One of the things that I’m passionate about in my role at Microsoft is helping developers ensure that their sites provide an awesome experience in every browser. Yeah, yeah I know it seems odd that a Microsoft evangelist would say this but it’s the honest truth and anyone who’s worked with me before knows that I take great pride in this effort.
We’re Not Supporting IE!
Today, I came across this article on HackerNews which caught me totally by surprise:
“We don’t support Internet Explorer, and we’re calling that a feature”
When I checked out the site entry, it directed me to a startup called Paydirt which has a very slick looking time-tracking & invoicing app. In their blog post, they rattled off a couple of reasons why they’re not supporting Internet Explorer and they were pretty upfront about it:
“That’s why we made a key decision at Paydirt: we don’t support IE – at all – and we don’t pretend to. You can’t even sign up for Paydirt using IE.”
They also mentioned that sensible browsers support cool features:
“Sensible browsers can do amazing things (canvas, SVG animations, CSS3, web-sockets, blazingly fast JS), and limiting usage to these lets Paydirt take full advantage of these new technologies.”
So of course, I had to dig in to figure out what magical features they were using that IE9, and especially IE10, didn’t handle.
Um…it works in IE
Sure enough, when I tried to sign up, I was blocked. So I used the IE F12 tools to fake the Chrome user agent string and, bam, I was able to sign up with no issues. The signup was painless and I waded through the app with Chrome & IE side-by-side. As I went through, I didn’t notice any differences in the functionality. Buttons worked as expected, data was being saved and even panels with fading functionality worked as expected. I was stumped because I couldn’t see what was preventing a user from using Internet Explorer on this site.
Chrome, IE10, IE9/Win8, IE9/Win7 Screenshots of Paydirt
Why is there code in here for IE support?
Next, I dug into the code and again, didn’t see anything that I could pinpoint as an issue. In fact, the Paydirt team had done a nice job of laying out their site using standards-based code. This is precisely what you want to see; developers using standards. As I dug through more files, I came across their main CSS stylesheet called paydirt-e08a29afb369fe41806659f40ff86301.css. When I looked in it, two things immediately popped out:
- They were using many CSS3 features supported since IE9, like transforms & border-radius, but just not adding the –ms vendor prefix entry
- They had the “-ms” prefix for CSS3 gradients which was added in IE10
- They added support for -khtml-box-shadow but not -ms-box-shadow???
This threw me off because in their blog post, their concern was the IE didn’t offer amazing feature support yet many of the features they implemented are readily available in IE9. In addition, they had actually done some initial work on supporting CSS3 features that are vailable in IE10 (e.g.: transitions & CSS3 gradients).
At this point, I’m completely stumped because:
- Their site works great in both IE9 & IE10
- They’ve included support for some IE features in their CSS
- They want sensible browsers that do amazing things and everything they’re doing is supported by IE9 & IE10. All they have to do is add the “-ms” vendor prefix just like they did for Webkit and Firefox. (They didn’t support Opera either unfortunately).
It works in IE so why not just Open it Up
So we’re left to wonder if IE issues are the real reason they didn’t support the browser. And trust me, I totally get the pain developers went through. I’ve only been with Microsoft for 2 years and was a web developer WAY before I joined them (even part of the jQuery project team). But IE9, and especially IE10, have come an incredibly long way since the IE6-8 days and I’m having a hard time rationalizing Paydirt’s decision here – particularly since it’s exactly this type of behavior (writing for or against specific browsers) that sets the Web back again.
Mind you, if their only reason was that their traffic didn’t justify the support for IE6-8, I could understand it. But one of the things that Microsoft has been advocating for is the concept of “same markup” and Paydirt is a great example of this. Despite them not wanting to support IE, because they used standards-based markup, the site just works in IE9 & IE10. And that totally contradicts Paydirt’s arguments about saving time and lack of feature support. Their site JUST WORKS!
Now, it’s not to say that if you really dig deep you may not find some issue. That happens actually in every browser and is not unique to IE. But seriously Paydirt, do you really want to use these arguments when your site actually does render great in Internet Explorer? Have you checked it out? Perhaps being a Mac-based company, testing may be the issue. If it is, then let’s discuss the options.
Claiming, though, that not supporting IE saves you time when your site is working just fine in IE9/10 is a tough sell. Let’s talk. I’m happy to help.
Perhaps they don’t want to commit to supporting IE. Perhaps they see the opportunity cost of supporting IE so high that they choose not to. That’s their decision based on their perception. IE has 10 years worth perception problems to shake off. That’s no easy task.
Hey Will. Their big argument was that supporting IE was tough and that it lacked HTML5, CSS3 & JS features. It’s hard to justify that statement when the site rendered just fine in IE9 & IE10 by simply changing the user agent string. It’s a huge contradiction.
Rey: again, it’s their perception. They have determined, based on their perception, that they don’t want to expend effort to support IE at all, much less even check if they could. They don’t want your money.
And that’s perfectly fine. My point of contention is with their argument that IE doesn’t support the features they need, which from the outside seems invalid.
You can rebut individual instances, but your post here only underscores that some people have simply given up on IE. They’re not even trying to test it on the browser. I wouldn’t be surprised if they, like me, don’t even have IE installed on a machine with which to test.
