Well beside its actual expiry (which seems to be 2020), people STILL use XP with an ancient versions of IE6 and IE7. Many Vista users stick on IE7.
So calm yourself and put flexbox back down. You've got a few years before that's supported enough.
If they are corporate users, I would be about now announcing that as Micirosft are EOL'img these IE versions only IE 11 and upwards will be supported. I'd probably give a grace period of 3 months officially, and depending on which can clients are affected I'd possibly give them 4-5 months to upgrade.
The reason why is that the fixed cost of a business are that: fixed. The profit is usually over-represented in the last few percentage points of the people using the service or buying the products so if you lop off some arbitrary percentage at the top that 15% of your users might be 80% of your profits. That's a very quick way to die.
What security issues does the latest version of Chrome have, incidentally?
Citation please. Are you really sure this applies to modern front-end web app with reasonable complexity?
Of course if your web app relies on all kinds of sexy stuff then you're going to have to convince yourself that those people don't matter.
Your original statement was broad and general, not qualified. Your anecdotal evidence is not sufficient to support this claim.
In the context of the discussion about supporting IE 8,9 and 10 and using your terminology flexbox (which IE 8-10 doesn't support) seems to be "all kinds of sexy stuff" and "not keeping things simple". Right, got it.
If the cost of supporting old browsers was more than the profit you make, then that spreadsheet will show a different story.
You could be a liar, idiot, troll, or just really love IE. Why would anyone take your word for it.
You omitted the possibility that I might be right, am not an idiot, am not prone to either lying or trolling and absolutely detest IE (and have not used windows for a really very long time).
> Why would anyone take your word for it.
Please don't. Ignore me and whatever I write here on the off chance that there is a grain of truth in it and that it will make you money.
Forgive me if I did not anticipate the Spanish inquisition demanding proof for something that to me simply makes good sense. Feel free to offend your users, to limit your products to function only with the latest browsers and to spend your marketing money to attract people that can then not use your product.
Branding a request for proofs to general, overarching statements as personal prosecution (Spanish inquisition, really?) is a sign of discourse grounded in ideology, not practical considerations.
If you choose to interpret that as ideologically grounded discourse then I think that says a lot more about you than it does about me.
Why would it have to be applicable to the 'absolute majority of HN startups/bootstraps/side projects' to be useful information?
It's pretty clear you're not a developer (or wouldn't be for long).
There are two really bad decisions here:
1) You are driving away up to 15% of people who actually use your product. How much would you spend to acquire that many active users? It's easier to keep existing users than acquire new users.
2) Sending users through a ticket system for errors you could have prevented with a simple warning is a bad experience for them and far more work for you. And most users will silently disappear instead.
disclaimer: my opinions are my own and do not necessarily reflect those of my employer.
Are you really willing to take a 10% hit on sales?
You can announce EOLs until the cows come home (Microsoft has and look how well it has done them). If your product stops working, they'll complain at you and then cancel their contract and find another provider who will support them.
I'm speaking from a gruelling 15 years experience developing enterprise SaaS. We've only just dropped IE6. It'll be years before IE8 goes away.
Entities who use legacy software have more problems than their browser. It's symptomatic of a much larger issue, and I love being able to filter them out!
At an entity level, they may be difficult to deal with, but at the level of an individual in the organisation they know their company has problems. These people move around, and personal relationships are really important long term. This is my main motivating factor.
As long as the client pays (20% of revenue...) and there are no other risks, I'm happy to employ a front end developer who has probably has years of experience making IE6/7 degrade somewhat gracefully. If it were my own product, my own time, my own cost, absolutely not, but if sales were assured with this type of customer, sure.
This is a personal opinion. No one client is the same. I find it interesting to contrast 2 clients I work with who are in differing fields of the same space: (massive names in) financial services: One does QA on only the latest stable version of major browsers (down to Opera, my personal favourite) browser on Windows and Mac; The other does QA on any browser or platform with 2% visitor share (measured by visitors to their website), and also accommodates customers who may be on something obscure but the customer demands it. The former's approach sounds pretty aggressive r.e. serving, or not serving, the customer, but the latter's caused them to avoid abandoning a really backward platform for 5+ years for all customers. Swings and roundabouts. Interesting how we take varying judgments.
Example: skin care website, they sell women's beauty and anti aging products. Their demographic is 15% IE8, but you know who are most likely to buy their products? Mature women, who are also more likely to be the ones in that 15%.
The next part of that puzzle was when they launched into China. China has a huge population still on XP and IE8. Literally millions of uniques from IE8. We even had a percentage from Netscape!
Fortunately the agency handling them was smart enough to research the demographic in both instances, but many agencies won't. They will brazenly force their opinions on their users who frankly, don't give a crap how you feel. They just want it to work so they can buy stuff.
You have to view every hit as a customer. Imagine if Walmart stopped letting in 15% of customers.
This isn't to say the majority of mature women use IE8 or that only mature women use IE8, just that they were over-represented inside that 15% and they were purchasing customers.
You are right that iPads and iPhones are the majority of their demographic.
> I would NOT warn the end users about needing to upgrade
People using your product are discussing it on review sites, on social media, with friends/family, etc.. If you silently give some of them a broken product, a few will contact support; but I imagine most will assume your product is just low quality. So that's the word they'll spread.
For corporate users: well, you'd need to talk with your customers. They might be facing a seriously expensive upgrade process, and could be extremely unhappy to hear you're setting an arbitrary deadline for them.
And they would be right.
So, you'd succeed in angering a lot of customers. Much better to try and inform all of your customers (not just corporate) of the upcoming changes, and provide support materials for them to follow while giving them a reasonable timeframe to do so.
Each business is different, and even then different areas of a product can have differing requirements and needs. What is 15% for you, may be <2% and declining for others. My current and former positions see iPad users in greater numbers than any other single platform/browser, as an example. It may be better served to create features that work for the larger audience and other platforms easily supported than grasping at a fraying edge.
There are lots of intermediate steps like warning about an inferior experience, but at this point it's necessary to test most websites on at least a few versions of IE9.
I've found that using react I've had very easy compatibility with older IE, etc. The only thing I've had to do is install a polyfill or two.