So I wrote this Python script (blogrollfinder.py, currently a mess) that takes a URL of a weblog and returns the weblog’s blogroll, by analyzing the HTML source and extracting lists of links separated only by HTML tags and whitespace. Then I got all funky and recursive and made the script compile a meta-blogroll of all the blogrolls of the sites on my blogroll, and then a meta-meta-blogroll of all of those sites’ blogrolls. And then it tallied up all the sites and sorted them by total number of links and compared it to my original blogroll and spit out a list of people whom my community is reading that I’m not.
| Name | Links |
|---|---|
| Dave Winer | 45 |
| Doc Searls | 35 |
| Jeffrey Zeldman | 23 |
| Evan Williams | 22 |
| Shelley Powers | 21 |
| Jason Kottke | 21 |
| AKMA | 16 |
| stavrosthewonderchicken | 16 |
| Wesley Felter | 16 |
| Mike Sanders | 16 |
| Meg Hourihan | 15 |
| MemePool | 14 |
| Meryl.net | 14 |
| Lawrence Lee | 14 |
| Gary Turner | 14 |
| Heather Champ | 13 |
| Garrett Vreeland | 12 |
| Meryl Yourish | 11 |
| Steven Vore | 11 |
And so forth.
If all these people implemented the LINK tag, then my script could integrate with the news aggregator of my choice and suggest a list of feeds I ought to be reading, just like Sam wants.