Considering the number of IE users, that’s a sad statement especially considering that installing a VM in Virtualbox (free) is straightforward.
Considering the number of IE users
Let’s compare apples to apples, are you referring to the number of users running IE9/10 under Win8, which is what you tested? How many is that, exactly, in relation to all IE users of all versions?
Don’t have exact numbers for you but hopefully these stats will help:
I’m pretty positive it’s not a small number and worth supporting.
I also tested in Win7 with IE9.
Did you test it under all six versions of Win7?
I wonder at which point of writing the IE 10-specific code they decided that IE didn’t support the things that were just working? I see sites getting free publicity by throwing IE under the bus, which is lame (speaking as a journalist, it makes me unlikely to cover them because good products & marketing stunts aren’t always close friends)
As a journalist, you probably haven’t had to deal with the unbelievable pain and anguish of attempting to support IE over the past decade. They have, and their fed up. They aren’t alone, it’s a common web developer perspective.
These guys didn’t test IE (6,7,8,9, and 10), and then decide that it was too much trouble so they’d better blacklist it. They simply didn’t test it at all, and put up a blacklist so that the 1% of their potential visitors using IE wouldn’t get the wrong idea.
Like Rey says, Microsoft has built up a lot of bad blood between the itself and web developers. They may be trying to turn over a new leaf with IE 10, but because they lack an auto-update policy 6,7,8, and 9 will still be with us until 2025.
Incidentally, unlike Rey, I wouldn’t lump IE 9 and 10 together. IE 9 doesn’t support CSS3 as well as Chrome or Firefox. For examples, visit caniuse.com.
(speaking as a journalist, it makes me unlikely to cover them because good products & marketing stunts aren’t always close friends)
What a perfect illustration of the issue at hand: at some point you made a decision not to pay attention to marketing stunts as an indicator of good products, which means that you will undoubtedly miss some good products that happened to use techniques you interpret as marketing stunts. Fair enough?
Now, replace “marketing stunt” with IE in your example, and websites for “products,” and you’ll understand where web developers are coming from. Historically, websites and IE aren’t always close friends.
huh, I have to reply to myself to thread here?
George, actually, I’ve done webdev & I know how awful IE has been to support in the past & how much better ie9 & 10 are. I know webdevs have a reflex of pain when they hear IE; but when I test my code for IE9 & Firefox, it kind of just works now. I’m not fond of caniuse & html5test as data points; I’d prefer folks either use the W3C test suite for HTML5 coverage, not something that aligns to a specific browser vendor’s agenda or just come out & say that your test is for how well other browsers match your favourite ;-)
If they didn’t test it at all, we’re assuming the ms- prefixes came from a library? in which case, pick one that minifies the code better, perhaps.
EH – I didn’t say I didn’t pay attention; I said it makes me unlikely to choose to cover them once I have paid attention.
Microsoft has to pay for the sins of the past. But if you don’t reward someone for improving their behaviour, what incentive is there to ever improve? If a stunt-driven product dropped the stunts & turned into a good product, I’d be interested in covering them. I think of it as evidence-based coverage, rather than simply attention based.
The reward is already there, as illustrated in the post: change your UA and the site will work fine. Sites that don’t want to support IE until they’re a known quantity can put a notice up with instructions on how to do it. This method incurs the least amount of developer overhead.
oh, yeah, getting users to change their UA string instead of writing feature detection in is absolutely best practice for web development ;-) when there are good ways to deal with this, punishing the users as a proxy for Microsoft isn’t a good response.
> not something that aligns to a specific browser vendor’s agenda or just come out & say that your test is for how well other browsers match your favourite
The metrics reported on caniuse.com is based off of a publicly available test suite. Are you saying that caniuse’s tests are biased and not representative? If so, can you provide some examples?
> Microsoft has to pay for the sins of the past. But if you don’t reward someone for improving their behaviour, what incentive is there to ever improve?
As a publicly-traded company, Microsoft’s incentive to improve is ultimately driven by their shareholders, and not by some murky idea of reciprocity.
Therefore, web developers, having seen Microsoft cede majority market share to its collective competitors, can simply elect to support the majority knowing that Microsoft must improve just to stay in the game.
While I can understand perception, the problem isn’t really what browsers they do and don’t support. It’s what features they do and don’t support. These days, browser sniffing gets you into trouble. It’s clear that Paydirt is simply browser sniffing (hence why Rey was able to get in by simply changing the user agent). If they change it from browser sniffing to feature detection (which is compliant with standards and what many projects such as Modernizr do), then you don’t have to worry about perceptions at all — just reality. If a browser doesn’t support the feature, then it makes sense not to waste time doing work-arounds for the it. But if a browser supports that feature, why wouldn’t you want that revenue? Seems to me that “sensible”, becomes a very relative term at that point.
How about just supporting web standards? Browser sniffing is pretty much always a bad idea. Build for features, not browsers.
So true Jon. The interesting part about this is that they took advantage of standards-based features and in doing so, ended up supporting IE anyways. That’s what really has me confused.
