Plain HTML: Back to the Roots

Forget application servers, content managers, JSF, ruby on rails... When I work on a web page or small web app for me, I try as far as possible to go with plain HTML and JavaScript. I think I need to justify myself now to why I'm mostly sticking to that tech.

- It's elegant. Probably my favorite reason - the code is all consistent because there's no other dependency than the browser. So the code can be capitalized correctly and formatted properly. It's very compact, so it has probably less bugs than in some more chatty format that many frameworks usually produce.
- It keeps you in contact with the underlying technology. And the web technology is advancing fast. Chance is your plain content will be able to do always more (CSS animations anyone?) with so less code than a non-native framework. Why using any framework function when the browser can do it alone? It's often faster and more reliable with the browser technology, and that way also doesn't need bridge technologies like external renderers (Flash! Silverlight!). Often, frameworks are a bit behind, so it's always good to be up to date and know when to do simpler.
- You can use your usual tooling. You never know when a framework will need a strange compiler or Java library that is cumbersome to get and use. That risk just isn't there with plain HTML/CSS/JavaScript. You can just start your favorite editor and push updates to the server with any tools you're comfortable with, from the file manager to an automated rsync script.

Of course, almost the same could be said of Unix tools - except that the Unix command line doesn't have ground-breaking changes anymore...

Edit 10/08/2011:
Thinking back about this post remembered me how Amazon made EC2 available as a static page host platform. Of course, EC2 doesn't imply a back to the root approach - you could use GWT's compiler, over even a homebrew code generator to produce a static site from dynamic data - but it fosters it. Github is also proposing static page hosting.