Saturday, February 3

Indexing by Likeness & Difference

Naturally, just a couple of minutes after posting an early draft of an indexing sheet, I come up with an even better approach. Instead of trying to create a complete genome for comics let's instead create a comic book index by comparing books individually and let the genome grow when an extra "gene" is needed.

Starting with any two random comic books we'll write down one thing that is similar and one distinct difference in these two books. We then pick another comic book and make sure that we find at least one likeness and some distinction towards the other two books respectively.

We keep doing this with an ever growing genome. Every so often we will have to go back to previously indexed books to consider newly found genes. As long as every comic book is uniquely indexed we don't have to add new genes, only when two books have the exact same index we'll look for new distinctions.

This approach has one major advantage in that we only consider properties that are actually used so the genome will grow organically. I'll try this out for a couple of hundred books and see where we land.


Robotacon said...

I didn't just write "a couple of hundred books" did I?

Anonymous said...

Good work in your comic book blog, it's a nice one. I have a new comics directory site hope you will like it as well comic books directory. Complete links of site all about comic books and more...