A new content management system has only then the right to be born, if it captures a new paradigm.
The paradigm shift can affect information architecture, interaction or technology. Let's have a look at current popular systems. Since I'm an information architect, I'll focus on the structural level of these systems.
Typo3 treats a website like a book. It’s a compendium of hopefully well organized, hierarchically structured pages. As a user, you have to engage into its structure before you can make sense of almost any page. Ideally, you are rewarded by insights in the end. It's a didactical approach.
A paradigm shift came with Wordpress. At its core, Wordpress is one database with nodes. Hence, it is good in two things. You can display a single node in a page. And and you can display a selection of nodes as lists of any kind. Single pages or lists, by default ordered chronologically. Additionally, you can filter and navigate by two different meta data types: tags and categories. This lead to different navigation paradigms and behaviour. Search, sort, filter rather than navigating the tree.
Drupal, by the way, is not much different from Wordpress concerning the display of content. It is more flexible, hence is more complex and more difficult to use, if not configured properly. Lots of content types, lots of meta types, complex database structures. Drupal beats at the level of user personalisation and community features in general. If you needed complex structures, or/and if you needed a platform for a community, Drupal would be the solution.
No Content Management Systems, though, could help you properly when it came to commerce. Their core architecture was not suited to an online shop. There are plugin workarounds for basic shopping. But as soon as it comes to e.g. personalised recommendations, another paradigm was needed. Thus, systems like Magento filled the void.
So if you intend to build a new CMS, answer the question first: what is the new paradigm you want to address? If you rather want to make the known paradigms better, you'll probably be better off contributing to the existing communities. Or you'll write your own extensions/modules/plugins (which I rarely recommend, since their maintenance costs are seldom calculated). For a very specific need, you might rather want a framework like Django or YII.
If you see a new paradigm on how to look at the user needs, the web, the content, the interaction - I'm curious to hear your voice.
(cc-by-sa) since 2005 by Konstantin Weiss.