My mom still uses the AOL browser and everything looks weird in it. But like you said, she’ll never see it on any other browser and I have to assume that a majority of the websites she views are also broken. It’s like people who have bad vision but don’t know it. They think it’s normal until they get glasses, everything is blurry not just some things. I just refuse to enable people who can’t take 30 seconds to upgrade their browser. Ya can’t say they don’t know any better once they have seen my warning message that tells them exactly what to do. Then it’s just being retarded. It’s 2010 for Christ’s sake, updating a browser isn’t like reformatting a computer. Click, install, surf.
ExNihil, have you considered using conditional comments? You don’t need complex hacks, a chrome window or any other yucky mechanisms, just a specially formatted comment which can be placed under your general stylesheet to ensure IE6 follows different style rules (overriding the ones which break and giving them an alternative). Granted it’s additional coding but it’s a clean validating method which allows you to hijack IE6’s glitches of using CSS and give it something less complex.
Yea, shame that most people won’t install the Google Chrome frame for Internet Explorer, partly due to the increased security risks it poses (as it’s hooking unnecessary ActiveX therefore leaving potential exploits beyond IE’s usual ones). It’s absurd to tell client’s to install the chrome frame, it’s basically the same identical thing as saying “Use Firefox or get lost”. It’s a very poor method of web designing, almost as bad as flash only websites. 
Wrong, those statistics are not accurate. No-one considers them accurate either, they only represent that single website which is mainly inhabited by web designers who use the resources to build websites. No normal people would visit the website therefore it’s own statistics are heavily biased and pretty useless.
Totally wrong, W3Schools is not associated with the W3C, nor has it ever been. In fact most of the information on W3Schools is out-dated and highly inaccurate, not something the W3C would allow to happen. The following Wikipedia article is considered the most accurate public collective statistics on the web, they get their numbers from several well known stat counter products and independent groups who produce demographics on web usage. Based on the information provided on that page it’s unsuitable to dump IE6 any-time soon, purely on the basis that most professionals would argue that anything over 1% of all web users could potentially be equal to several hundred thousand users (if you account for ALL people who make use of the Internet worldwide).
Personal pages. As in ‘written in my free time.’ I only have a limited amount of free time, and I’m not spending any of it pulling my hair out dealing with IE 6. As for additional security holes, stop drinking the M$ kool-aid, spewing their FUD for them and prove it.
As for work…
- WE OWN AND CONTROL THE CLIENT MACHINES -
If you’re reading comprehension is to poor to grasp that concept then you’re comments are of no value to me or anyone else. Yes, it is a unique situation that doesn’t apply to most developers.
Alex, you are without doubt one of the snarkiest, smarmiest and passive aggressive posters on this board. Newsflash kid, we are not all dumber than you are. You often have good insights which is the reason I haven’t put you on ignore but I would appreciate it a great deal if you would cut the attitude.
Microsoft has lagged behind in web technologies and internet explorer is one big headache to designers. Hopefully people will stop using IE6 and we can ignore IE6 users once they reach a small minority.
Michael, firstly: it’s perfectly fine for you to do what you like with your own website however if it were me and I had little time I would simply provide an IE6 specific stylesheet using conditional comments which retracts all the style thereby giving IE6 users a passable but uninhibited experience, it’ll be basic but like you said… who cares. At least then there isn’t a pre-requisite for dependency on a third party component which I would argue is totally a cop-out. If you don’t like fixing bugs in IE6, leave them there, adding the frame is nothing more than an extra barrier for the end user. And that was the only point I was trying to make that respect.