Because whether or not IE works today, based on MS pattern and practice, they are worried about whether or not IE will support it in the *future* and don’t particularly want customers who are die-hard IE users who will bitch and moan about it hard when IE’s featureset breaks their website, especially when those reviews often get the most press.
IE is a moving target, though. The chance intersection of developer aim and platform position can’t be relied upon.
*all* the browsers are a moving target. trying to support websockets in the dev channel of Chrome was practically the definition of ‘moving target’ last year ;-) It’s just that it’s OK to complain when IE makes a developer’s life hard; complain about the other browsers and someone says you’re a wimp. For extra points, explain to the average dev exactly what’s in HTML5 on any given day & tell them to hang out in IRC to find out how that’s changed a week later…
When people say “IE” they don’t mean just the latest version – many end-users are still using IE 7 and 8,and a few even 6, so “supporting IE” means supporting all these versions, with all their poor support for standards..
I know many web developers that if they would be able to make a choice, they would happily drop support for IE, but unfortunately it’s still the dominant browser being bundled with Windows.
Yep in most cases, you’re correct. In this case, the argument was about lack of feature support, especially relating to HTML5, CSS3, & JS. If your app is advanced where you need these features, then the options are available across all major browsers including IE9 & IE10. Otherwise you could either offer a degraded experience in non-modern browsers or simply explain that the browser doesn’t support these modern features.
Or they could just not support IE and not worry about the users using IE7 & IE8. I have a pretty good feeling the lack of features comment is more directed at those versions, and users still stuck on using them. Why not just respect their decision to not support it and move on?
Tyler, they obviously have that option. If IE6-8 are their concerns, there are obviously ways of advising those users to upgrade.
Using the argument that IE doesn’t have the features they need is inaccurate and disingenuous.
Advising them to… what, replace their operating system? If they’re on IE6-8, they’re probably on XP.
Or that they won’t be able to use the site until they have a modern browser. Or maybe offer a degraded experience.
And like including -ms extension in CSS, is yet another offloading of work from the IE team to the web developer. “Well all you have to do is this extra thing!”
Huh? Vendor prefixes are a W3C best practice supported by all major browser vendors. It’s not IE-specific. Am I not understanding what you’re saying?
Disingenuous, seems a bit presumptuous. Could be a sincere communication problem due to the way you and they scope IE differently. They seem to be focused on 8 and below, where you focus on 9 and up. If they had said we don’t support IE<9, the above is pretty moot. If their analytics show a majority of IE is less than 9 (totally plausible), then for them IE support is (effectively) support for old versions of IE.
Possibly. I hope that lack of understanding is the case.
Correct, you aren’t understanding what I’m saying. I’m saying that the opportunity cost of beginning down the path of ms-specific functionality is not worth it, so best to just ignore even the basic stuff. Sometimes it’s all bathwater and no baby, you know?
Hmm not sure if I agree with that. All you’re doing is heading down the road of an incompatible web for no real cost savings, especially when there’s feature parity in most cases.
All you’re doing is heading down the road of an incompatible web for no real cost savings
OMG, the irony.
Every solution you’re talking about requires more developer effort specifically for IE.
It’s not the reality of every company. Actually it’s almost never the reality of companies.
Extra effort to IE = money loss
Works on IE 10? great! we still don’t support it, if you have problems, it’s because of IE :)
> Or maybe offer a degraded experience.
The thing about offering degraded experiences is that those experiences still require effort to support.
You’re right George. It’s an option though and both progressive enhancement & graceful degradation are still preached heavily.
By acknowledging my point that a degraded experience has associated support costs, aren’t you undermining your own argument that they could just throw open the gates and let IE 9 & 10 users in with no impact to their productivity?
Not really. They could similarly not allow IE6-8 users if they needed a modern web experience.
OK, so gating IE 6,7, and 8 is fair. What about IE 9? Throughout your responses to comments you’ve grouped 9&10 together. Consider the caniuse.com matrix of which modern browsers support which parts of HTML5 and CSS3,
Based on that chart, it looks like any developer that has to support IE 9 & 10 will have to do special, different, custom work (e.g. finding and testing shims and polyfills) for IE 9 apart from all other browsers. So, it seems disingenuous to claim that IE 9 is in the same ballpark support-wise.
If you look at IE 9 compared to the latest mobile browsers, there is more parity; on the other hand, it looks like there is a special flavor of IE9 shipped with Mango whose features aren’t reported on caniuse.com. So, maybe it’s not fair (to IE9 Mobile) to compare baseline IE9 with the other mobile browsers.
Nevertheless, is it fair to gate IE 9 if a site uses modern HTML5 features such as Web Sockets, HTML5 forms, IndexedDB, or WebGL, or CSS3 features such as gradients, animation, or multiple column layout?
Now here’s the thing. IE 10, as great and modern as it looks to be, hasn’t been released. It’s still in preview/beta land. Most sites don’t support unreleased browsers. Most users don’t have unreleased browsers.
Consequently, I (and many others) don’t understand your position given the facts. Every released supported version of IE has merely-frustrating-to-crippling support costs which exceed the costs of just supporting the competitors’ (auto-updating!) products while the only version of IE that achieves some form of feature parity is in beta.
