Archive 1. June 2011

CouchDB: The Definitive Guide. And since it fits so well, here is the link to a book on CouchDB for online reading. It is also available online in German (and of course in English. And also in French). However, the German version seems to be incomplete - you keep encountering English pages.

Simple JavaScript Applications with CouchDB - CouchApp.org. Interesting approach: entire applications are implemented in HTML+CSS+JavaScript and stored directly in the CouchDB, also persist their data in the CouchDB. Reminds me a bit of Frontier, only here CouchDB and JavaScript are used instead of the Frontier Roots and UserScript. The big advantage: if the application actually gets by with the elements used, you get the easy scalability so to speak for free - because CouchDB can be easily scaled by starting multiple instances. Maybe it becomes clearer what it means in this tutorial.

Its About The Hashbangs. Blogged because it is a pretty good description of what is wrong with these #! in addresses. The part after the # is only client-side, the server never sees it - any form of server-side redirection and server-side routing is completely out of the question. Also, server-side access controls are largely out of the question, the server only sees the main page. For single-page applications like TiddlyWiki, this is all fine - everything is already in one file and it is the declared purpose of these applications to be structured this way. For websites like Twitter or even news sites like those of Gawker Media, it is simply absurd. Twitter can perhaps guarantee due to its banal structure that the URLs to tweets also work with #! in the long run, in that a corresponding JavaScript part remains in the homepage forever (which is already pretty silly), but news sites will sooner or later simply drop the old #! URLs - and thus produce massive link rot.