A Non-Business Case for Supporting Old Browsers | Hacker News Reader

A Non-Business Case for Supporting Old Browsers

ovl.design

33 pointsriledhel24 comments

I think that the best reason for not supporting old browsers is that most developers dislike it enough that it's not worth the drop in moral.

Never mind IE11. http://motherfuckingwebsite.com still works on IE6, and why shouldn't it? It's plain HTML with no CSS or JS.

There is no reason why any web document should not still be readable in Lynx, Mosaic, or even WorldWideWeb.

Web applications are a bit more complicated, but that complication is self-inflicted. We developers choose to to hijack users' devices for work that should be done on the server in the name of providing a better user experience, and we make our work harder in the process.

Supporting old browsers isn't that hard. Build a solid foundation of HTML, HTTP GET, and HTTP POST, even if it means doing full page refreshes. You can always layer your AJAX calls and such on top. It's called "progressive enhancement", remember?

That website doesn't do anything.

It's a document, not an app. What were you expecting?

Browsers aren't about just reading documents, they are about interacting with the web. That's why there's such a good argument for upgrading browsers.

It doesn't need to. The "everything must be a web app" mindset is completely unnecessary and actively harmful to accessibility. I've seen far too many "redesigns" turn a site of static pages or even forms into a horribly slow and buggy limited-browser-support SPA that excludes access for no real reason.

I think you missed the point.

This website doesn't do anything but others do. Browsers have to support both.

It delivers information. Arguably more effectively than most websites I visit for information.

It seems to be delivering almost zero information.

The information delivered is the author's philosophy on website design and implementation complexity.

I can easily tell that the author does not approve of overly complex websites, and thinks they lead to poor experiences for the developer and user.

Contrast this with a website for some hot new web 3.0 buzzthing, where "what is this thing" is not answered in lieu of presenting a shitload of "why we/this thing is so great". An example of this is npmjs.com. Nowhere does it actually state "this is a package manager for javascript", if you didn't already know it.

To its credit, at least npmjs.com is still readable without JS. However, it is lacking in terms of what it's about, and reminds me of GE.com - the latter maybe somewhat excusable because GE makes so many different things, but I don't think NPM is quite so large...

Then you're missing the point that the web is used for interactions not just information.

I think you are absolutely right with "self-inflicted" --- it brings to mind the frequent discussions here about Slack and their web-based client using hundreds of MB to gigabytes of memory for what is essentially "IRC with some images". Not too long ago they "stopped supporting" older browsers, which was basically a kick in the face to everyone for whom it was working perfectly fine before.

I remember using "webchats" in the late 90s/early 2000s with IE6 (or could be 5/5.5, not sure) on machines with probably 1/100th the processing power and memory of ones today, and they worked perfectly well; complete with things like animated emoticons (they weren't called "emoji" yet) and GIFs that didn't even lag the system unlike what happens today, and that was on a single-core CPU!

A possibly fun project is to create a Slack clone that works with that resource constraint, and uses their API; they have a WebSocket interface which older browsers are obviously not capable of using, but I believe there's a classic AJAX API too.

Use a Matrix bridge if you really need Slack, and pick one of the web-based Matrix clients, or build your own. Some are extremely lightweight.

i like how the mfwebsite loads a google analytics tag at the end which is larger than the actual website. lol.

Oddly enough, this page doesn’t appear to scale properly on Safari on iPadOS. It’s too wide, requiring horizontal scrolling. Requesting the desktop version works. I haven’t tried any debugging, so don’t know why.

Well, it's got js now.

     <!-- yes, I know...wanna fight about it? -->

Although I don't see why in this case. Owner can grep/wc the server logs to get a hit count.

As with all things it's a cost benefit analysis. It costs money to support old browsers and a lot of money doesn't come from them do do the math.

I was trying to buy movie tickets online the other day.I tried Firefox then Firefox without ad blockers, then ie, then edge, then chrome on two different desk top machines.

Then my gf bought them on her phone browser.

Thanks mobile first development. How much money must you be costing your adherents?

It costs money to support old browsers

One of the arguments the article makes is that it doesn't --- thanks to the backwards-compatible nature of the Web, anything that works on an older browser is likely to continue working on the newer one.

So I have to design and build for the older browser then hope that the newer browser handles it properly, whilst also constraining myself from using any newer functionality?

That seems backwards.

No. You have to build on the new browser, because that's what is on your new computer, then hope it runs on the old computer with the old browser.

The article is saying that you don't "build for the older browser" --- or indeed any browser specifically, but instead start off with basic HTML and adding just enough markup and styling around the content as you desire. Maybe you'll later tweak the site to look exactly the way you want in a specific browser, but this process means that even if it doesn't look perfect, other browsers will likely be able to use the site too.

The article is wrong. It absolutely does cost money and time to test on old browsers. You need multiple computers or vms set up. You don't build websites using old computers running ie6. so you have load the site on that old browser after you write the code.

It's an obvious fact to anyone who does front end web development.

Ah the mystery that is cinema websites - they always seem to be so bad. VUE's website in the uk is a complete disaster.

Reply on news.ycombinator.com