Really? Are you in enterprise software or something? We (WePay) see less than 1% of our traffic and revenue from IE<=7, and that includes the data from our embeddable checkout flow. And of that tiny amount, most is related to our API.
We still officially support 7 for our offsite/embeddable stuff, but from the chatter I hear of the devs that are working on that stuff, the maintenance cost sounds higher than the revenue. Then again, that cost to us is yet another reason that people like our API, since its one less tedious thing for other devs to worry about in their checkout process.
International ecommerce, including China, I can't speak to the exact numbers but the revenue from 6 and 7 is in the millions so we're not quite out of that maintenance cost to revenue hasn't quite tipped yet. Though we do display the ie6countdown.com banner along (and I may have messed it up a bit to display for 7 as well.)
I don't think it would be THAT hard to add back in support for IE7. Also, I'm sure that not long after the release someone will make a forked version that supports IE7, since so many of us still need to support IE7.
In fact, it will be hard! They will use `box-sizing: border-box;`, which changes everything about width calculation and IE7 does not support this property.
There's a great reason. Browsers are advancing, and it makes sense to gradually take advantage of features as most of the mainstream browsers start supporting them.
The hacks required to make older versions behave tend to pile up with time, increasing complexity and maintenance overhead for the developers. There gets to be a point where the complexity and overhead is no longer worthwhile to appease a very slim demographic.
Just use older Bootstrap versions if you want older browser support. I don't want your legacy users slowing down the progression for the rest of the modern world :)
This 'very slim demographic' is more than 1 in 20 people on the web still, and catering to them has typically required very little (maybe a few extra lines of CSS; otherwise, modern websites work fine on IE7, just without the nice rounded corners and such). I'll be using Bootstrap 2.x until that statistic drops much lower.
Additional data point for all the people arguing about it deeper in comments: Firefox 3.6 had its last security update 9 months ago (2012-03-13) and hit official end of life from Mozilla a month later.
It's the 16th most popular user agent seen across the ~60,000 websites tracked by W3Counter. IE7 is the 7th. Dropping support for both will affect 1 in 20 visitors of the 'average' website (your particular audience may vary).
Hospitals and big companies lock down the computers, you can't install anything. If it comes with IE7, that's all you got.
And it's not because of these companies.
E.g. they need to run a specific custom version of SAP (accounting), which on turn crashes if any other browser is present. Upgrading would cost over $30M, so forget it for a while.
Or the hospital needs to validate the hardware + OS + software. Until GE does not support feature X they are stuck with the old version which requires a specific list of software to be installed.
You don't want to hear "sorry about the bad news last week that you have cancer - actually it was just a rendering glitch on your MRI due to the new Chrome version".
So life is a bit more complex than "just upgrade"...
To a degree, definitely. As someone who works in healthcare (both as a practitioner - paramedic), and whose other job is healthcare IT...
I know your example is exaggerated for effect, but there is (or should) be a difference between "general use" PCs in a hospital, and Patient Care / Management PCs (EHR, workflow automation, digital diagnostic imaging, pharmacy and the like). That in itself is a failling of healthcare IT policy.
Sure, your critical systems should all be be "certified", but even that is an area ripe for disruption - witness DrChrono in EHR, and I myself am working, or brainstorming on, better "field reporting" (i.e. 911 response laptops / tablets - most software in this field is horrific for usability, though admittedly there is pretty cool functionality, the ability to transmit 12-lead ECG to the hospital for prepping cath labs is fantastic) - definitely willing to talk to people interested in such a thing.
Edit: as an aside, I'm yet to see MRI software that wasn't driven by a Solaris backend, or even Irix, though that does demonstrate how this area works.
Not a lot, but they are there. I just pulled the Google Analytics numbers for one of my sites and over the past 30 days (3.3M visits) there were 3,003 visits with Firefox 3.6 (1.2% of all FF visits and .09% of all visits).
IE 7 is more common, there were 67,841 IE 7 visits. Worse still is the 12,188 IE 6 visits. Amazingly there were 78 IE 5.5 visits, the web must be a crazy place with IE 5.5.
I was using it this spring (2012) because I had a netbook running Ubuntu 9.04 (9.10 had failed to work properly). But after Google Docs began refusing to let me edit, I finally gave in and upgraded the netbook to 12.04, which has an up-to-date Firefox.
Still, I don't doubt that there are lots of public computers (in libraries, for example) running Firefox 3.6 today. I've seen even older.
- Add the box-sizing polyfill: https://github.com/Schepp/box-sizing-polyfill
- Use 'display: inline; zoom: 1;' (with the * hack) wherever 'display: inline-block;' is used