Microsoft Sends Broken Stylesheets to Opera

An anonymous reader writes "The Register has a story that the MSN homepage serves a different style sheet to the Opera web browser that makes Opera appear to be broken. Is this deliberate or a mistake? Who can possibly say? Opera's own take on the situation can be found here." This is not the first time.

Comments

No fear of prosecurion, no problem!

[Sun Tzu]

It's back to the bad old days at Microsoft... Sounds a lot like how they killed DR-DOS, but on a smaller scale.

Send us your Linux Sysadmin articles.

Re:No fear of prosecurion, no problem!

[Strange Ranger]

>back to the bad old days at Microsoft

Drat, I must have missed the good days.

Re:No fear of prosecurion, no problem!

[ichimunki]

Forced... because I have a hotmail account?

But is someone forcing you to use Hotmail?

Re:No fear of prosecurion, no problem!

[critter_hunter]

MSDN has a similar behavior. I don't give a shit about MSN, but I needed to download the DirectX 8.1 SDK (to use OGRE) the other, and it was hell. I fact, I needed to identify as Mozilla 5 to see more than a few unrelated links on this page (try it if you have Opera. Change your identifier and reload the page)

Standards schmandards.

[Boogaroo]

Seriously, isn't this why the W3C tries to make people follow standards? So it doesn't matter what browser you use, it should all work?

Anyone, including Microsoft, who writes a site that serves seperate pages to different browsers is doing a disservice to the public.

Re:Standards schmandards.

[u38cg]

Well, not neccesarily. I don't think it's unreasonable (as a provider, not a web zealot) that a server doles out pages that are renderable. Where it's possible to predict what needs to be changed to get a browser to render that page properly, you might as well do it. Of course, it does have the potential to be abused.

My browser is set to send nonsense as its id strings; it doesn't seem to do my surfing experience much harm.

Re:Standards schmandards.

[Enahs]

Um, I suppose you didn't read Opera's take on this; MSN's webserver(s) uses a special broken CSS for Opera. When using the stylesheet intended for IE6, Opera displays the page just fine.

Re:Standards schmandards.

[Anonymous Coward]

> Seriously, isn't this why the W3C tries to make people follow standards? So it doesn't matter what browser you use, it should all work?

How the HELL did this get modded to 5?!? RTFA, the problem is MSN sending a perfectly-compliant, but deliberately flawed in values, CSS style sheet *only* to the Opera 7 browser. Note that the sheet values were chosen to instruct O7 to misrender the pages. Nothing the W3C can do about this, standards compliance wasn't the problem.

Re:Standards schmandards.

[questionlp]

W3C calls their CSS and HTML specifications as "Recommendations" rather than "Requirements" or "Standard", per se. Instead, they provide specifications on how user agents (be it browsers or cell phones) are recommended to follow.

Of course, it would be lovely if all browser makers were to forced to follow the recommendations down to the nitty gritty, but even the recommendations don't always provide strict requirements on how a property or class should be rendered.

The fact that Microsoft is pushing out (delibrately) a broken style sheet is just wrong.

Re:Standards schmandards.

[rknop]

Anyone, including Microsoft, who writes a site that serves seperate pages to different browsers is doing a disservice to the public.

While I agree with the philosophy, unfortunately it's unrealistic. Reason: so many browsers, worst among them Netscape 4, try to support CSS and fail so miserably that a standards-complaint CSS page is likely to be unreadable. And, unfortunately, some people still use NS4 and old versions of IE.

What I've done some places is write some SSI that detects the browser. If it detects Netscape 4 or lower, or IE ... probably 4 or lower, I forget at the moment ... it sends a "dumbed down" style sheet that will present only a faint echo of the layout of the page, but which will leave the text readable. Any other browser, you get the normal "standards compliant" style sheet. Note that here I am sending specific style sheets for specific browsers-- but I assume that any version of Opera, and any version of Netscape or Mozilla 5 or greater and any recent IE and any other browser that may come is standards complaint.

-Rob

Re:Standards schmandards.

[Fembot]

Actualy thats an interesting point. Should I bother to make a special version of my page for older browsers (Like NS4 and IE4) which have frankly awful CSS and DHTML support? Especicaly give how few people still use those two browsers. Or should I maybe have one primitive version which runs in even mosaic etc, and one which works with all the more common browsers which conform to the latest standards? And why cant there be some nice easy javascript way to determine if browsers support CSS2 and other fancy standards?

Re:Standards schmandards.

[FuzzyBad-Mofo]

Here's a tip for including CSS without having to worry about NS4 and (I think) IE4 screwing it up:
Use the CSS2 @import rule to import your style sheet.

The version 4 browsers will ignore this tag, therefore you don't have to worry about crashing NS4 with your perfectly valid CSS.
Example:

<style type="text/css" media="all">@import url(main.css);</style>

Opera should respond

[pubjames]

Opera should respond by automatically translating any page on the Microsoft web site into German and back again with Babelfish.

