Even not considering that HTML5 History API, is it hard to just consider "/1234/spam" and "/#!1234/spam" synonymous and use the latter when you navigate with JavaScript enabled?
I.e.:
1. You visit "/1234/spam", and get served full HTML page with some (optional) JavaScript for progressive enhancement.
2. You click on "/4321/ham" link, but JavaScript hooks it up and replaces URI with "/#!4321/ham" (if your browser don't support HTML5 history API, of course). Yes, there is one full-page reload.
2.1. (Alternative) Or - even better - you can be redirected to "/#!1234/spam" on step 1, so you won't notice the glitch on your first click.
3. You continue to navigate with AJAX, now without reloads. You can bookmark pages and so on (and if you somehow happen to lose JavaScript support you could just remove "#!" to get a valid URI).
Very simplified implementation cost:
1. `$("a[href]").attr("href", function() { return this.href.replace(/^\/(?!#)/, "/#!"); });` on every page
2. `if (location.pathname === "/" and location.hash.match(/^\/#!/) { $("#content").load(location.hash.replace(/^#!/, "/") + "?format=content-only"); ... }` on / page.
3. Ability to return undecorated ("article-only") content on `/1234/spam?format=content-only` requests.
You just described exactly what you're _supposed_ to do in a progressively-enhanced single-page Javascript application. Build the existing site first without a single line of js, develop the hash-bang/HTML5 history syntax on top of it, and replace the links if javascript exists.
The Gawker family of sites failed at doing all of this.
I.e.:
1. You visit "/1234/spam", and get served full HTML page with some (optional) JavaScript for progressive enhancement.
2. You click on "/4321/ham" link, but JavaScript hooks it up and replaces URI with "/#!4321/ham" (if your browser don't support HTML5 history API, of course). Yes, there is one full-page reload.
2.1. (Alternative) Or - even better - you can be redirected to "/#!1234/spam" on step 1, so you won't notice the glitch on your first click.
3. You continue to navigate with AJAX, now without reloads. You can bookmark pages and so on (and if you somehow happen to lose JavaScript support you could just remove "#!" to get a valid URI).
Very simplified implementation cost:
1. `$("a[href]").attr("href", function() { return this.href.replace(/^\/(?!#)/, "/#!"); });` on every page
2. `if (location.pathname === "/" and location.hash.match(/^\/#!/) { $("#content").load(location.hash.replace(/^#!/, "/") + "?format=content-only"); ... }` on / page.
3. Ability to return undecorated ("article-only") content on `/1234/spam?format=content-only` requests.