It would be really useful for users, developers, media and MPR itself if there was one universally accessible source (API) for searching all MPR content by location, time, keyword and article. This is a simple first step at trying to quantify what that API would do.
API Methods
The following methods describe a very early vision of what the API should be able to do. These are simply ideas, and an attempt to capture as much information from the session as possible. When implementing the output, other known standards (see links below) should also be consulted and considered.
Location
This would be the primary method that allows a user to search by their own geographic location to find out what content is available from MPR.
Input parameters
Output parameters, per media item returned
- media_id - unique identifier for this particular media item
- media_type - type of media, is it an article, an audio file, a photo, video, etc
- geocode - physical location on the planet in latitude and longtitude
- timestamp - when did this item occur?
- uri - more information about this item
- summary - summary description about this item
- author - author of item
The url might look something like this:
http://api.mpr.org/location?zip=12345&radius=5&last=30&keyword=
The results returned would have one entry per media item found in the repository. This media item would be tagged by keyword, and each keyword would have a corresponding lat/long descriptor (geocode) as well as timestamp. Each media item would also have a direct URI to more information about that media item utilizing other API methods.
Media
This method would extract details of a particular media item. It would be referenced by the location method.
Input parameters
- media_id - unique identifier for this particular media item
Output parameters (would vary depending on media type)
- media_id - unique identifier for this particular media item
- media_type - type of media, is it an article, an audio file, a photo, video, etc
- geocode - physical location on the planet in latitude and longtitude
- timestamp - when did this item occur?
- uri - more information about this item
- summary - summary description about this item
- author - author of item
- full_content_text - actual text content of the media item (if an article)
- full_content_audio - base64 encoded audio of media (if an audio stream)
- what else?
The url might look something like this:
http://api.mpr.org/media?id=1234567890
Use Cases
Search by location
A client may provide an address and radius of interest. The API would return all MPR media content for the last 30 days in the containing circle around that address.
Search by location and time
A client may provide a geocode as a location, a radius of interest and also a time frame of 7 days. The API would return all MPR media content for the last 7 days around that particular latitude/longitude location.
Search by location and keyword
A client may provide a location, perhaps only zip code and a keyword to filter by, for example, "Obama". The API would return all MPR media content for the last 30 days (default time range) in that particular zip code but filtered by keyword "Obama".
Search by article
A client would provide a unique media item ID which was extracted from a location search. The API would return that particular media item, including all content and summary information.
Possible Tools for Automating content tagging:
Links of Interest, References:
There are a number of potential standards in use:
- IPTC NewsCodes Genre - The IPTC standards are used by newspaper industry, primarily for wire service syndication. Unclear about interest in other publishing formats.
- PRISM Genre - PRISM stands for the Publishing Requirements for Industry Standard Metadata, a standard of the Industry Digital Enterprise Alliance. It is strong in use by magazines, trade publishers, photography publishers. Version 2.0 of the standard was released February 2008. The Nature Publishing Group authored a PRISM module for RSS in 2004.
- Google News SiteMaps schema - Google is evolving the Sitemaps for Google News schema, but hasn't been clear on how open this process will be.
- Microformats -- Unclear whether genre is covered.
- Structured Blogging -- defunct?
- RSS Extensions wiki -- A collection of RSS extensions, none of which cover genre. (It has been spam-riddled for the past year.)
- Reuters Calais -- An ontology for business reporting. Does not include genres, but otherwise is an example of an ontology that benefitted from a well-publicized launch in early 2008.
Brainstorming Session Reference
Here's a photo of the whiteboard where we brainstormed the ideas for this API. You could also look at a higher resolution version here.
Comments (0)
You don't have permission to comment on this page.