Hacker News Reader

Comment by darrhiggs | Hacker News Reader

>> E8 and IE9 still make up 32% of the browser market share and that is a lot to give up if I was to revert to the flex solution

Yes, if you have an even worldwide visitor share. Can I Use[0] puts USA IE9 usage at 3.86% and IE8 at 7.28% (11.14% in total). If your user base is different it would be wise to examine the data before discounting flex-box based on worldwide stats.

If anyone has access to netmarketshare's USA[1] (used in the article) data, it would be nice to see what it is.

EDIT: Also the article's stats[2] are for desktop only. It seems[2] that mobile support is relatively high with Can I Use[0] stating that total USA support = 86.46%.

[0] http://caniuse.com/#search=flexbox [1] http://www.netmarketshare.com/browser-market-share.aspx?qpri... [2] http://www.netmarketshare.com/browser-market-share.aspx?qpri...

View on HN

netmarketshare also tracks users based purely on the user agent, regardless of how much browsing they do. As a result, it skews toward older browsers like IE8 and IE9 whose users typically don't surf very much at all. IF you track browser usage based purely on page views per browser IE8/9 represent a far lower percentage of all browser usage.

I was going to comment on that as well.

That depends on your target, my blog gets about 1% of traffic from IE8/9 so using flexbox would be ok if I wanted. If you're working on gaming/tech things for example, chances are most of your users will be on chrome/firefox and you can safely use it.

And flexbox is so nice.

Agreed. A friend of mine suggested Flexbox for a web-app I was going to build him. I glanced over the requirements and then asked him his traffic for <IE10, and it was nearly 10%. (Website has 1m+ pageviews per month).

Not nearly as bad as 32%. And the number continues to drop.

That said, if I told people on this site that I was offering a PaaS with only 1 nine of reliability, I'd be laughed out of the park. And yet that's exactly what I'd be saying if I were to advocate for ignoring <IE10.

Sure, wasn't making any comment as to whether it's okay to ignore 10%, just that 30% <IE10 is probably an overstatement.

As for whether it's okay to ignore... I guess it really depends. I actually ended up writing a mobile web-app with Cordova, so obviously my target market had no old IE in there. It might also be true that people on old IE are people on old hardware, the type like my dad who never does any ecommerce at all, and who may not be very lucrative visitors anyway.

Anyway I wonder if someone has some actual data on this, I can make up some assumptions but I really don't have a clue. I do often read that China and Russia are big IE markets (lots of windows XP and lots of websites that still use activex stuff). It'd be interesting to see which markets are high on <IE10 and what the average value (e.g. ad or ecommerce) per user from different browsers is.

Just checked on my own website, target audience are mostly 30-40 year old moms in OECD countries. <5% IE, and virtually all IE10/11. Just 5% IE9 and barely anything below. (mostly West-European audience.)

I hardly think you're comparing anything meaningful. The hypothetical PaaS has a certain level of reliability regardless of the client your user choses to use to connect but your user's client doesn't support the server's feature set.

Anyways, I have no respect for anyone using <=IE9. I currently have to support IE8 (and 9 but that's not a big deal) because we deal with Australia and we deal with banks. At this point it's ludicrous that they haven't caught up. IE8 doesn't have the ActiveX lockin that IE6 had and current IE is immeasurably better in all conceivable ways from what's come before.

... something something inhouse-app on ie9 install base ...

Reply on news.ycombinator.com