This entry originally was a response to Heath's comment on my last entry on my ChickenScratch project, but it started to get long, so I figured I'd make it an entry.

He asked about my ideas for relational algorithms between topics.  The first phase of "auto-linking" is based on matching phrases in content with available topic names and meta-data like categories/properties with the content of entries.  Later, I might do things like Google-type analysis of content and present these “related“ topics in a way similar way to the “ads“ you get in Gmail.  We use FlexWiki at work to manage all our projects and their bugs, feature requests, and infrastructure projects, and more than once we've had “happy accidents“ in auto-linking within that data that have led to better design, and FlexWiki uses a very simple topic matching algorithm.

He also mentioned hierarchy.  Hierarchy is a key reason for me to create something new rather than retrofit one of the Wiki implementations available like the excellent FlexWiki. I love the "magic" linking and organizational ability of Wikis, but I don't like giving up the ability to direct organization through hierarchy.

Security and permissions is also core to the goals of ChickenScratch.  It's always frustrating to me when I have to separate my public information from my private information from the information I only want to share with friends just because a program doesn't let me.  My security model is working pretty well, and was pretty easy to implement thanks to the new Role and Membership providers in Whidbey

I'm definitely open to collaboration.  I'd like to solidify my ideas a bit more.  I've changed my parsing model about 4 times now.  What I've got going on now appears to be working out pretty well.