External Web Presence - overview
Building a high quality web presence that shows off what is available inside IBM Research and is easily navigated by people who don't know our organization is a daunting task. Building one that can be maintained as people and their roles and their research changes is even harder. Doing so in a consistent manner without hiring a large number of editors to patrol the output of 3,000 people is an interesting software challenge, but one that—if solved—we hope can help many other organizations.
This project is the work of various volunteers building upon a community source base. We wanted a system that automatically puts in as many cross links as possible so that, once you've found one person, paper, or project, you can find related ones. Another goal is to let IBM Researchers create a web presence for themselves that they can be proud of in only a few minutes.
The design of this system has been an interesting process of taking an initial vision coupled with continual refinement and heroic coding efforts punctuated by deep confusion. We realized that no content management technology we knew would suffice. The linkages within a dynamic organization like IBM Research are just too varied for using a technology like friends to link all the appropriate people. Many IBM Researchers would have trouble limiting their list of friends to, say, 200. And listing 200 friends is almost as useless for another person using those friends to find other interesting people as listing no one at all.
Our design uses common elements, perhaps a joint paper, a joint project or other joint activities to form almost a labeling on the links between friends. Our UI design had to make it easy for users to add that sort of labeling. Those common elements are interesting in their own right and we wanted as well to show the linkages between those elements.
As you explore our web pages you will see many links. What you won't see is that a very large percent of those links are automatically maintained.