[abalashov]

"With the maturation of browsers, many features that used to require dynamic code running on a server can be moved entirely to the client. Want comments on your website? Add Disqus, Isso or Facebook comments. Want social integration? Add Twitter or Facebook’s JavaScript widget to your website. Want real-time data updating live on your website? Add a squirt of Firebase. Want search? Add Swiftype. Want to add live chat support? Olark is there. Heck, you can even add an entire store to a static website with Snipcart.

The list goes on and on, as a whole ecosystem of purely browser-based add-ons to websites is emerging. Apart from that, modern web apps built with Ember.js, AngularJS or React are often deployed entirely as static websites and served directly from a CDN with a pure API back end that’s shared between the website’s UI and the mobile client."

--

I'm not sure I understand. It doesn't seem to me that a fully single-page, AJAX web site is truly "static". If much of the utility and content must be paged in via client-side JS calls, that too will contribute to load time and the same problems that are attributed to dynamic document generation. It may be all asynchronous and fancy, but from a UX point of view, the content isn't there until the data's retrieved. How's this any different than arguing for a grid of IFRAMEs?

After all, if your page is a minimal HTML DOM harness for a bunch of JS, can one really be said to have "loaded" the page simply in virtue of having loaded the stub HTML?

Or is this argument based mainly on either the implicit premises that (1) not all the functionality and components are used at once? or (2) that much of any given site's functionality can be off-loaded to third-party components (e.g. Disqus) which can all be loaded in parallel from different network sources?