Rethinking a Web Operating System
The mythical WebOS has always been a grail. I remember back in college in ‘03 or so I made some progress on one, using PHP on the backend with lots of XHR polling.
(Aside: DHH hadn’t even announced Rails as a thing yet, just to make myself feel old here. Pretty sure people were still talking about DHTML as a thing.)
In the late ‘90’s and early ’00s, a WebOS was literally an operating system ported to the web. Remember g.ho.st? Desktop.com? Yeah, me neither.
The problem was that we were trying to replicate lots of the functionality that a user perceives as the “Operating System” in the browser. Clocks. Multitasking. Windowing. Search. Things that the guest operating system already gave us.
Moving some computation onto a server made sense, but lots of stuff was just silly. g.ho.st’s window management would be redundant three times over today - I’ve got my native OS providing me windowing in addition to my browser providing multiple tabs, in addition to whatever app I’m using giving me an interface.
It’s time to re-think what an operating system provides for an end user, and think about bringing that into the browser level or the webapp level. I’ve got some ideas - I’d love to hear yours.
