[Jimmie]

How hard was it to interface with HN? Do they provide an API to work with?

Reddit has a very nice API, one of the killer features is that if you take any reddit page and add ".json" to the address it'll spit out pure json at you.

For example: http://www.reddit.com/r/all becomes: http://www.reddit.com/r/all.json

[duck]

There is an unofficial api at http://api.ihackernews.com/. It use to have a lot of timeout issues, but seems to be running a lot better now.

[joe_the_user]

Having sampled it, I'm not sure of the value of a broken api that might improve later.

[marketer]

Screen scraping. HN doesn't have a JSON feed, but it has fairly simple and predictable HTML.

[joe_the_user]

How did you process the html?

[joe_the_user]

I'm playing with the Reddit api.

It works but what it gives you is complete dump of their page, 150K, say.

It would be nice to get just the tops of the comment trees instead. But hey, its there...