PayDirt’s user-agent sniffing might be a sloppy hack, but their thought process certainly seems sane.
I don’t have the ability to check this off hand, but I will make the assumption that if I try to run this website in Firefox 3.4, or one of the first versions of Chrome, I will get more display and compatibility problems than IE9 and IE10 get on that website.
The circular reasoning of web services saying “We still support IE6 because companies use it” and companies saying “We won’t update IE6 to higher because web services still support it” is really starting to get old. Chrome and Firefox now update automatically, and people rarely come out to say “I have Firefox 2, why won’t you support my old browser?”, so why should we extend the same courtesy to IE6?
IE is, and seems like it will always been, the worst thing about web development/design. If you ask me what the most annoying, stressful, time-wasting part of the last few apps I built were, it would be dealing with IE-only bugs that make no sense at all. In IE8, you can’t even repeat a background image on a body element. Microsoft owes me a lot of money for wasted hours making my sites “work” in their browsers.
Unreal… it’s almost as if they’re intentionally shooting themselves in the foot. Does it matter that a small percentage of users are using IE? If it already works with the latest versions of IE, then simply allow it to be supported.
Where are the costs in that?
I think that what they’re doing is simply bad business.
“Working in IE” and “Supporting IE” are not the same thing.
The fact that they used standards-based techniques allowed it to just work in IE, like any other standards-based browser. This is the way we should be coding.
There has never been a guarantee that IE would follow standards, so you can’t blame them for not assuming that it would.
Also: IE being “like any other standards-based browser.” LOL
“Working” means that it works now. “Supported” means if it doesn’t work in the future, you can file a bug report.
Not saying they shouldn’t sniff for ie 9+, but it costs money to support browsers, especially when you are not developing on windows.
I think this is a marketing trend first noted here:
Startups who are appealing to a niche that doesn’t use IE. Given the odd attitudes of Silicon Valley types it almost makes sense. I don’t think their goal is to serve the largest number of customers — it’s to get bought. This is one (odd) way to get attention of a certain type of user/investor/buyer.
I actually know the Paydirt guys, and that is most definitely not their intent(and they’d be offended by the suggestion).
It’s two hard working guys bootstrapping a business. They simply don’t have the resources to ensure that their app works in IE.
Hey Alan. I’ve seen their site and it’s definitely sharp. And not having resources, especially as a bootstrap, I get it. But their post was specifically how they saw not supporting IE as a “feature” and listed technologies that IE supports. That’s where it gets a bit confusing. I also offered to chat with them when they want to. Be happy to get on the phone with them anytime.
They’re using LESS or SASS or something similar such that when you use a box shadow or similar it outputs the appropriate CSS for each browser. I doubt it was an actual “choice” to include it.
I’m playing devil’s advocate here obviously, as I don’t necessarily agree with them :)
When they’re in I’ll point them here (although I’m sure they’re already aware!!)
Thanks Alan. I’m really happy to chat with them and get their feedback about IE.
They specifically mentioned web sockets. I would say that this is their BIGGEST concern with users using their product.
try that on IE9….
works just fine in IE10;) Now that really argues for feature detection vs UA to me; given how, um, varied sockets support is in the full range of ‘modern’ browsers, there’s going to be quite the tree of user agent strings to parse.
“Perhaps being a Mac-based company, testing may be the issue. If it is, then let’s discuss the options.”
I’d love to have the options discussed publicly. I’m a Mac-based HTML developer, and one of the reasons I don’t test stuff in IE as much as I’d like is because there’s not a slick way to do it on a Mac.
(a) install Oracle VirtualBox
(b) install Windows in a virtual machine … you can probably even get a free Windows 8 consumer preview, and I’m sure IE9 and IE10 work on that
(c) configure networking for your virtual machine … actually, almost no configuration
(d) test from your virtual machine
Yep. Note that I tested the site on Windows 8 using IE10. I used the F12 to change to IE9 mode to ensure it also worked. Finally, I went to my Win7 machine and tested with true IE9. All seemed good to me.
You forget the bit where the “preview” is just that, a preview. How long before it expires? Then what do I use? (This is precisely what we do btw)
Relying on Microsoft’s release cycle to get access to a “preview” OS is not exactly a sound testing strategy. If MS is serious about startups supporting IE (and therefore windows) they need to provide the tools to allow us to do it without having to buy a PC and a Windows license JUST for testing.
Hey Alan, I mentioned all this in other comments I’ve made. You’re right that the preview will end eventually. But there are the Internet Explorer Application Compatibility VPC Image which allow you to test as well as services like BrowserStack.
I think Microsoft needs to look at better solutions for allowing Mac developers to test IE.
And then every time you do a release…
e) Fire up your VM and spend more even more time testing it in IE
f) Spend hours hacking around trying to work around IE specific issues.
g) Give up cos IE is only 1.6% of Paydirt’s traffic.
My eye got hooked on that number in their original blog post. Of COURSE, if they specifically block IE users from signing up, their IE traffic will drastically go down. If the only traffic you’re giving to a browser is a landing page and a message that says “your browser is not supported”, people either turn back or switch browsers, so we can easily infer that this is the reason why their IE browser is significantly lower than it should be.
