I think designing services first is the more appropriate thing to do. Don’t think about the UI first. UI’s are a dime a dozen and should be lowest common denominator (browser), up to native, high performance, best of breed (platform specific; iOS, Android, etc.) Because mobile is so important today you can’t leave that out of any discussion about web based services. Most web apps I’ve run on my phone are pretty horrible performance wise, but I can run them. If I have a choice to get a native mobile client I always opt for it because of that.
Dave accused me of calling him stupid.
Far from it, Dave. I know you’re not stupid, I just have a different idea than you, that’s all. A differing of opinion, and that’s ok. I think you’ll be very happy with your new model and out of that we’ll see some great browser apps.
I’ve talked about it before. I’d like to see the browser evolve way beyond what it is today if this is the new “OS” for the web based world. It must if we expect to create beautiful, highly usable, fully functional applications like we did on the desktop before the invention of the browser.
When the browser can evolve to a faceless shell that is language agnostic and allows us to control it like a native application controls the desktop environment, man, then we’ll have something special, until that time it will provide the lowest common denominator gateway to the web.
I’d love to see the browser community embrace the idea of a full CLI implementation and create a way to hit and endpoint URL that pulls a fully built application to the desktop that’s hosted by an “invisible” browser. [UPDATE: This part, about pulling a full app plays nicely with what Dave is talking about.] The browser should serve as the new OS giving the developer full access to a virtualized machine.
We’ll get there, some day.