would be a number of advertisers on both sites that would start to pull there advertising dollars away. I think that it would be very good to just mark sites initially and notify all their advertisers that they are "going to start to lower the search position of any non-compliant site as of *some-date-here*". This would strengthen their position to be able to do it, and not piss of their paying customers.

Re:Boycott

[the_duke_of_hazzard (603473)]

BOC Customer: "Why does my site not work in the new version of IE?" Me: "Well, Mr Big, I'm afraid Microsoft does not conform to the CSS standard, so we're boycotting them." Customer: "What the fuck are you talking about? Make my site work or you're fired." EOC

Re:Boycott

[Anitra (99093)]

If your site doesn't work, they'll just move on to one that does, not complain to Microsoft that xyz.com doesn't render properly.

If you're lucky, they'll complain to someone at your company that the site doesn't work...

As a web developer, I can't afford to ignore IE. It is what 95% of my clients use to review their sites. "But it works in every other browser!" won't encourage them to keep their business with us.

As a website visitor, though, I use Firefox and Safari. And I complain to the webmaster@blah of any site I that tries to force me to use IE.

Re:Boycott

[kevin_conaway (585204)]

Maybe you weren't around then, but it didn't bother people one bit to put "Best viewed in Netscape" or "Best viewer in IE" on their site.

You're referring to the golden era known as HTML 3.2?

"Best viewed in any W3C compliant browser"

Thats the problem, none of the browsers fully implement any of the standards. Some are just better than others.

De-facto standard not difficult for them

[SirTicksAlot (576078)]

It will not be MS that will make it the de-facto standard, but the people that code websites. Most commercial websites "code for IE" only and therefore force it's customer base to have IE wether they want it or not. The only workaround is to not use that company's service. But then again the people that actually use these services may not have a say as to which services they use because these services are mandated by the companies they work for.

Hopefully this will change soon.

Re:De-facto standard not difficult for them

[Tokin84 (919029)]

Hopefully places will stop coding for IE since they dropped Mac support. While the Mac user is not the biggest user, it is a percentage, and coding to IE will certainly remove their ability to use the site. Just stick to open standards... is it really that hard?

Don't ask

[Bloke down the pub (861787)]

Don't ask what Microsoft can do for IE7; ask what IE7 can do for Microsoft.

200...5 article?

[cpct0 (558171)]

Unless they are mistaken, this is a 2K5 article. And it talks about the beta 1 release, I got beta 3.

Now on the topic of better CSS, I think IE7b3 is better than what is advertised in that article. It's still far from perfect though.

"Embrace, Extend, Extinguish"

[Futurepower(R) (558542)]

"Is Microsoft purely incompetent and tone-deaf to customers - or simply counting on IE's non-compliance remaining a de-facto standard?"

Microsoft's business model is heavily dependent, not on actually giving customers what they want, but on tricks like "embrace, extend, extinguish". Microsoft will make more money if everyone follows Microsoft's non-standard way of doing things, because then everyone will need Microsoft software to see web sites.

If it weren't for the fact that it is temporarily possible to trick users who have little technical knowledge, Microsoft might be only barely profitable.

--
Will the violence of the U.S. government will end the 3,000 years of violence in the Middle East, or increase it?

Why don't boycott Vista all together

[guruevi (827432)]

Really, they made a good step in the direction with W2K and later with XP it was nice for end users and W2K3 is bearable but wtf are they doing with Vista. There is nothing exasperatingly new (like usage of the NT kernel in W2K) or a breakthrough in GUI (as with the speed of the GUI in XP) or a sysadmin-friendly environment (as in W2K3). In fact, nothing has really changed, a little GUI painted on but that's it.

IE7 still not W3C compliant or anywhere near there, still giant loopholes in the OS. Still using NTFS instead of the promised WinFS.

I was really (as an MS hater) looking forward to maybe a change within Microsoft since WGates left (and we all know a lot of work goes before the actual announcement) and Vista coming out and having promising features announced, but I can't see anything of that in their new OS.

As for a change, Stevie is announcing stuff at some convention and I am astounded. I mean, I didn't know they could do a lot more improvements in 10.4, but look at the Leopard Sneak Preview and a versioning file system and all kinds of other neat stuff... and that's right after a devving freeze in Vista which was supposed to copy some neat features out of OS X 10.3, maybe even 10.4, heck they could even copy stuff out of KDE for all I care, it still look better.

Microsoft (Gates or Ballmer, whoever has the power): I am very disappointed in you guys. I work in a mixed environment (Linux, Windows, Mac) and I have heard things in that my company (which has a bigass license with you) moving to Mac's for some non-critical users (that only need Office and to surf the intranet). If Apple pulls it off and actually builds in Win32 support in their OS, you are going to become just another SCO within a few years.

What are they doing for IE7?

[Red Flayer (890720)]

I'd rather know what they would do for a Klondike Bar.

Hopefully nothing like what is said here [hastypastry.net]. Warning: there there be crappiness.

Paul Thurrott?