As for your business, it’s fine for an Intranet where you control the machines, no issue there. However it’s wrong for you to claim that it doesn’t increase the potential security implications, coming from a security viewpoint one of the main points for contention against the use of ActiveX is out of the fact it executes additional code beyond that of what IE already offers, therefore any flaws or exploits within webkit (or the activeX component) also apply on top of those applicable to the browser. It’s a very simple principle to understand on the basis that having added executable code running within the environment provides another vector for penetration and attack. Claiming that it’s just Microsoft scaremongering is quite frankly an argument from ignorance, not fear uncertainty and doubt. Sorry if it sounds rather abrupt but I come from a background initially in software development and know all too well how added layers of code injected into an application (through API or other vector) can increase the risk of penetration. As for Microsoft’s own claims, they were also backed up by Mozilla and Steve Gibson (world known security expert).
Alex, you are without doubt one of the snarkiest, smarmiest and passive aggressive posters on this board. Newsflash kid, we are not all dumber than you are. You often have good insights which is the reason I haven’t put you on ignore but I would appreciate it a great deal if you would cut the attitude.
As for these comments, I would appreciate it if you didn’t resort to personal attacks. I only stated on the basis of what I know to be factually correct, it has nothing to-do with being snarky or having an attitude, my post was simply to deter future readers of the post from thinking it was indeed a good idea to follow suite (for whatever reason). The fact remains the potential security flaws do exist and claiming that the Chrome frame does not increase the risk to the end user as another point of injection is both irresponsible and poorly sourced. Sorry to be harsh but you aren’t the only person who takes issue with FUD being spread around.
Whoa, what century cave have I been living in? I’ve never heard of this thing??
Michael, I’m curious what this thing is you’ve put on your page… is it for IE6 users only? Or does everyone load this thing?
Well, I suppose I could stop being lazy and google it, but if your site has a working thingie I’d like to see it myself. Link?
Everyone here has different posting styles, and while I think sometimes someone’s particular style comes off sounding patronising or whatever (plenty of people claim that about a few posters around here really), I usually never believe that was the actual intentions behind the words. Esp Alex has a sort of streamofconsciousness style that gets a little hard to read.
A few people have posted on the forums in a way that’s irritated me, but I always try to remember that forums aren’t the best place to show the mood you’re writing in. I wouldn’t let how a post looks get to me so much. I’m pretty sure I piss off people regularly shrug
There’s an ignore option?? I missed that!
http://code.google.com/chrome/chromeframe/ <<< Basically you link to the frame trigger in the header (like triggering IE7 compat) and if you don’t have it installed, the page generally redirects demanding you install Google’s plug-in before you can view the website. Pretty abusive and potentially risky (security wise).
I like that word, though I tend to write through my mindset because I think it’s best to keep the conversation flowing as logically as possible 
and if you don’t have it installed, the page generally redirects demanding you install Google’s plug-in before you can view the website.
Ooh, nasty. Like when I get stopped from viewing any content without JS or Flash. Back Button!
But I’d still like to see it in action on a real site, so I hope Michael posts a link.
[ot]I think your posts would improve greatly by judicious use of a few newlines now and then : )
…yeah this coming from me who’s always scoffing at people who can’t read lines longer than the magical 80 characters… but whitespace does indeed help with reading : )[/ot]
where does that wikipedia article show version number? I’ve seen in many, many places that IE6 is below or at 5% and dropping. Sure, 59% may use IE but most of those are IE7 and IE8…
Clicking on the “Usage by version number” right below the pie chart for Internet Explorer:
http://en.wikipedia.org/wiki/Template:Msieshare1 (20.07% of all browsers)
Again, I work with some pretty huge websites like national food stores, a national auto chain, some big forums etc. and we don’t see anywhere near those numbers. We currently have 5.5% of IE6 for January 2010 across 1.3 million hits on three servers, and most of them are from a specific forum. Do you think a majority of those 20% they found are aggregated in old school labs, libraries and outdated company machines? I mean even my mom using AOL has IE7 on her machine somehow and she doesn’t even know how to copy/paste a pic
It seems unreliable that 20% of all web users are using a 10 year old version.