The windows 8 preview is free, and if bootcamping it is too much trouble parallels is only $79.99. You can also test your site using browserling.com
I’d love to see Microsoft work with a service like Adobe’s free BrowserLab which makes it way easier to do cross browser, cross platform site testing; I find it hugely useful. (You don’t need Dreamweaver either)
There are multiple reasons why it would work in IE9/10. And why they would have the IE prefixes.
a) They may be using compass (compass-style.org). With compass you get mixins which create rounded corners and use browser prefixes. So compass knows how to just make it work.
b) IE9/10 are pretty decent at rendering, so you get the benefit of a decent browser that probably just works. But if there’s a bug, they don’t give a shit.
c) It’s expensive getting hardware to test on, taking time to test it, making sure you have a windows installation, they may just have macs and test on Chrome/Safari/Firefox/iPad. Chrome makes the claim that if it renders on Safari, it probably renders on chrome and vice versa. So lots of freebies.
d) DEBUGGING: To debug in any browser I can do it on a mac (decent dev env) to debug in IE I need a dev environment in windows. Enjoy setting that one up.
Just realized how skitzo that was. First 2 point to the reasons why you would get “free” IE support. Second 2 is why they don’t give a shit about IE.
I guess that’s a big sign for Microsoft (hanging there for a big while) to force upgrading IE to the 9 or 10 version (even on XP). I know they made that decision but probably it should be done earlier. And now it should be done faster and around the world.
Consider the source. You found a headline that was overstated and patinaed with cynicism on a social news site. Sensational headlines (by definition) do have a tendency to get passed around more than modest ones.
Not to mention, I hadn’t heard about paydirt until you posted about them. Odds are you wouldn’t have posted anything about them had they not made this claim. It’s hard to argue against sensationalism. On the internet, doubly so.
When there’s a version of IE that runs on OS X (which a huge number of front end devs use as their primary OS), and doesn’t have *really sucky* dev tools (F12 is like pulling teeth compared to Chrome’s dev tools or Firefox+Firebug), maybe developers will start to get over their instinctive antagonism towards IE. Maybe IE10 can deal with all the features they’re using, but supporting it doesn’t just mean adding a bunch of -ms prefixes and turning off UA sniffing, it means taking time out of each iteration to test every feature in another browser (with sucky dev tools).
Historically, IE bugs have a strange habit of being incredibly difficult to fix, and often impossible without detracting from the experience for users of other browsers. Whether that applies to IE9/10 or not is moot – the painful memories make us not want to risk it.
IE9 and IE10 are still way behind other browsers in terms of standards support. (Source: http://html5test.com/results/desktop.html) It’s safer just to not bother supporting IE at all. For most people IE support is a necessary evil, but Paydirt decided that didn’t apply to them. I applaud them.
No, see, you don’t understand: the web developer just needs to install a virtualization environment, into which they register for, download, and install a version of an operating system that may or may not (I haven’t checked) expire in some not-too-distant future. Then just add the relevant -ms- tags to your CSS and tweak as needed in order to be compatible with the very latest versions of IE (earlier version users: you’re on your own).
Opera started reading -webkit- prefixes to work around such [redacted] attitude. Microsoft will be next.
Are you excited that soon you’ll have to workaround difference browser vendors’ implementations of -webkit- prefixes?
oh yay, Opera just took a dependency on bugs in someone else’s code. double yay.
Although i hated IE for years when i was developing front end stuff back in the day i must say this is a pure publicity stunt.
The whole purpose of the article was to get some traffic on the site. It’s a well known fact that web devs (their core cutomers) hate IE, so it’s an easy and cheap way to get their attention – some of them will check out the app and maybe become customers.
disclaimer: i’m involved in a company that makes a time tracking app and we fully support IE.
Attack the messenger all you like, but the simple fact is that IE isn’t even available for the two most popular web development operating systems.
Not attacking anyone.. why are you saying that? I’m just stating the obvious. Also, it’s a stupid move to lock out users when the site actually works. Let people decide if what browser to use…
You’re attacking Paydirt on the basis of this being “a pure publicity stunt,” meaning that it has no practical value outside of attention.
People can still decide what browser to use, Rey demonstrated such with his post here. It’s not any more difficult than developing under Virtualbox.
“It works” != “It’s supported.”
“It’s supported” means they’ll allow people to report issues, and they’ll address them. It means they’ll test their site on IE, a non-trivial thing thanks to IE being a browser only available on Windows machines.
I applaud Microsoft’s embracing of open-source and open-standards. It’s going to make life a lit easier, and I look forward to IE10. That their stuff works on IE shows how much of a commitment Microsoft is making with IE. But just because it works doesn’t mean it won’t be supported.
And also, thanks to libraries like Compass, the -khtml prefix is automatically inserted. So the library they use for their styles may just be giving that away for free.
Conveniently, you neglected to mention that IE9 and IE10 are FAR FROM having wide-spread adoption in the browser market. According to w3schools.com IE has a total of 18.3% browser marketshare for 4/2012, and IE9/10 account for a COMBINED 6.5% TOTAL browser market share for 4/2012. That means thats 64.5% of all IE usage is still IE6-8.