[R.Mo_Robert (737913)]

In the linked article, he describes CSS as "an HTML-like technology that Web developers use to create Web sites." That's really a stretch, especially on a site like Windows IT Pro. (Couldn't he have said, for example, that it's used to style pages?) But I digress.

In any case, he can complain about IE being stuck in the 90's all he wants--I get as frustrated with it as the next Web developer--but has anyone looked at his site (or Windows IT Pro, for that matter, except I doubt he has much control over that one)? It's a mess of tables, inline Javascript and CSS, and it doesn't even have a DOCTYPE. And he's complaining about standards? IE's buggy rendering and the compatibility mode in Firefox and other browsers is probably the only thing holding that site together.

The article reads like just another attempt to bash Microsoft. It's even a bit hypocritical (see my last paragraph)...

ACID2 - Whoopdeedoo!

[kthejoker (931838)]

Seriously, they're complaining about the Acid2? The most irrelevant web standards test ever devised?

Seriously!?

IE7 fixes the Holly Hack, the box model, PNGs, the pixel jog, the double margin float, child selectors, position:fixed, the XMLHttpRequest object, XML degradation, the phantom box, percentage vs. auto, the PEEKABOO bug (Oh My God - line-height bug, too!), EMACScript degradation ...

IE7 is waaaaaaaaaaaaay closer to Firefox and Opera than IE6. And because they have a new product, they're going to work harder on CSS2.1 for the next year while they claw their way back into their 90+% market share.

I could honestly care less about ACID2 compliance, and the people who do are impractical pedants. ESPECIALLY when IE6 fails so many more basic standards tests than ACID2, all of which IE7 is fixing.

It is like complaining that you passed calculus without knowing how to use a slide rule. Ridiculous.

Re:ACID2 - Whoopdeedoo!

[John Fulmer (5840)]

Fixing long-standing bugs =! Standards compliance.

jf

Re:ACID2 - Whoopdeedoo!

[Bogtha (906264)]

IE7 fixes the Holly Hack, the box model, PNGs, the pixel jog, the double margin float,

All of these are bugfixes, not additional support for CSS.

child selectors, position:fixed,

Yes, these are improvements to CSS support.

the XMLHttpRequest object,

This is only part of draft specifications at this stage.

XML degradation

This is a workaround for proprietary behaviour that gives false positives in Internet Explorer 6. Doctype switching isn't part of any specification, it's intentional misrendering. Not to mention the fact that it wouldn't even be a problem if Internet Explorer supported XHTML in the first place.

the phantom box, percentage vs. auto, the PEEKABOO bug (Oh My God - line-height bug, too!),

More bugfixes, not additional support.

EMACScript degradation ...

What are you referring to? They haven't made any changes to their JScript engine, which is their implementation of ECMAScript.

All in all, I see a lot of bugfixes, but hardly anything in the way of adding missing support for parts of CSS. Sure, they added selectors, but they missed out tables and generated content, which are huge parts of the specification. Sure, they added a workaround for people using faux XHTML, but they didn't actually add XHTML support. And I don't know what you mean by "ECMAScript degradation", but they still have a non-standard event model instead of the DOM event model.

IE7 is waaaaaaaaaaaaay closer to Firefox and Opera than IE6.

Come off it. Bugfixes are not a great leap in functionality. Sure, it's great that we finally have them, but to characterise this as closing the gap between the browsers in any meaningful way is exaggeration beyond belief.

I could honestly care less about ACID2 compliance, and the people who do are impractical pedants.

Er, some of the things that Acid2 tests for are things you are describing as fixed in Internet Explorer 7, so obviously some of the things in Acid2 are important to you.

And, wearing my impractical pedant hat, I have to point out that you are saying that people who care about Acid2 less than you are impractical pedants, which makes no sense.

Re:ACID2 - Whoopdeedoo!

[nine-times (778537)]

It is like complaining that you passed calculus without knowing how to use a slide rule. Ridiculous.

Not to be pedantic, but isn't it more like complaining that someone passed calculous when they have shown an inability to pass a calculus test?

ACID2 is not the end-all and be-all of web standards compliance, but it does give an indication of how well a browser is rendering certain kinds of CSS with reference to the W3C standards. It was devised on feedback from web developers to be a collection of common rendering inconsistencies between the major browsers. It's not completely meaningless.

Re:ACID2 - Whoopdeedoo!

[Bogtha (906264)]

Unfortunately, from what I've seen, they have added more hacks to the clunky rendering engine from IE5 (or earlier), instead of developing a new rendering core from scratch.

Backwards compatibility with all the crappy proprietary behaviours of older versions of Internet Explorer is pretty important to Microsoft, which is why they are still using their older rendering engine instead of replacing it with something better. They can't make big changes because they are afraid they'll break things.

Internet Explorer 8 is where you're likely to see a change like this. From what they've been saying, I think it's likely that they'll not add a further doctype switch, but implement a new rendering engine for XHTML only. Everybody using text/html will be stuck with Internet Explorer 7-level support for CSS, and everybody using application/xhtml+xml will get the new rendering engine. This has the added advantage of zero regressions - so Microsoft won't have to worry about backwards bug-for-bug compatibility.