Re:Opera should respond

[evilviper]

"Welcommen to SmallDelicate.com"

Re:Opera should respond

[Anonymous Coward]

Pornoliz it http://www.pornolize.com/

Standards and lies

[Andy Social]

I love the bald-faced lying that MS pulls out for this behavior. "We're heavily invested in following standards." or "We're trying to produce the best site for all viewers." Yeah, right. Explain why there would be any reason at all to force every child entity 30 pixels to the left of its parent. For that matter, why does MSN still use the tired old hack of sending different pages to each browser? I don't need 4 versions of my site to handle every viewer. Amazing.

Re:Standards and lies

[Andy Social]

They obviously do something wrong. Check out the spiral demo in various browsers. The ones that actually follow the CSS standards can render it perfectly. Want to guess which ones render it well? Mozilla-based browsers, and Opera 7 are the only Windows browsers to work correctly. That means the only major browser that does not follow the standards is the most popular one.

Extend and enhance (also known as extend and extinguish) is not the way to go about making a standards-based system. There are standards. They are not mutable, they are not extensible (except where stated).

In this particular case, they purposely serve a messed-up CSS stylesheet to Opera. If you browse with CSS turned off, the site looks fine. So, regardless of their adherence to standards (which is not very good), they purposely try to monkeywrench Opera. That's the point of this story, really. Every page should render identically on every browser. All information should be visible on every browser. Purposely hiding your text under a graphic is unacceptable behavior.

Re:Standards and lies

[curtisk]

No it doesn't......check the menu bar on left....its not transparent to the shell in the background...and the shell in the background of the blue area is fixed and moves with the scrolling of the page, it shouldn't....get Opera or any of the browsers that the page mentions and you'll see the difference. I just looked with IE6 /WinXP Pro and Opera7 / WinXP Pro and there is a difference
Yes, you can see the page, but it's not doing WHAT it SHOULD.

Re:Standards and lies

[pacc]

Anyone using CSS at all would be aware of bug in
every browser, even Opera: Real-world example

There are no chance that they would have gone through the process to server different code to different browsers without testing it out afterwards.

IE on macintosh is reported to work very good, and there are XML engine updates for Windows to download. This all points to the fact that microsoft is very capable of actually supporting the standards, but we also know that standards would give people no reason to prefer IE over some other browser.

Sites del. diff. content to different browsers...

[eddy]

..is fucking irritating. Don't mess with it!

NO! It it not necessary. It just makes things worse in the long run, so if you're doing this _you're_ part of the problem, so don't complain about how you have to treat browsers differently.

Sheeesh. Write to the standards, not browsers.

(And no, this isn't "insightful", it's totally _obvious_ to anyone with a clue)

Oddity to me

[unterderbrucke]

Why did they pick Opera, and not Mozilla or Netscape, not to mention Safari?

Re:Oddity to me

[sulli]

My theory is that they particularly hate Opera because it's commercial. To their customers (OEMs and corporate licensees) they can bash Mozilla for being open-source and therefore unreliable, Netscape for being too tied to AOL, and Safari for being too new and half-baked, but Opera proves that there's something better, separate from the OS, that people are willing to pay for, and this must really piss them off.

Of course, I am happy with Moz and never think of using MSN. But that's just me.

Won't help!!

[Baki]

Just tried www.msn.com with Opera 7. No matter what I set the agent identification to, msn keeps sending the wrong (faulty) stylesheet.

It looks like MSN uses more advanced techniques to find out what the client is than just the agent identification, in order to sabotage Opera in this case.

logically speaking....

[theBrownfury]

....this doesn't make much sense. MS makes a lot of money based on the popularity of their MSN portal. this portal links to a lot of their other properties as well and it is against their best interest to make it difficult for users with a different browsers to access this page.

one would think that since they want people coming to this page and accessing it regularly they would make it easier for them to get here.

conspiracy theory aside this doesn't make sense from a business point of view. i have a feeling this is a mistake of some sort.

Not Just opera broken

[cluge]

Older version of IE were also purposely broken in the same way; forced obelesence? As a regular Opera user I notice the same problem on some portions of the Microsoft web site as well (not just MSN).

To me this just proves that the remedy isn't working, that MS as a company prefers dirty tricks to competition and that the states that have not agreed to settlement had better press MS hard. (Wow holy run on sentence batman). It's sad that a company as successful and as full of talented people as MS has to resort to this type of behavior when a competitor comes out with a good product.

I'm reminded of a famous quote "Can't we just all get a long". I guess if your MS and you can't or won't compete the answer is no.

MSDN didn't work with Mozilla UA for a while

[Malc]

For a while I had to change my User-Agent string under Mozilla to that of IE6 under WinXP when visiting MSDN pages. Thanks Mozdev's uabar, and later Xulplanet's prefbar! Content served to Mozilla UA strings was unreadable with much of the text over-lapping. This went on for almost a year, but it seems to have been okay for about a month.

