I have been talking recently about the MVC, model-view-controller, pattern (here and here), and I wanted to summarize my observation about it and what I see as its significance as a pattern of the web and of microcontent.

My observation is this: given any structured information accessible on any layer of the web, web users and systems will look to explicitly capitalize on model, view and controller functions expressed in that information.

The development of these uses will tend to push every layer of the web to become more loosely coupled between model, view and controller, with each more directly accessible.

My recent starting point for thinking about this is RSS, which I believe, in its original context, is technically a "view". That is, originally, an RSS document was not designed to be treated as a database or used (much) as an application in itself.

Now, however, as an example, Technorati treats RSS documents as databases and exposes them to queries. And, some of my RSS feed reader's operations are affected (controlled) by URIs embedded in RSS documents.

So, my point is that whatever information is served on the web, in whatever format it is in, however micro the content gets, we will look to parse and query more out of that information directly, manipulate the display of that information directly, and execute linkages and other operations out of that information directly.

I think the suggestion of the Model 2 MVC pattern in the web application / controller layer is also an example of one layer of an MVC pattern being further understood itself as having an MVC pattern. I believe parallels could be drawn in the database / model layer with regards to the relationship between base tables, views, and stored procedures, constraints and permissions.

As another example, search engine spiders capitalize on model and controller functions in HTML web pages. Search engines spiders use the page structure (including the structure of its human languages), which is all in one sense a view, to treat the page instead as a database; and search engines also allow page URIs to affect (control) their own operation.

The URL is perhaps the true atom of web MVC, though it can itself be broken down further. But, putting that aside for a moment, a URL is a view, a URL is a controller, and a URL is a model.

What is powerful and contagious about the web as a system is that each bit of it is potentially its own functioning MVC micro-system, in the context of a web browser or server. And, the functioning of these micro-systems very often links together other micro-systems.

I think one liability at the moment is actually in the boundary between MVC functions between web browser and web server. I think a lot of development around the idea of microcontent is around integrating these so that the act of browsing can capitalize more directly on each of the MVC modules (as an example, many RSS-type feed readers run as local web servers).

So, with the iCite net, I think of all of this as Fractal MVC. Each component of the iCite net tries to expose direct interfaces to each MVC module of its functions and allow each of those to be aggregated from one iCite to another. The assumption is that however atomic the information gets, someone should be able to create a system to further parse and query it, manipulate how it is displayed, and execute links and operations on it.

Finally, I should quietly note that I think that access to more and more complete MVC functionality across the web will have interesting affects on traditionally relatively centralized systems dedicated to model and controller functions ;-)