How does your web dev team decide on browser support?

Comments

[Antonin J. (they/them)]

We support the usual "last two browsers". The best support goes to browsers that the devs use and the top browsers that our users use so:

1. Latest version of Chrome -- most developers use it but also, most clients use it!
2. Firefox -- I use FF exclusively and have forced prime support for it which is great. None of our clients use it as far as I know.
3. IE Edge -- this one is actually pretty easy to support since it follows the specs. Some our clients use it.
4. IE 11 -- I think someone uses IE 11 but we don't explicitly check for support. We fix issues as they come up. So far? We had one issue 1-2 years ago regarding input box sizes.

But that's about it! Typically, there aren't any issues we have. Most of the styling we do hasn't changed over the past year or two which means we've worked out all of the cross-browser styling bugs. JS browser-related bugs have been extremely scarce.

[edA‑qa mort‑ora‑y]

I think it's fair that unless you support >95% of the "web" platforms available you aren't doing web development, but in stead "Chrome", "Firefox", or "IE" development.

It's really not that hard to support the major browsers, especially for basic apps. I don't really see an excuse* for it not working (sure, glitches are fine, as long as it works).

*This applies to Google as well, who seem to not care anymore that some of their crap only works in Chrome. The problem of vertical integration.

[nayman898]

I know this is a year later, but I disagree. Internet Explorer NEEDS to Die. The reason it hasn't completely died yet is because of this idea; "Someone uses it, so I guess we have to make it work". IE is a security risk, it's a Javascript nightmare, and it reduces performance for ALL users just to include the <8%. Supporting it is a bad idea, and it's not just an opinion. There is factual evidence as to why it is bad, and there is not enough people using it for the inclusion of them.
Old browsers don't die fast enough because devs keep supporting them. The sooner the support is ended, the sooner it will die.

[Moe]

It's really not that hard to support the major browsers

Wow, It's quite difficult for me. A lot of time I feel like I have to rewrite entire features because of differences of these platforms. Each platform even performs differently depending on your OS (IE11 on win10 vs win7), so I have to account for that. It's really not simple.

[Ben Halpern]

I'm sensing some frustration in the early answers to this discussion 🙃

[KhoPhi]

My team builds for financial institutions, and so we support as far back as possible, making Internet Explorer, at least 8+ a target.

We use Angular 5 and Bootstrap primarily, so a handful of compatibility issues are handled seamlessly.

However, we ensure there's a cross browser support.

[Shi Ling]

Ouch, IE8.

[Jesal Gadhia]

We QA test the last two versions of IE Edge, FF, Safari, and Chrome. We try our best to support old IE and other older versions but don't spend a whole lot of time if we run into some hard to fix issues.

[Ben Halpern]

Seems sensible

[Isaac Lyman]

I feel like there are two sides to this one. There are a lot of questions that technical management asks themselves, but only one question that they should be asking.

Questions they ask:

• Who are our users? (To be fair, a lot of B2B companies are forced to support old versions of Internet Explorer because they have enterprise customers who give their employees super locked-down computers and won't invest the money to update them.)
• What does our competition do?
• What kind of development resources do we have? (If you're a big company that moves slowly anyway, legacy browser support may seem like an easy win because it doesn't require a lot of creativity or debate.)
• What will our egos allow?
• Do we need a monopoly? (If you're Google Search or Facebook, you have to support IE5 because your ability to sell ads at a premium is extremely dependent on your dominance in the market. But you're probably not Google or Facebook.)
• How complex is our front end? (If you're willing to be Craigslist, then heck, you can support Netscape Navigator. On the other hand, if you want a shiny, cutting-edge single-page app with cool animations and complex interactions, it's hard to support non-evergreen browsers. You either have to pick one or spend a lot of money.)

What they should be asking:

• What's the marginal cost of supporting another browser, versus the marginal benefit of gaining people who use that browser as prospective customers?

If you're doing analytics, you can see exactly what browsers your users and prospects are using, and if you know what each user is worth, you can determine if it's worth the cost of supporting their legacy browsers. It comes down to a subtraction problem.

In many cases, Chrome + Firefox will net you 75% or more of your potential market in one fell swoop. So that's something to think about.

[Leighton Darkins]

For applications for consumption inside an enterprise: It's not a decision, it's typically a mandate, delivered reluctantly by whoever the head of the IT Department Happens to be. This can be a blessing and a curse. One time I only had to support the last 12 months of versions of Chrome - I thought I'd died and gone to heaven.

For applications facing the outside world: Our focus is always maximum accessibility. That means looking at trustworthy browser usage stats and targeting whatever doesn't fall into the "other" category on the chart - hell, we'd even target "other" if our clients would pay for it.

Typically this manifests as targeting the last years worth of releases for Chrome, Edge, Firefox and Safari (the web's new problem child) and the most prevalent screen-reader.

These days, this combination seems to result in nice, broadly accessible sites.

Just don't get too ambitious and try to implement things like automated screenshot comparison across browsers, for maximum consistency - you'll have a bad time. Things are always minutely different between browsers, accept it and move on 😣.

[Moe]

My current application is used by our Sale reps nationwide. Since Sale associates only use IT managed computers and our IT Department only supports IE 10,11 (Not Edge) and chrome. I only have support IE and Chrome.

It's still frustrating when things work perfectly in chrome, then everything explodes in IE. Things like Local Storage, page rendering, onloading, and CSS/Style are implemented differently in both platforms. Its a compromise trying to realize requirements on both. I often have to talk to the Bussiness Analyst and amend requirements based on this compromise.

[Moe]

While I was writing this, I got an email from my team telling me that our styles are not working in IE! -_-

[Doug R. Thomas, Esq.]

We only do lead-gen websites, so it's not as difficult as say, a React app.

We support down to IE10 based on a combination of web analytics and what we can reasonably test in-house - for a typical site, we see it leaves about 2% of IE users (about 10-20% of traffic) out using that strategy - so maybe .1-3% of total users.

Over the past 4 years that I've been here, we started at supporting IE8, and when it dropped to less than 1% of IE traffic, we dropped it. We dropped IE9 for similar reasons, as well as the fact that some of the Google APIs like YouTube and Maps API aren't compatible with lteIE9.

We're less concerned about browsers like Amazon Silk, which is only .5% of traffic to begin with, and then heavily fractured from there. We also don't have a high percentage of developer users, so stuff like Brave and Opera are pretty rare.

[Weston Wedding]

Depends a lot of the target audience of our client and whether they have preferences.

It tends to be "support browser versions that have come out in the past 2 years"

Or, at least, that's the goal. There are frequently times where towards the end of development someone with an IE machine will test a site even though we were pretty explicit early on that we wouldn't support it and the client will suddenly care about IE.

[Ben Halpern]

There are frequently times where towards the end of development someone with an IE machine will test a site even though we were pretty explicit early on that we wouldn't support it and the client will suddenly care about IE

Ah humans.