Rather coincidentally, it was fixed shortly after I filled out a MSFT survey that appeared as I tried to leave the site - I claimed I was leaving because I was fed up with changing my UA string. Of course, I'm not conceited enough to think they fixed their problems because of me :D The whole time though, I had no problem with the MSDN subscriber downloads site, which even had a message for Netscape users.

Actually

[hafree]

According to the web logs on the various sites I host, Mozilla and other Gecko-based browsers make up almost 15% of web traffic now, and Opera has a signficant enough user-base that it also makes it into the top 10 user-agents on web sites that get 1M+ hits per month from 100+ countries. I think the problem is that people need to move away from Microsoft web deveopment tools until they can learn to play nice and output standards-compliant HTML code. Ever try using the "save as HTML" feature in a Microsoft product? A 100-row table becomes a 2MB plaintext file by the time it makes it to the web...

It is pretty easy to do

[MetalShard]

I used to have different style sheets for different browsers in an effort to make my websites look good for all of them. More than once I updated some of the html and only tested the pages in IE where they looked fine (I know, I know, but programmers are naturally lazy.) It turns out the style sheets for the other browsers totally made the pages look broken. I'm not defending MS. It would not surprise me if they did it on purpose, but I am saying it is easy to do. Now I just have one style sheet and I made sure to use simpler html that would look good on all browsers. Sometimes simple is better.

www.wannabrowser.com

[KalvinB]

You can see how webpages react to various browsers at www.wannabrowser.com

I'm not going to bother posting the results here but it's easy enough to see for yourself what the differences are.

Ben

Re:Aww, come on...

[Shalda]

I remember back to about '97 or so, at the height of the browser wars, every time I went to the Microsoft web site with Navigator 3.0, Navigator would GPF. Of course, the same thing would happen if I went to the Netscape site with IE 3.0. Which was ironic, because I was going there to download Navigator.

Re:Clearly This Sucks but....

[wrenkin]

If you read the article you would see that Opera is perfectly capable of rendering the page that MSN sends to IE. If you change Opera to identify itself as IE, no problem.

The problem here is that if you've set Opera to the report the true user-agent, MSN sends a page with a broken CSS file that tells the browser to render the content so that the page becomes unreadable--Here, they set a negative margin on content in some divs so that the first couple words in any column are overlapped by the div to the left, frustrating the viewer. Even IE chokes on the page they give to Opera:

http://deb.opera.com/howcome/2003/2/msn/opera7.p ng

This is sabotage.

Read the original report here:

http://deb.opera.com/howcome/2003/2/msn/

Re:Sites del. diff. content to different browsers.

[Blimey85]

Write to the standards, not browsers.

This is fine for a personal or hobby site but for e-commerce, you need to write to users, not standards. It makes no difference to the user that your page is coded to standards if he/she can't view it. Telling them they need a different browser isn't the answer either. Showing them what they want, in a manner that works correctly with their browser, is unfortunately the best solution if you want to be profitable.

I've had to code drop down menus differently for different browsers to get things to look the same, however when I'm done, you get the exact same page, with everything the same size and in the same place in IE, Netscape, Mozilla, and Konqueror. I've never used Opera so I don't test that one, but I guess I probably should.

i dunno

[sydlexic]

but it really looks like an honest typo in the style sheet. i can understand how they'd be pissed, but did they try contacting MS to get it fixed first? the fact that the server sends a client-specific style-sheet isn't exactly damning. it's a very common (though misguided) practice.

never attribute to malice that which can be explained by stupidity.

Re:i dunno

[ink]

did they try contacting MS to get it fixed

It took microsoft SIX MONTHS to fix a one-liner that prevented Mozilla from working with Passport (buggy browser "detection" code). See bugzilla bug #141279 if you are curious. Interoperability and open standards are not placed anywhere near the top of the queue at Microsoft. In fact, the dragging of feet would point to more sinister motives... but of course there's no proof of such (without Halloween memos, at least).

Re:i dunno

[KoolDude]

...but it really looks like an honest typo in the style sheet...

The point is that the page rendered exactly same as IE provided the stylesheets are same. Unless MS thinks there is something wrong with the way IE(or Opera7) displays the page, why type out a different stylesheet and commit a typo in the process ? If it ain't broke, fix it to break it ?

Re:i dunno

[stinky wizzleteats]

but it really looks like an honest typo in the style sheet.

Really? Let's look and see...

MSIE stylesheet:

margin: -2px 0px 0px 23px;

Opera stylesheet:

margin: -2px 0px 0px -30px;

I don't know how you go about typing, but I'd have to throw silly putty at the keyboard from the other side of the room to hit the "-" key instead of the "2".