It very much depends on the demographic of the audience to what kind of results you’ll get (to be honest), it’s very hard for there to be any accurate statistics in relation to browser usage, even with the packages your clients sites will use. Detecting a browser is only applicable when that user has scripting enabled, doesn’t have a user-agent tool cloaking or adjusting the value or doesn’t have any other mechanism at work which might incorrectly report the browser in use (noted and cited in the Wikipedia article under the factors which may affect the results). However what I will say is that if those statistics show 20% of all users are running IE6, I would consider it as close to factual as we can get. Purely on the basis that NetApplications and the other sources Wikipedia use aren’t basing their statistics on a single (or few) websites but on many thousands (possibly millions) of websites to whom they independently report the usage statistics for. I would say it’s safe to assume that while the ratio will differ between websites if you want a general picture of web usage then take those results to be entirely meaningful.
PS: I just checked, NetApplications (alone) have over 3 million sites being monitored through over 3 thousand active partners (a fair number).
berliner_kindl, them stats will reflect just those who visit your site. It just means that those who visit your site are less likly to use ie6. That 20% is from a much wider base. You say:
We currently have 5.5% of IE6 for January 2010 across 1.3 million hits on three servers, and most of them are from a specific forum.
What sort of demographic is that forum aimed at?
18 to 35 males. It’s an automotive forum, ~ 235,000 users.
I wonder if Alexa keeps any sort of browser records?
Well, you’ve seen that others here sure can relate, I don’t know who those “other developers” are that have things going “relatively smoothly for them”, but I’m guessing that they’re either mature professionals or don’t know what they’re talking about.
Unfortunately, most of the posts have not been providing much useful information for how to actually deal with the problems you’re having.
Personally I find the whole idea of “cross-browser pixel-perfect” a bit foolish. I mean how many view the site in more than one browser at a time? Is the design so rigid that a pixel here or there is going to even be noticed let alone damage it in any way?
[OT]*If you want to IE6 bash, please consider posting here instead http://www.sitepoint.com/forums/showthread.php?t=658367
Please make me happy and keep this thread on topic. Thanks.[/OT]
I agree that you shouldn’t strive for anything near pixel-perfect in a 10 year old browser. I’d be more concerned with structural problems like div elements dropping down below each other because IE’s box model problems (with padding etc.) and how IE sometimes does crazy stuff with floats. What sucks is when you have a really nice design with a 1px border on everything and IE6 renders it with 1px on the left, 4px on the right, 1px on the top and none at the bottom. That drives me nuts.
What apps do you guys use to check older browsers, I know there are a bunch out there, I am on OSX BTW, is there anything you can run as a desktop app to do this accurately.
Yes, there are IE6 “features” that can cause a bigger problem than a “pixel here and there”. And it can be aggravating to do “extra work” just to get things to work out for such an old browser. But the fact is it’s still around. Generally, I try to get everything OK, but not equal. I mean it’s kind of like expecting a horse and buggy to look and act like a ferrari. IE6 is just not capable of some things that more modern browsers are. So as long as it “moves” I’m happy.
But I guess if you absolutely must get it to look like a modern browser, you shouldn’t muck up a perfectly good style sheet with IE hacks. Use conditional comments to load a separate file for those unfortunates and keep your mark-up and CSS clean.
For testing? I use the actual IE6 browser. I won’t surf with it anymore, but I use it for testing locally.
Can you run multiple versions of IE on windows? I work on a mac but have an XP machine here in the studio.
You can use something like VMWare Fusion to run multiple IEs on your Mac. For the PC, there are quite a few options, such as:
Alexa is NOT a good comparison of statistics information, the only way they get records is from people who install their spyware ridden toolbar. Therefore they have a very small user-base and only those affect the percentages of people browsing websites (it’s like Nielsen ratings in that only 0.1% of everyone dictate what’s considered popular on TV). If you want demographics, you want the largest independent user base possible, hence why I use the Wikipedia sourced information because it uses what is basically among the 6 largest stat providers. If Google offered stats it would be an even better solution.
This is one of the better solutions: http://spoon.net/browsers/
It allows you to emulate the exact browser through their virtualized sandbox.