I am working on something at work that is going to be a little less than a full-blown portal (which full-blown portal, thankfully, teams of other people are always already working on here and everywhere), and I thought it might be a useful exercise to imagine this little thing I am working on being built on something straightforward like blojsom at the core.

The core function is basically many people updating different categories of information that gets displayed through a website, like articles, events, news, etc.; and the aggregation of that information into singular "portal" pages.

Blojsom can handle single and multi-user blogs, and I am thinking about different blogs (and/or different blojsom blog sub-categories) representing different content types and/or sources. Some blogs might be for events, some for news, and some might represent one department or another. In front of this, I would create portal view pages using a simple servlet filter that pulls from different blogs / blog sub-categories.

I would also create some blojsom flavors that display things like "events" in aggregator friendly format, e.g., just titles and links, that could just be simply included on portal-like pages. So, the blogs / blog sub-categories package their own content in different display flavors, and portal pages simply pull from those flavors as needed.

There would also be some interesting possibilities in using blojsom's metadata categorization to cross-reference things, or to support content types (e.g., news) from multiple sources (i.e., separate blogs in blojsom) being aggregated together.

(Note: with the iCite net, which is totally separate project from what I do at work, but might be interesting to consider for this kind of thing someday, I hope to implement a creative way for multiple sources across multiple servers to be categorized and aggregated in this way. For example, rather than a central blojsom portal, imagine different individuals and groups having their own personal portals on their own machines, and all of them being related in an essentially peer-to-peer fashion.)

Part of the idea of using blojsom would be to make use of the already existing network infrastructure for user access control, rather than creating a new one. Because blojsom uses the file system and people simply create files and store them in certain directory folders, adding users and controlling their access is relatively straightforward and can also be scripted.

What I think would make this very interesting would be integrating blojsom with Jakarta POI for reading files in Office format. Most of what goes on where I work involves people creating files in Word or Excel and saving them on hardrives. So, I have been just imagining a web interface to all those files on drive shares—also, capitalizing on individual's efforts to organize their files in directory folders, which could appear as blojsom sub-categories.

Finally, throw in a little Lucene search across all the documents and web pages—I think that would add to up a very lightweight implementation for a k-log / portal / CMS that builds on existing files and directory structures.

Now, of course, we already have Microsoft SharePoint 2 at work for much of this. It can do much of this, and it has tight Office integration. But, the question at work is to what degree we need to build a more usable interface layer outside of SharePoint (and thus, my fantasy about using blojsom) versus being able to build it in SharePoint. I shall see soon . . .