Re:i dunno

[Anonvmous Coward]

"Why would Microsoft care about Opera? Or think that misrendering the MSN home page is a good way to undermine it?"

Just wanted to say that I'm glad you took a moment to take a step back and and say "Why?" as opposed to jumping the gun and saying "Damn MS trying to enforce it's monopoly as usual". Frankly, I'm tired of the wild assumptions that MS works that way.

As an Opera user and a Windows user, you can understand that I've run across exactly this problem. I'd like to share with you a few observations I've made on this topic:

- As an Opera user, I find myself having to deal with a number of sites that just don't care about me. Having IE available as a backup is just part of my everyday Opera life. I don't see MS as being very different here. Some sites block me totally, like the site I use to send payments to my credit card.

- Because of my having to keep IE on hot standby, it doesn't even occur to me anymore to email MS (or any other site) and complain about lack of Opera testing. If they don't get feedback, they ain't gonna fix it.

- Website maintenance is a perpetual, priority based job. Often problems are ranked by how many people are affected by them. Truth be told, Opera's just not significant today in light of other things going on. I wouldn't be a bit surprised if nobody there had Opera installed. Anybody who's ever done web development at a Dilbert-esque corp would probably understand this.

- Wouldn't it be weird that MS would break Opera, but not Mozilla, Netscape, or other browsers?

- What possible benefit could that bring them? Despite my comment earlier, Opera handles the vast majority of sites(*) just fine. When I run across a site that doesn't work with Opera, it feels like the operators of the site were moronic, not that Opera is incompatible. In other words, MS's site not working right with Opera makes MS look incompetant, not Opera. * Sites that I personally have visited, other people's experiences may vary.

- MS's site is a marketing tool. Head on over there and you hear all about TabletPC's, PocketPc's, MS's latest server stuff, Windows XP, etc etc etc. Breaking their site means potentially shoo'ing off customers. I seriously doubt any PHB would want to do that.

If other Opera users share my observations, then it actually makes sense that MS just doesn't care. But the idea that they're doing it to enforce a monopoly is not so evident.

Please don't flame me for not jumping on the "MS is like OCP!!!" bandwagon. I'm just the type of person that'd rather look at all the details than try to find details to support a bias.

M$ claims W3C compat...mistakenly?

[horatio]

From the old article, one of Microsoft's marketing directors should get his facts straight:

"We supported the latest W3C standards when developing the content and services delivered from MSN," ... He added that Microsoft wants users to visit the Web site "regardless of the browser they choose."
But Visse recommended that for the best experience with MSN, customers should use a browser that tightly adheres to the W3C standard.
"If customers choose to use a browser that does not tightly support W3C standards, then they may encounter a less then optimal experience on MSN," he said.

except, that if you ask the W3C validator, it doesn't work!

www.microsoft.com
www.msn.com

Microsoft has a long history of intentionally breaking compatibility with other products to promote their own, as early as (and maybe earlier) the Windows 3.1 -> 3.11 "upgrade" which conveniently broke the diagnostic and repair software PC Tools.

Re:We need browser masking.

[rknop]

As an unfortunate side-effect, this would reinforce webmasters' belief that everybody in the world uses MSIE.

Yep. What we really need is too late to accomplish. What we really need is a protocol that forbids you from identifying which browser you are, but only allows you to specify to which standards you conform.

Then maybe webmasters would write their HTML and such the way they're supposed to, and what's more the browsers would have to really support the standards they claim to support.

But, unfortunately, that's an ideal world, not the one we live in.

-Rob

Slashdot and w3.org...

[jonr]

On (semi) related topic. Why has slashdot.org blocked validator.w3.org? Are they embarresed by the results? After all I can always do a "Save as" and then upload the page to the validator.
Pretty childish, if you ask me.
J.

Re:Slashdot and w3.org...

[Anonymous Coward]

I predict a mysterious but sudden moderation to -1 Offtopic in your future...

Re:Slashdot and w3.org...

[Nicopa]

Dialog:

Slashdot: Microsoft Sends Broken Stylesheets to Opera.
The people: So what, you send broken HTML to everybody!

Comment removed

[account_deleted]

Comment removed based on user account deletion

the reason they are targeting opera

[a7244270]

MS lost out bigtime by failing to convince the cellphone manufacturers to adopt their embedded OS - most of the bignames plan to use Simian (is that how you spell it) which uses Opera as its browser.

The reality is that most windows users will never change their browser from IE to something else, so they are not afraid of Mozilla, konq, Safari, etc.

The cellphone market on the other hand is HUGE, and given recent advances in wireless bandwith, has the potential to be highly lucrative.

More than likely its probably safe to say that a significant percentage of all web browsing in the future will be on cellphones.

They are attempting to ensure that non MS cellphones can't surf the web properly, in an attempt to make consumers prefer buying MS enabled webphones, which in turn will generate more revenue in the embedded market for them, which they desperately need.

