Shanti’s Dispatches

I’m becoming convinced that del.icio.us may be Knowledge Management’s first Killer App.

It’s simplicity lends itself to Killer-App-Dom. Not some souped up $3M J2EE-Oracle-Web_Services piece of crap a group of 15 overpaid consulting services guys would come up with for a Fortune 500 Company. Sure, management may read the latest book on KM (knowledge management) and think they need that. Sorry - you’re employees are going to hate you. Give them an Intranet-only del.icio.us server, and let them go crazy.

There’s one piece that it’s missing for uber-KM, that perhaps would be better addressed in a seperate web app.
- SuperMemo-like functionality. More about SuperMemo here.

Ironically, I came across SuperMemo while browsing del.icio.us’ popular links page.

What this uber-cool KM++ web app would be, is, I would call, a “social factdexing service.” Instead of bookmarking, you enter each fact as a discreet question & answer.

This service would provide functionality to quiz yourself on your knowledge base, and would use SuperMemo-style learning & memory retention algorithms to ensure that you don’t forget too much of your knowledge base. This would be an optional feature. It could also just be used as a braindump resource, and you could use it as a reference-only KM tool.

This factdexing service would have the following features:
- ability to tag facts/knowledge with keywords (a la del.icio.us)
- ability to set a user-default of private, as well as a per-post setting of private (i.e. for personal KM u don’t want to share
- the stream of facts (as shown on its main page) would all be public, and with a little user-moderation, globally applicable.
globally applicable mean, for example, that you have created a piece of knowledge (a question/answer pair) that anyone in the world could potentially make use of. i.e. in the public stream of Knowledge, we don’t want an item like “Where are my scheduling documents located on the Intranet?” - this should be tagged as private.
- ability to create groupings of knowledge, i.e. US Capitals, CSS Fundamentals, etc. This would go beyond the distributed tagging. You could have namespaces, etc.

The inherant flaws with this kind of system:
- Duplicate facts/knowledge. Whether this would be a “flaw” - I’ll leave up to the court. It’s an important data-model issue, and goes to the scalability of this kind of system. But it’s important that I be able to create any piece of Knowledge on the fly and add it to my KM stream. Search functionality could alleviate some of this. Perhaps you want to add, “How do I enable directory-security in Apache using htaccess files?” Type in a search for “apache htaccess” and you discover that someone has already entered this item. Test it out, make sure the piece of knowledge is correct/accurate, and if so, add this knowledge item to your stream.
- Users adding personal-specific, not “globally applicable” knowledge to the public stream. This could alleviated some by user moderation. If you see a non globally applicable piece of knowledge in the main stream, flag it as “private knowledge” and over time the system learns.

I’ll come back to this post & update it… I think there’s something here, and someone really should implement this, but I’m not going to be that guy, at least not right now. My plate is kinda full right now. Not to mention my day job.

You can bookmark this, digg this or check the cosmos