There’s been some buzz about ‘AJAX’ apps this week, and the one thing they got right is that — long term — proprietary web layers like Flex and Laszlo are dead in the water (the jury’s out on Avalon-based web apps simply because of Windows’ huge potential reach).

That said, AJAX is to rich web apps what C-based CGI apps were to web apps, period: an early, messy, easy way to give new features to users. I’ll be disappointed if in five years we’re still writing pages of Javascript just to get an autocomplete dropdown. If anything, the web of the future has already been imagined, and it wouldn’t surprise me if the Googles and Yahoos weren’t working towards it now (we know Microsoft already is, with Avalon’s blurring of the desktop app/web app boundary).

The interesting this is, the unsolved pieces of a standards-based rich web app become much more solvable given a popular open-source browser (check: Firefox), browser/plugin vendors working together (check: WHAT-WG and the plugin alliance) and a powerful compilable open-source runtime environment that binds it all together better than Javascript can (here’s where an open source Java — or a mature Mono — could help).

As Joel Spolsky noted, the pieces of the puzzle that remain unsolved are a modern widget set and compiled code that can access the browser DOM. To this I’ll add the problem of offline operation. The first two aren’t rocket science and are (say) 2 years work for a company like Google. The third is a deeper problem (and the fact that Alchemy is still not available publicly shows how far we have to go); however looking at Lotus Notes/Domino’s relatively successful replication feature we can say it is not unachievable.