Just my opinion tho - can never tell what does guys are up to...

Make Opera appear broken??

[MongooseCN]

If I went to MS's site and the webpage they sent was broken, I would think MS had an incompetent webmaster who didn't know HTML. I wouldn't think Opera was broken.

Re:What is the alternative?

[DickBreath]

MS are free to serve up whatever they like on their servers.

Are you sure?
Without consequence as to what?

Okay. I'll use your logic, and the same logic as some other posts here from Microsoft agents.

I want to start serving stuff from my site that takes advantage of all known exploits in IE browsers. After all, it's my site. I can serve whatever I want. It's my business.

If users don't like it, then they should use Mozilla or Opera.

If you're a Microsoft user, why would you want to come to my site anyway?

It's just an accident. Give me the benefit of the doubt.

I'll probably get modded redundant since my above four arguments have already been made.

Re:Quite the contrary

[Isofarro]

I don't know where you got this myopic view of the Web from, but it is certainly trollish from a standards POV. Obviously the technique of augmentative authoring has eluded you.

If you are creating multiple copies of resources for different user agent strings, then it is a prime indication you haven't understood the very simple concept of the World Wide Web.

For various reasons ( including access to the reading disabled) every site should, at the very least, serve a different page to pure text browsers than it does to graphical browsers.

Making a website accessible does not mean text-only. This is a myth, and a badly misinformed piece of strawman fluff. Text-versions of websites should only be a last resort, when you've reached the point where you admit your design and markup skills are inadequate to do even a competant job, let alone a good one. Accessible websites can also be well designed, there's no mutual exclusivity.

The whole *point* of identifying browsers at all is to allow the server to serve optimized pages for different browsers.

If you so strongly want to believe this nonsense, please post a reference to either a standard or recommendation that states that User-Agent is a mandatory HTTP parameter. You know as well as I do that User-Agent strings are optional, and relying on them to determine presentation is so typically short-sighted that its now laughable.

You cannot succeed over the medium to long term adopting a browser-sniff route. It is folly.

No problems with Mozilla as Opera

[Midnight Thunder]

I just made my install of Mozilla pretend it was Opera, by adding the following to the user.js file:

user_pref("general.useragent.override", "Mozilla/4.0 (compatible; MSIE 6.0; MSIE 5.5; Windows NT 5.1) Opera 7.0 [en]");

I restarted Mozilla (1.3a), checked the about page (it shows the user agent) and then visited the MSN page. The page showed up fine. I thought that maybe that maybe MS had changed the CSS. I downloaded the style sheet in Mozilla and saw the -30 there. From what I can tell Mozilla must have a check to ensure that text does not appear outside of the cell, not matter what the css indicates. If Mozilla can do it, then the guys over at Opera can do it too.

Note - I am not saying that this clears MS, as any well implemented web site should only need one version of any page, unless they have localization. What I am saying is that this is a fixable issue on the part of Opera.

Re:Clearly This Sucks but....

[Okojo]

Actually I spoke too soon and am going to take back what I said. I loaded up the page in Opera 6.0 and the margin -30 is supposed to fix a bug in Opera 6.0's rendering of lists. In fact, it's the very same problem I ran into while designing some webpages a few weeks ago that annoyed me to no end. Basically, Opera 6.0 indents list items by about 30 pixels to the right, unlike other browsers. Thus that -30 value is there to correct that problem. Opera 7.0 doesn't exhibit that tabbing effect (thus consistant with the latest IE and Mozilla browsers). Apparently MSN is serving Opera 7.0 the same CSS sheet as Opera 6.0 even though 7.0 works best when it's served the same style sheet as IE. Thus, saying that this problem is browser sabotage is too strong of an accusation.

Opera uses a bad disguise

[Rui del-Negro]

Opera always includes "Opera" in the ID string (ex.: "Mozilla/4.0 (compatible; MSIE 5.0; Windows 2000) Opera 6.05 [en]"). Which sort of defeats the whole point of identifying as a different browser. It'll only fool scripts that first check for "MSIE" and, if they find it, don't even bother looking for "Opera". All other scripts will still see it's Opera.

