O...kay? Why is this news-worthy?

Was there XML previously? They're the same thing, they just use different parsers.

Was there no programmatic access previously? How about a link to where they announce this omg-new-way-to-read-News-River? Scraping looks like it would be simple enough, so I doubt lack-of-JSON-access prevented people from writing anything for changing News River's display.

Is this meant to be public? Will it go away? Why do we want News River in the first place? What is the point of this post?

Don't know the point of this post, but anything that's code or output of code gets my vote over the "everything I need to know about startups I learned in Kindergarten" submissions.

Why support the lesser of two evils when you don't need either?

Because we can only upvote.

And not-upvote.

It was not XML previously.

I will write a blog post soon explaining, but I wanted people to have a look first.

I believe it's totally newsworthy, but of course YMMV.

The blog post I'll probably end up reading if it makes it to the front page of HN. I don't often look at the new stuff (sorry, but it's true).

But as-is? It's a 1-1 duplicate of the main page (though it does separate by source) in a different format, and a strange extra layer in the JSON:

  { updatedFeeds:{updatedFeed:[{feed item}]}, metadata:{} }

Why updatedFeeds.updatedFeed[]? There's nothing else in updatedFeeds.

edit:

There's also the inconsistency that updatedFeeds.updatedFeed[x].item may be an array or an object, depending on how many items are by that source. Inconsistency == increased edge-case code == less happy users of the data.

It seems a barely-thought-out representation of the data, nothing I'd want other people to see.

Just pushed something more generic to heroku last week..

Http://jsonr.heroku.com

Converts any XML feed to JSON or JSONP. Docs coming very soon.

The host is refusing connections on port 80.

Had to fix something for a minute. It's working again. Sorry for the outage.

Pity we can't go back in time and specify something like this in JSON as the actual format for RSS. It would have prevented a lot of drama. (Not all of it, but a non-trivial amount of it.)

It's working for me now. Can bodies and titles contain HTML markup?

You won't see any HTML markup in any of it.