PLoS ONE comment categorization via crowdsourcing

Last year PLoS released data on the comments and annotations left by readers of PLoS ONE papers. It's a large dataset and it'd be interesting to be able to do a proper analysis of the different kinds of comments that scientists have been leaving.

We already did something similar with a dataset from BioMedCentral, the results of which can be found on Nascent, NPG's web publishing blog.

The PLoS dataset is too large for me to go through on my own. It'd be awesome if you could help me out, crowdsourcing style. Here's how:

1. Sign in with your Google ID below (or register for a new one)
2. You'll be taken to a page that shows details of an article and all of the comments and annotations associated with that article
3. On the right hand side of each comment choose the category you feel suits it best (just ignore any edge cases)
4. Once you've categorized all of the comments on that page, click the 'next article' link
5. You'll be shown another article. Repeat from step 3.

Once we've got enough coverage - two or three opinions on each comment - I'll pick a user out of the database at random and send them some Nature conference goodie crapola (everybody likes keychains, right?).