RMN
~~~

Truly unrecognized browsers are treated similarly

[Noel]

I messed around with a few other UserAgent strings, and it gets a little clearer:

"Mozilla/4.0 (compatible; MSIE 6.0; MSIE 5.5; Windows NT 5.1) Opera 7.0 [en]" -> site.css

"Mozilla/4.0 (compatible; MSIE 6.0; MSIE 5.5; Windows NT 5.1) Oprah 7.0 [en]" -> site-win-ie6.css

So far, exactly as reported in the article

"Mozilla/4.0 (compatible) Oprah 7.0 [en]" -> site.css

"Mozilla/4.0 (compatible; MSIE 5.5; Windows NT 5.1) Oprah 7.0 [en]" -> site-win-ie5.css

"Mozilla/4.0 (compatible; MSIE 6.0; MSIE 5.5; Windows NT 5.1; Opera 7.0) Oprah 7.0 [en]" -> site.css

It's pretty clear what they're doing:

if string contains "Opera", use site.css
else if string contains "MSIE 6.0", use site-win-ie6.css
else if string containse "MSIE 5.5", use site-win-ie5.css
else use site.css

In other words, it doesn't matter what Opera claims to be compatible with - they always get the default sheet, just like a completely unrecognized browser does.

I'm trying to apply Hanlon's Razor here, but it's hard...

this is typical, the .NET framework does this, too

[sirshannon]

My biggest complaint about the .NET framework:

the .NET framework does a check to see which browser you use and then sends formats aspx pages for the capabilities of that browser. So if you use abs positioned divs, you'll get those for modern browsers but Netscape 4.7 (for instance) will get the same page (theoretically) but formatted via tables. This is great, if only MS were honest about it.

I constantly have to hard-code formatting for controls because MS treats Netscape 6 as a 'down-level' browser and doesn't bother sending out certain formatting tags. So some pages look bad in Netscape 6, the reason behind it would be that the formatting tags weren't sent out because Netscape doesn't support them, but this is false because when I add them by hand, netscape handles them fine and my pages look the same in both browsers.

I have to believe that MS does this so people say "this page looks like azz in Netscape" and assume that it's Netscape's problem.

the framework has been out for too long and this is still not fixed, so I can not believe that it is an honest or innocent mistake.

I doubt it

[Wraithlyn]

They specifically designed their web site to send a different style sheet, (and '30' is not a typo.. '33' or '34' or something I'd believe) AND a larger page with less content, JUST to Opera. That seems pretty far from "an honest typo". This is MSN's HOME PAGE. You don't think they know what it looks like in different browsers? I work at a 4 person company, and we know what all our websites look like in IE, Moz/Netscape, and Opera. Furthermore, they have a motive to make it look better in IE, and they've shown in the past OVER and OVER that these kinds of underhanded tactics are their bread and butter. Someone at MS knew about this, and also knew it could never be proven in court.

By the way, the full quotation is:

"Never ascribe to malice, that which is adequately explained by incompetence"
- Napolean Bonaparte.

I think one of Microsoft's new unwritten policies is "When accused of malice, always hide behind incompetence".

"No no... we'd love to, but we simply CAN'T remove IE from Windows." Sound familiar?

Breaking my own sites.

[JWSmythe]

Honestly, this is not an effort against Opera. If I choose to break my own site, so be it..

In my industry, just about every site does video of some sort. There's always some group that feels they were intentionally blocked because of whatever reason. I've seen sites that stream exclusively Windows Media, and some that use propriatory plugins like "Emblaze".. Some were using the Netscape "Push" method (send a multipart header, and then send a new mime delimiter between frames). Netscape "Push" doesn't (or didn't) work with MSIE.. Windows Media doesn't work with Linux. (with a few exceptions).. Something doesn't work with something else.

If I choose to make my site not work with MSIE or Netscape, and only let Opera viewers see it, well, it's my site.. If Slashdot decides tommorrow that they like a feature of Mozilla 9.999, and it doesn't work with any other browser, including MSIE, how many of you are going to be bitching for MSIE compatability?

I'll get a bunch of comments back "Microsoft Sucks", but I'd *LOVE* it if they'd put the REMOTE_USER_AGENT string beside your name in the comments.

For those curious, mine is:
Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.3a) Gecko/20021212

I'm not defending Microsoft. It's shitty that they did it, but honestly it's their site. Try doing a Windows Update from Netscape, that doesn't work either.

