What is newsfeed.shtukas.net ?

newsfeed.shtukas.net is a news-feeder that I put together few hours after nytime.com published an article called Using Twitter as Your Database [open.blogs.nytimes.com, Thursday 16th June 2001], in which they explained how they used Twitter to keep track of their news stories. Inspired by this article newsfeed.shtukas.net was born a searchable database of news stories. You can think of newsfeed.shtukas.net as a Twitter service for computers. Ideally, publishing companies will register their news items to this site as part of their publishing workflow. Allowing people to have a clear, ordered, timeline of what has been published all around the world.

The service is just being born and has to be considered an alpha release. In particular submitting and searching tags has not been implemented yet.

How can I use it ? (part 1: format of the feed)

The feed

• http://newsfeed.shtukas.net/feed.php
• http://newsfeed.shtukas.net/feed.php?pretty=true (pretty print version)
• http://newsfeed.shtukas.net/feed.php?text=true (text version)

returns the last 1000 news items that have been posted to the server (in reverse chronological order -- newest first). The feed is free for all to use. The answer is a JSON string, encoded using UTF-8.

A typical item is

    {
        "record_uuid": "6b97c9af13526807f987df2ec49840df",
        "record_unixtime": 1308242512,
        "record_datetime": "2011-06-16 16:41:52",
        "client_name": "physorg.com",
        "newsitem_clientuid": "",
        "newsitem_unixtime": 1308242568,
        "newsitem_datetime": "2011-06-16 16:42:48",
        "newsitem_title": "The Port of Athens was once an island",
        "newsitem_url": "http:\/\/www.physorg.com\/news\/2011-06-port-athens-island.html"
    }

The keys of every individual items are as follow:

• record_uuid: a unique key generated by the server which uniquely identifies this item.
• record_unixtime: the unixtime of the item addition to the server.
• record_unixtime: the timestamp of the item addition to the server (refers to the exact same point in time as record_unixtime), it is given in UTC.
• client_name: the name of the client.
• newsitem_clientuid: this is a key set by the client themselves. It could be the primary key of this news item in their own database. It is not meant to have any particular meaning for consumers of the feed.
• newsitem_unixtime: the publishing time of the item as declared by the client.
• newsitem_datetime: same as above in user fiendly format.
• newsitem_title: title of the news item.
• newsitem_url: url to the online article.

I am a (private or coorporate) publisher and I want to post a news item to the feed, how do I do that ?

You will basically need an API key as it would be unrealistic to let anybody register anything without previous registration. If you want an API key, please just send an email: newsfeed@shtukas.net. The API's url will be given to you together with your key.

How can I use it ? (part 2: specifing subsets)

This section specifies how to querying the feed to get subsets of what the standard answer would be (note that the specification is a draft and may change unexpectedly, refer to this page for the latest definitions). You can specify GET variables and POST variables in your query.

GET variables

• size: controls the size of the answer up to 5000 items (five times more than what the defaut query returns). To get 6 items you will call http://newsfeed.shtukas.net/feed.php?size=6
• pretty: if equal to true, displays a pretty json. Example: http://newsfeed.shtukas.net/feed.php?pretty=true
• text: if equal to true, displays a text version of the data set . Example: http://newsfeed.shtukas.net/feed.php?text=true

POST variables

• ...part of the specification in progress...