There in lies the problem. It’s not that Microsoft isn’t baking in less suck than they traditionally have in older IE versions; it’s that they have a distribution and fragmentation problem. IE9 and 10 might be the best thing since sliced bread; but no one is using it.
Furthermore, Microsoft has left a nasty taste in my mouth regarding supporting Internet Explorer. Development tools up-to-and-including IE8 are garbage. They’re a lot better than older IE versions, but they’re still WORLDS behind Chrome and Firefox. I can’t even debug any version of IE without a VM of Windows installed, which again is a giant pain in the behind.
Microsoft has several problems to solve in the browser game. Primarily, making their browsers not be complete trash. Secondly, FORCE USERS TO UPGRADE! I don’t care if your browser is the most performant browser in the world, if no one uses it – I’m not wasting me time. Thirdly, I NEED TO DEBUG CODE! I can debug all the major browsers on my Mac with minimal effort; no I don’t get complete coverage, but it’s an excellent base. I can’t even LOOK at a site in IE without booting a VM, which Microsoft doesn’t even *really* make available.
Hey Jim. The topic of stats is a heated one and here’s a totally alternative view:
If we solely look at that, it seems a compelling reason to continue to support IE. But that’s not what I’m focused on in this post. If this was solely an issue with a company saying that their stats support their decision, there’s no argument that can be made. It’s their “business” decision. That’s not what Paydirt said though. Please reread their blog.
And yes, I agree with you that we need to figure out a way to help Mac developers have a better test experience for IE.
Actually, the adoption rate of new versions, not the growth rate, is the the important metric.
Notice how abnormal IE’s adoption rates are? That’s because IE is tied to OS version which means that developers need to remember 10-15 years worth of hacks to keep all of the various “in flight” versions of IE working. Developers supporting browsers with faster adoption rates might need to remember only one year’s worth of hacks.
While it’s true that I pretty much hate IE, I have to give MS props for stepping up support for web standards in a big way with IE9/10. Sure, if your company cares about being streamlined, don’t test or offer support for IE. But sniffing UA strings in order to lock out IE users seems like a dumb business decision. Maybe they think the “controversy” will get them more savvy customers? Questionable at best.
Oh come on. Every other rendering engine can easily be tested on any platform (Windows/Mac/Linux), but IE requires setting up Windows. On a mac this means getting a Virtual Machine (and buying or pirating windows), using another computer, or using a cloud-based solution like Sauce Scout, which requires doing extra work to test a site that’s running locally.
You conveniently left that part out of your post.
My post wasn’t about testing, it was about feature support. Wasn’t a matter of “conveniently” leaving something out.
With that said, yes, testing IE on Macs is a drag and yes, we (MS) need to figure out how to make it easier.
How can we know if IE properly (for the dev’s purposes) supports a feature without testing on IE? I simply don’t understand how you can try to separate testing from feature support like that.
You install a virtual machine w/ Windows on it or use any number of cloud-based solutions such as BrowserStack. Right now, you can download Virtualbox (free) and install Windows 8 Consumer Preview (free for now) and test IE7, 8, 9, & 10.
I’ve already commented previously how I think MS needs to consider better IE testing options for Mac developers but the above scenarios work, albeit aren’t as convenient.
Just Mac, eh? Why does IE always require something extra?
My point was that indeed your post *was* about testing, unless you’re totally excluding developer effort in producing sites that work well with IE. Feature support and testing are the same thing on the construction side.
Nope you’re right, it’s partially about testing because had they tested their site, they would’ve seen it worked. Considering that they used standards-based techniques, it rendered great in IE9 & 10. But overall, it’s more about their statements regarding feature support.
Ievms is also an option for Mac developers. I use it and it works okay for running the official IE App Compatibility images in a VM, although having to reset the windows authentication every few months is a bit annoying.
Had Microsoft bothered to spent half the time developers spent fixing websites for IE. This would probably be the best browser around. Although Microsoft made incredible leap supporting standarts. I still think Microsoft thinks of IE as a side product.
Nearly all my friends who have PC’s have Windows XP even on new machines. So unlike other browsers such as Chrome of Firefox, users can’t update to IE9 or IE10. So in that sense IE doesn’t support all the modern features PayDirt wants. If Microsoft allowed users to upgrade to IE9 and IE10 that would be a different story.
This company decided not to support IE. While your quick search showed no issues that doesn’t mean there aren’t or more importantly that there won’t be in the future as they add new features. One would think why not allow all IE users to access the site and if something doesn’t work oh well. But if they did that then all the IE users would email support anytime something didn’t work as expected or look right. That’s more time and money to deal with the users and the support requests.
Once again the big issue in my eyes is Microsoft requiring operating system versions (that have hardware requirements) based on what browser you have installed. For people using IE7 or IE8 and such they can install Chrome Frame however (no admin access needed) and that will add the necessary support and also assuming the developers added the necessary code to their site they can allow IE users with Chrome Frame installed. This line is included in HTML5Boilerplate.