Want more fun? Try installing a nice fresh copy of an older Microsoft OS (say WinNT 4.0), and get yourself up to day.. Years ago, they broke the Microsoft pages, so you couldn't get the updates. But I can't say that I've ever seen a /. story on that.

Where I work, we try our best to make our pages render correctly on our machines.. That means, keep everyone in the office happy, and hopefully it will make the majority of our customers happy. We have enough varity by choice to keep things interesting. here's the short list of the browsers we use:

Win98/Win2k/WinXP:
MSIE 5.0 -> MSIE 6.1
Netscape 4.7 -> Netscape 7.01
Mozilla 1.1 -> Mozilla 1.3a
Opera (unsure of version)

Mac: OS/9, OS/X
MSIE (unsure of version)
Netscape (unsure. various versions)
Mozilla (unsure. various versions)

Linux: (Slackware)
Mozilla 1.1 -> 1.3a
Netscape (various)
Konqueror 3.0.1

But sure as hell, we'll have some sort of rendering problem on some browser, and someone will scream that there's a conspiracy against them specifically..

Our sites don't require any special browser. They all work. We don't know of any compatability issues right now, but I'm sure someone will find that Konqueror v1.0 won't work with a particular page, if they try hard enough. Our site has average users browsing. Some advanced users, lots of regular users..

In the last 24 hours we had 17,017 different REMOTE_USER_AGENT strings sent to one of the servers, in 1,949,023 requests from 116,273 unique IP's.. If I take the list and:

cat list.txt | cut -f 1-3 -d ";" | sort | uniq -c > work.txt

less work.txt

Here's the top 10 results:
474500 "Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1
317359 "Mozilla/4.0 (compatible; MSIE 6.0; Windows 98
140794 "Mozilla/4.0 (compatible; MSIE 5.5; Windows 98
91425 "Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.0
66331 "Mozilla/4.0 (compatible; MSIE 5.0; Windows 98
31072 "Mozilla/4.0 (compatible; MSIE 5.5; Windows NT 5.0
29963 "Mozilla/4.0 (compatible; MSIE 6.0; AOL 8.0
26778 "Mozilla/4.0 (compatible; MSIE 5.01; Windows NT 5.0
25426 "Mozilla/3.0 (compatible
20841 "Mozilla/4.0 (compatible; MSIE 5.01; Windows 98

And in comparison, we'll look at some other top 10's.. Here's the top 11 Linux clients (11, because the first Opera was #11)

grep -i linux work.txt
1563 "Mozilla/5.0 (X11; U; Linux i686
387 "Mozilla (X11; I; Linux 2.0.32 i586
161 "Mozilla/5.0 (compatible; Konqueror/3; Linux
145 "Mozilla/4.7 [en] (X11; I; Linux 2.2.13 i686
96 "Mozilla/5.0 Galeon/1.2.7 (X11; Linux i686; U
72 "Mozilla/5.0 Galeon/1.2.5 (X11; Linux i686; U
67 "Mozilla/5.0 (X11; U; Linux i586
64 "Mozilla/4.78 [en] (X11; U; Linux 2.4.7-10 i686
56 "Mozilla/4.78 [en] (X11; U; Linux 2.4.9-34 i686
46 "Mozilla/5.0 Galeon/1.2.6 (X11; Linux i686; U
39 "Opera/6.11 (Linux 2.4.2 i386; U

And the top 10 Opera clients.

127 "Opera/6.01 (Windows 98; U
118 "Opera/6.05 (Windows XP; U
104 "Opera/6.05 (Windows 2000; U
74 "Opera/7.01 (Windows NT 5.0; U
72 "Opera/6.05 (Windows 98; U
60 "Opera/6.0 (Windows 98; U
56 "Opera/7.0 (Windows NT 5.1; U
49 "Opera/6.0 (Windows 2000; U
41 "Opera/7.0 (Windows 98; U
39 "Opera/6.11 (Linux 2.4.2 i386; U

Ok, lets give better Opera numbers. It seems Opera has a few different formats for its browser string. Thanks guys. That helps me a lot..

The top 10 browser string with "Opera" anywhere in it are:

cat list.txt | grep -i opera | sort | uniq -c | sort -r -n -k 1

---
752 "Mozilla/4.0 (compatible; MSIE 5.0; Windows XP) Opera 6.05 [en]"
627 "Mozilla/4.0 (compatible; MSIE 6.0; MSIE 5.5; Windows XP) Opera 7.0 [en]"
617 "Mozilla/4.0 (compatible; MSIE 6.0; MSIE 5.5; Windows NT 5.1) Opera 7.0 [en]"
378 "Mozilla/4.0 (compatible; MSIE 6.0; MSIE 5.5; Windows 98) Opera 7.0 [en]"
277 "Mozilla/4.0 (compatible; MSIE 6.0; MSIE 5.5; Windows NT 5.1) Opera 7.01 [en]"
271 "Mozilla/4.0 (compatible; MSIE 5.0; Windows 2000) Opera 6.05 [en]"
246 "Mozilla/4.0 (compatible; MSIE 5.0; Windows 98) Opera 6.05 [en]"
222 "Mozilla/4.0 (compatible; MSIE 5.0; Windows XP) Opera 6.05 [de]"
194 "Mozilla/4.0 (compatible; MSIE 6.0; MSIE 5.5; Windows NT 5.0) Opera 7.0 [en]"
156 "Mozilla/4.0 (compatible; MSIE 5.0; Windows ME) Opera 6.05 [en]"

Or more specifically, lets find every Opera browser regardless of OS type.. That's just about as big as we can inflate your numbers.

cat list.txt | grep -i opera > work.txt
cat work.txt | grep ^\"Opera > a.list
cat work.txt | grep -v ^\"Opera > b.list

cat a.list | cut -f 2 -d \" | cut -f 1 -d " " > opera.id
cat b.list | cut -f 2 -d ")" | cut -f 1 -d \[ >> opera.id

And then a little cleanup in 'vi' to fix the leading space, and the space versus slash in the two types...

cat opera.id | sort | uniq -c | sort -r -n -k 1
---
2565 Opera/6.05
2488 Opera/7.0
678 Opera/7.01
549 Opera/6.01
