What Portals Should Be

I've worked on the implementation of several portals, but always from the content side, never the back-end technology. I don't know what goes into the guts of a portal that makes it work, but I know it's complex. Most portals cost about $1 million and take about a year to implement. Customization and personalization are the keys to a portal, particularly the selection of portlets that pull targeted content from anywhere on the network into a box on your screen. Given these options, any employee can collect a page of portlets that contain the information, data, and processes he needs most often. A lot of portal projects fail because the infrastructure works just fine, but there aren't enough portlets for most employees to make that level of personalization worth the hassle. I've been playing with a service called del.icio.us and another through which you can conduct online chats. The second kind of portlet is the one that intrigues me. It's an RSS feed. Add any URL to an RSS feed, and a new portlet window displays the feed's most current contents. The combination of "flakes" that are programmed to deliver rich content and feeds works extremely well. You can create tabs in which to collect your portlets. The default setting gives you one called "work," one called "fun," and one called "reading." You can edit these and add new ones. On any given page, you can drag your portlets into different positions, tailoring the view of that page's portlets to your interests.