I guess I’m a contrarian, I think “all websites work perfectly” when they are allowed to be web sites instead of “apps”.
“Heavy amounts of testing” comes from trying to use the web as a combination of magazine publishing design style and Flash animation interactivity. If you used hypertext with semantic markup, you’d work in anything including Lynx in a terminal and text readers for the blind. Add a minimalist CSS such as https://milligram.io/, change font, spacing, colors of elements, and be done with it.
If you use the web just a piece of paper to convey information, think Economist or The Atlantic instead of Cosmo or Us magazine, it all pretty much “just works”.
By contrast, if you use the web as a SPA (single page app) engine, you’re gonna have a bad time. For that, test in Safari for iOS, Chrome on Android, so you work in their native embedded engine. Given they’re both originally WebKit / KHTML, you’re not too far apart.
This approach downscopes your test footprint dramatically. You’re still worried about “duplicates code into different paradigms” but (a) you should have different paradigms for Android and iOS, they have a very different user expected feel, and (b) its really only the 2 paradigms with 2 modalities each: iOS/Android, phone/tablet.
Recognize this is a contrarian view when these days it’s popular to try to shoot for a single experience on every form factor, and easy to forget most content on the web is still not in app form, it’s just stuff to read. HTML 3.2 / 4 with CSS wasn’t a terrible place to live.
Fun fact: “Internet Explorer 5.0 for the Macintosh (really!), shipped in March 2000, was the first browser to have full (better than 99 percent) CSS 1 support, surpassing Opera, which had been the leader since its introduction of CSS support fifteen months earlier.”
“Heavy amounts of testing” comes from trying to use the web as a combination of magazine publishing design style and Flash animation interactivity. If you used hypertext with semantic markup, you’d work in anything including Lynx in a terminal and text readers for the blind. Add a minimalist CSS such as https://milligram.io/, change font, spacing, colors of elements, and be done with it.
If you use the web just a piece of paper to convey information, think Economist or The Atlantic instead of Cosmo or Us magazine, it all pretty much “just works”.
By contrast, if you use the web as a SPA (single page app) engine, you’re gonna have a bad time. For that, test in Safari for iOS, Chrome on Android, so you work in their native embedded engine. Given they’re both originally WebKit / KHTML, you’re not too far apart.
This approach downscopes your test footprint dramatically. You’re still worried about “duplicates code into different paradigms” but (a) you should have different paradigms for Android and iOS, they have a very different user expected feel, and (b) its really only the 2 paradigms with 2 modalities each: iOS/Android, phone/tablet.
Recognize this is a contrarian view when these days it’s popular to try to shoot for a single experience on every form factor, and easy to forget most content on the web is still not in app form, it’s just stuff to read. HTML 3.2 / 4 with CSS wasn’t a terrible place to live.
Fun fact: “Internet Explorer 5.0 for the Macintosh (really!), shipped in March 2000, was the first browser to have full (better than 99 percent) CSS 1 support, surpassing Opera, which had been the leader since its introduction of CSS support fifteen months earlier.”