Rey, you clearly have too much time on your hands. Nice ploy do divert attention from paydirt to you. Clearly worked.
Haha! Wow, okay.
Not supporting ie seem like a good idea. It’s pretty much the worst browser there is. 10 editions and still lacking some of the css2.1 support, no updates on the fly like opera/safari/chrome and firefox and ridiculous 10 year life span per edition. Say whatever you want but ie slows progress and when people are pretty fed up when competitors are far far ahead ie wobbles at the end. We just stopped supporting ie6 still we need to deal with turds like ie7 and 8.
Open it in ie7 or 8
I’m focusing on this statement they made:
“Sensible browsers can do amazing things (canvas, SVG animations, CSS3, web-sockets, blazingly fast JS), and limiting usage to these lets Paydirt take full advantage of these new technologies.”
you say “… installing a VM in Virtualbox (free) is straightforward.” It isn’t very straightforward. Also while Virtualbox may be free, Microsoft Windows (to you know, install on your VM) isn’t. Ignoring the very real option of stealing Microsoft Windows, it should be noted that buying an operating system (and to really make certain your test environment matches your user base you probably need both XP and 7 so it’s more like two OSes on two VMs, still sound straightforward?) from a company that is already making your life difficult by having distributed the hideously non-compliant IE6-8 to the masses just isn’t terribly appealing. Of the five major browsers, (Firefox, Opera, Safari, Chrome, IE) all of them are available for all major OSes except one. If Microsoft wants web developers to support their browser they ought to make it available for the OS that comes with web developer’s MacBooks. The point remains that Paydirt is doing a good thing by pushing users towards modern browsers and Microsoft needs to give up and start distributing Windows with Chrome or Firefox (I’m assuming Safari would be way too much of a political concession).
Microsoft’s provided pre-built copies of Windows 7, Vista, and XP for browers testing for a long time: http://www.microsoft.com/en-us/download/details.aspx?displaylang=en&id=11575
Doesn’t anyone realize they just wrote that article and disabled ie9 to attract hipster freelance designers and developers? The whole thing is basically a PR stunt since they realized 99% of their hits were from Mac anyway. Especially posting their own article to Hacker News was just to get attention from the large freelancer following on HN.
I’m *no* fan of IE. However explicitly blocking IE9 is just idiotic for any other reason. They could, at worst, just have a subtle notice letting people know it is optimized for Chrome/FF.
However, for the reason of getting Mac using hipsters excited, it was pretty brilliant.
(btw, I’m not a Mac or Windows user)
You know what a concern troll is?
No, I don’t….?
Ok, just looked it up: http://www.urbandictionary.com/define.php?term=concern+troll
Not sure how that applies to me. The whole article is getting up in arms about this assertion that ie9/ie10 are easily supported and I’m just pointing out that the assertion wasn’t ever made in earnest to begin with. It was just a ploy. That’s fair. I think, isn’t it?
In fact, I think it is fair to say that calling someone a concern troll when they aren’t really is in fact being a… concern troll! LOL!
So at least one of us is being a concern troll. Possibly both. Which one is it??!
Oh wow! I just searched on this page for “EH says:”.
Looks like you were really enjoying this debate! Sorry to point out that the entire thing is a big joke on everyone who bit on their ploy.
What? This topic is low-hanging fruit. I should be less bored, but then again my sites work well without having to renew my MCSE in order to learn a particular browser’s nuances.
I, as most other devs had to struggle a lot with IE in the past and probably will have in the future but what they did is either really silly or a marketing stunt.
IE9 and IE10 probably dont have a high market share at the moment but they will have, so whats the problem in adding prefixes. As you said they also “forgot” opera which forces them to add the -webkit prefix.
And for testing IE simple use https://github.com/xdissent/ievms – i dont see any problem here – nobody compla
Ins about buying an ipad on the other hand.
IE 9 is really solid. IE 10 is pretty amazing, IMO. Not supporting IE 6-8 makes a lot of sense to me. But blocking IE 9 and 10 users seems silly.
It’s a shame Microsoft didn’t make IE 9 work on XP. That’s a huge hole for Microsoft and leads to problems like this. Leaving half of the Windows world behind is going to haunt you all for years.
Tantamount to testing under virtualized Windows, they could just add a page item that says “IE users please change your user agent to (Opera|Firefox|Chrome|*) for the richest internet experience.”
I’m just wondering how hard is it to add auto update features to ie7+ windows updates pretty much roll out every week how hard is it to roll out browser update with it?!?! Ok ie10 is meant to fix this issues but in 2029 when ie 9 still be with us how many more features we will need not to use to support microsofts “browser”.
I think the big point here is that as IE is growing up we will stop seeing a lot of these issues.
You don’t have to support older version of any browser. But there is also no reason to discourage people because they aren’t using your personal favorite.
Hot thread! I a member of the IE team at MS and just came back from Future Insights in Las Vegas. If there is one take away is that it is natural and efficient to develop experiences that work beautifully across browsers, devices, and overcome any accessibility barrier: after all we want our content to be seen at its best by the highest number of people.