537 Opera/6.0
438 Opera/6.04
336 Opera/6.03
105 Opera/6.11
63 Opera/5.12
47 Opera/6.02
47 Opera/5.0
43 Opera/6.0/\xa4/
32 Opera/5.02
30 Opera/4.0/Beta/4
28 Opera/5.11
27 Opera/5.01
21 Opera/6.01/~/
14 Opera/5.12/\xa1\xe8/
13 Opera/3.60
12 Opera/5.12/OCV2/
9 "
7 Opera/6.1
2 Opera
1 Opera/6.01/OCV2/

Now honestly, who should I be designing pages for? the 2,500 hits from Opera 7.0, or the 474,500 from "Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1 ?

**WE** do respect peoples ability to choose what browser they want, and *WE* won't limit it, but I'd bet with these numbers in front of them, most bosses would have the pages designed for the majority..

If the decision were presented to me, wether to include a really great feature that works in Netscape and MSIE but not Opera, or not, and I did exactly what I just did, and saw that 8,092 of 1,949,023 hits came from Opera, that's .0004% of our hits, I'd have to say "Do the change, ignore Opera".

If Microsoft had half a clue (which I'm sure someone there does), and they checked to see what browsers were viewing, and *THEY* saw that .0004% of the browsers hitting them were Opera, they wouldn't waste the time to do make special pages specifically to break Opera.. It's simply a bug.. It's not worth the effort.. If someone did anything, I'd bet they were trying to make a better page for the Opera people, and failed.. Probably a newbie was given the job. Who cares if you mess up the page that no one sees..

Re:Clearly This Sucks but....

[Mitreya]

Please mod the parent up if you have any points. As much as I enjoy the discussion that bashes MS, this is the most insightful/informative post that I have seen.

And, not to defend MS or anything... but how come EVEN THE OPERA.COM SITE NEGLECTS TO MENTION THAT OPERA6.0 HAD A +30 BUG WHEN RENDERING CSS SHEETS?? Who is sabotaging who here?!?

Re:Bad autodetection

[Mitreya]

The fact that it only sends the broken sheets to Opera 7 would indicate that someone at MSN wants to hurt the reputation of Opera's latest offering.

If you read the discussion and set your filter below "+5 for loudly bashing MS", you would notice that Opera6.0 had a problem that was FIXED BY -30 OFFSET. So what you're looking at is the inability to distinguish between opera6.0 and opera7.0. Or, essentially bashing for the (somewhat screwd up) fix of opera6.0.

What truly amazes me is that opera.com description of the problem fails to mention that. I have to say that MSN guys come out looking *far better* than the Opera.

READ THIS FOR EVIDENCE AS TO WHY IT HAPPENS

[spectecjr]

Microsoft didn't do anything of the kind.

1st piece of evidence:

Amaya and Netscape Nav 4.7 both get fed the same stylesheet that Opera gets. Indicating that the site checks for Netscape 6 and above, and IE 6 and above only, providing a default style sheet to all other browsers.

2nd piece of evidence:

Mozilla gets the Netscape 6 stylesheet, which has the SAME bug that the default (passed to Opera) stylesheet has. The same -30px margin is passed to it, but Mozilla renders it correctly (latest build).

3rd piece of evidence:

Netscape Navigator 4.7 MANGLES the front page of MSN if you set the margin-left property to 0px instead of -30px. Here's NS4.7 showing the page with a modified site.css stylesheet:
http://home.earthlink.net/~simoncooke /ns47.png

Whereas here is Netscape Navigator 4.7 using the unmodified stylesheet (the same one passed to Opera):

http://home.earthlink.net/~simoncooke/ns47orig.p ng

Now, if you take a look at most sites, you will see that the most popular browsers are IE, followed by Netscape Navigator 4.7, followed by Netscape 6.x (including Mozilla), and finally trailed VERY FAR BEHIND by Opera.

http://www.sla.org/stats/conf2003/conf2003_sep02 /l astmonth_07_b.htm

Now if you were to realistically act as a site designer, you would go out of your way support IE, Netscape 6.x and company, and Netscape 4.7 -- which is the 2nd most used browser in the world.

And guess which browser needs a bugfix so that it doesn't crash when you pass it a stylesheet it doesn't understand, and so that it doesn't screw up the layout?

Yep, that's right, Netscape 4.7. Our 2nd place winner, and the one that this "horrible, Opera breaking stylesheet" was *actually* written for.

You know, a little research and a little critical thinking might not have set you down this path in the first place.

Simon

Re:Clearly This Sucks but....

[howcome]

This is not correct.

Opera6 shows the page that MSIE6 receives just fine. I even included a screenshot of it on my page -- scroll down to the second image.

If you still believe this is Opera6's fault, please provide a test case showing how it fails.