I got it done, but it requires a password, and isn't very friendly to use. Also, there's a better way to download comments then the one I used (screen scraping the 'read comments' page). Especially since my method doesn't get all the comments in a post that's been highly commented on.
I don't have a problem with trying to make one that doesn't require a password because, as you pointed out, it just automates a manual process. Heck google probably does it weekly.
But, the time I have for such a thing is very limited. And, I'm not sure if there are any nice programmer oriented interfaces for getting all that stuff for a user other than the logged in user. I made use of a couple of such for my journal downloader.
I could probably be persuaded to post the code for mine up someplace. Right now, the username and password are embedded in the code, so I could either fix it to take them as command line parameters, or just blank them out. It's written in Python.