Fortunately – thanks to the emerging tools, libraries and best practices – it is becoming easier and more efficient for front end developers to do so. I remember a similar case that surfaced on the tech memes a few weeks back, and I loved the rebuttal from Sean Gerety. http://techcrunch.com/2012/04/16/4ormat-rebuttal/ I think it applies very well to this case too.
Even if this is coming from the vendor, I’m glad to see someone did this check. All the group-think, anti-MS, FUD throwing Hacker News are annoying. It’s like SlashDot*100. At least 10 years ago we all had valid reasons to complain about the state of MS web tech. People really need to get over their mindless hate toward today’s excellent technology.
You can setup IE VMs using this script: https://github.com/xdissent/ievms
Or you can just recommend that users change their UAs, with a link to instructions.
Somehow nobody has mentioned the easy, no-buying-an-MS-OS-way of testing in IE on Linux and OS X – IEs4Linux.
Have a look at the installation instructions: http://www.tatanka.com.br/ies4linux/page/Installation
It is very easy to set up and allows you to run IE 6 and 7 without installing Windows on anything.
Don’t let the name put you off, it also runs on OS X (http://www.tatanka.com.br/ies4linux/news/49).
It’s not all that hard to design for IE. Of course I’m speaking from a freelancer pov – I’ve got time to expend during night.
I understand designing for startups requires you to make some sacrifices due to time constraints among other things, but completely dropping support for a browser is borderline bigotry.
No, it’s not bigotry so much as learned behavior. Microsoft has earned this.
It wouldn’t be confusing if you had half a brain. No one wants to be locked into supporting IE in the future with a site that may need to change and might potentially not work with IE. It sounds unlikely at best that you actually tested every single feature and that you know for sure that it looks and works exactly the same in IE and other browsers.
No one wants to be locked into IE. I think that’s a fair remark. But by browser sniffing, you are locking yourself into the support models for specific browsers rather than doing feature detection and future-proofing your site against ANY browser lock-in.
That said, no one is asking to support a particular browser. If your site is just doing feature detection rather than browser sniffing, you can use an objective measure such as “sorry, this application requires feature which isn’t currently supported in your browser.” Instead, there is a subjective test for the browser itself even though it might very well support those features.
IE sucks! Worst thing Ever! I hate IE!
Nitpick: It’s hardly surprising that they haven’t added -ms-box-shadow. IE doesn’t actually support it. Doesn’t need it either. IE9 simply handles box-shadow and only the unprefixed version (which is as it should be, imho).
If the world was a decent place, MS would be forced to financial compensate all web developers and web designers for hours and hours around a piece of software that is pure crap. The day that a global legal action against IE, for professional and moral damages, takes place – I will be among the first to sign it.
“Vendor prefixes are a W3C best practice supported by all major browser vendors.”
Hey Rey, its the opinion of many that vendor prefixes are actually doing more harm than good and should be discouraged altogether. See http://hsivonen.iki.fi/vendor-prefixes/
Well, I have old Windows-XP with (disabled) IE-6, but always use Firefox. If anyone is (for some odd reason) using some version of IE, would they somehow neglect to also have Firefox? Such neglect would be just plain stupid. If they can’t get through with IE, they could simply try with Firefox.
If they are too stupid to use this simple solution, maybe you do not wish them to be using your product, since their stupidity will cause them to make endless service requests, and issue many complaints when they cannot get things to work.
> This is precisely what you want to see; developers using standards.
So what is Microsofts excuse?
Reasons why IE, even IE9, is much harder to support than other browsers:
1. The IE tax: Testing IE requires a VM license ($50-$100), plus another $100-$150 for a copy of windows. Plus a machine with enough RAM+disk to host the VM. That’s $250 out of pocket, just to test a single version of IE. Nevermind trying to test versions 6, 7, 8, and 9 (and, now, 10).
2. Inferior debug environment. Yes, IE’s debugger has (finally!) improved past the days of the Script Debugger. But it’s still inferior to what other browsers offer.
3. Lack of automatic updates. Often in other browsers one can simply wait a few weeks or month for a bug to get fixed and released. Not the case with IE. Yes, IE10 will have automatic updates (finally), but supporting IE today means supporting at least IE9 at a minimum. And unfortunately that browser is still subject to the vagaries of user adoption, meaning developers will continue to have to code for the lowest common denominator feature set.
3.5 That lack of automatic updates is a killer. Sure, IE9 has most of the standards support of other browsers today, but come back in 6 months and will that still be the case? Or will it be standing still while Mozilla and WebKit race ahead with their standards development and support. If so (and that’s almost surely going to be the case), are you suggesting that web developers should commit to not taking advantage of these standards coming down the pipe simply so their sites will continue to support IE9 in the months ahead?
Your claim that Paydirt supports IE today, while technically true, is a pretty hollow argument. Until IE10 is in production, and Microsoft is **aggressively** forcing users to upgrade to it, it’s not going to change the decision-making landscape significantly.
Guess nobody here has heard of the Microsoft Action Pack subscriptions?
A windows license + 10 minutes setting up a VM. Big deal.
The only publicity these guys got is for being cheap and lazy.
On a separate note IE9 isn’t available for windows XP so this entire post is assuming not only browser version but also operating system.
Comments are closed.