I spent some time today looking at the Technorati API and coding part of a personal tracking application. The API server was spotty throughout the day, making testing difficult. I cleaned up some of the wiki documentation, stored my own copies of API responses, and used my own servers to pull the data.
Some things I noticed:
- Cosmos query never returns
atomurlelements even though the data is stored. Use the
urlas a parameter in a bloginfo query to pull this data.
langelement in bloginfo is undocumented. 31411 for English, 1065 for Japanese, not sure what else.
- Technorati picks up a lot of duplicate references. A weblog entry will often be counted twice: one for the home page and once for the individual entry pages.