Unfortunately, to do this, they actually need to implement XHTML...

MS has specifically stated that IE7 will not support the application/xml+xhtml mime type. This is a simple thing that most people overlook the importance of.

No, it's not. I know it looks quite similar when you are writing it, but supporting XHTML isn't just a case of adding "application/xhtml+xml" to the list of media types that get chucked through the HTML rendering engine. Apart from the obvious fatal-error-on-malformed-documents behaviour, there are changes to the DOM, changes to CSS, changes to page structure, and so on. For instance, the following code means different things in HTML and XHTML:

<table><tr><td>...</td></tr></table>

In HTML, this code creates four elements. In XHTML, it creates three elements.

There's all kinds of subtle ways in which XHTML differs from HTML, and if Microsoft don't get it right, it's going to cause a whole load of problems further down the line. XHTML is a golden opportunity to leave cruft like doctype switching and stupid CSS bugs behind once and for all, and if Internet Explorer 7 includes premature broken support for XHTML, it will be a squandered opportunity, and it will cause all kinds of problems further down the line.

This article is a year old

[epohs (775630)]

Am I taking crazy pills, or is this article not over 1 year old? [ August 02, 2005 ]

For the best browser experience...

[IGnatius T Foobar (4328)]

...go to http://www.ie7.com [ie7.com]

(Seriously. The best browser is there.)

Woof woof

[davmoo (63521)]

Whether you all like it or not (and note that I use Firefox myself), Microsoft Internet Explod...um...Explorer is the proverbial tail that wags the dog. In the minds of the great majority of computer users, Microsoft *sets* the standard, not breaks it, and you will not convince them otherwise. You can whine and moan all you want, but I got 5 bucks that says when IE7 rolls out, we start seeing a new round of sites that work *only* in IE7, and when you complain the response will be words to the effect of "get a real browser like everyone else uses".

What we need is a Gecko ActiveX control

[Schraegstrichpunkt (931443)]

I would love to see an automatically self-updating Gecko ActiveX control. Any IE user who visits my sites (or dozens of other sites that mandate it), would simply have to click "Yes" once (ever), and then the user would be using the newest version of Gecko to render the pages automatically.

IE could be effectively marginalized that way.

How to defeat IE

[onlyjoking (536550)]

The crucial argument against IE is its terrible CSS support but it's very difficult to get this across to ordinary users. Here's my suggestion. Create your site with as many features as possible which fail in IE but render perfectly in Firefox, Safari etc. Next insert Javascript or CSS IE browser detection into your home page which inserts into the IE rendered page something along these lines:

This site will display better in a browser which supports web standards [webstandards.org]. Here's an example [homepagescreenshot.com] to show you the difference.

The example is a link to a screenshot of the home page rendered in Firefox and a link to the Firefox download page should also be added. This way we don't lock out IE users but make IE's shortcomings as obvious as possible thus dispelling the pernicious M$-cultivated illusion that sites with IE workarounds are the standard. For this to work it needs to be a standard response developed by the web standards project so that it becomes familiar when users see it on different sites. The only way to defeat M$ is to play them at their own game.

IE7 CSS Job Interview

[Temujin_12 (832986)]

I interviewed for a IE7 CSS job at Microsoft about four months ago. Coming from a web development background, I was curious as to how they would present their goals/problems with meeting CSS standards. I was well aware of the "code it to standard, view in IE, and cry" web development cycle.

One of the team leads (sorry don't know how high up of a team lead he was) actually said that often when people say IE is rendering something incorrectly it is actually IE that is doing it correctly while all of the other browsers are rendering it incorrectly. I could tell he was looking at how I would respond to that statement. I just sat there and didn't move. While in some cases that may be true, I knew that was an arrogant lie, and was just enough for me to stop caring about the interview. Needless to say, I didn't get the job. Fortunately, I had already interviewed for another job, which I've since been hired at, which is much better.

Two points here:
1- With team leads holding that kind of attitude (and touting it during interviews), no wonder IE is the quagmire it is. They're more used to making standards, not adhering to them.
2- Yes, recent college CS grads can find a job! I actually had 2 1/2 offers after only 4 interviews. Just develop your skill set (more than what they teach you in class) and learn how to communicate in *English* not just C, C++, Perl, etc.

IE's most egregious offense

[phiber_phreak (719398)]

Does anyone know if IE7 will fix the absolute worst behavior in IE -- closing TCP connections with RST rather than FIN?

This bad behavior:

--exists in IE6 and earlier

--violates RFC 793 sections 3.4 and 3.5

--ties up LOTS of memory in zillions of stateful devices (firewalls, VPN gateways, L4 and L7 load balancers, and on and on)

--does not belong to the MS TCP/IP stack, since other applications (eg, telnet) close connections properly

I haven't played with IE7 yet. Someone please tell me MS has finally addressed this abomination.