« Please join us for the 2009 CrossRef Technical Meeting. | Main | CrossRef Labs »

nature.com OpenSearch: A Structured Search Service

opensearch-triptych.jpg
(Click panels in figure to read related posts.)

Following up on my earlier posts here about the structured search technologies OpenSearch and SRU, I wanted to reference three recent posts on our web publishing blog Nascent which discuss our new nature.com OpenSearch service:

1. Service
Describes the new nature.com OpenSearch service which provides a structured resource discovery facility for content hosted on nature.com.
2. Clients
Points to a small gallery of demo web clients for nature.com OpenSearch which all use the text-based JSON interface.
3. Widgets
Introduces the new nature.com search desktop widgets which interface with the nature.com OpenSearch service via an RSS feed. (See also the screencast posted to YouTube.)
We hope that this new search service will prove to be useful and may also provide a model for other implementations.

Comments

Very very cool. Count me in as a fan. OpenSearch is cool as hell, we're experimenting with this as well.

Are you guys looking into representing search results as an Atom feed?

Hi Gudmundur:

Thanks for the comment. Regarding your question about ATOM you will see from the OpenSearch description file at

http://www.nature.com/opensearch/opensearch.xml
or from the SRU entry point at
http://www.nature.com/opensearch/request

that we are supporting multiple media types - both XML based:
  1. SRU
  2. RSS
  3. ATOM
and text-based:
  1. JSON
  2. JSONP
  3. HTML
From an OpenSearch perspective you (or your user agent) would just select the ATOM template, i.e. the URL template with the associated ATOM media type. From an SRU perspective you would just submit an SRU request with an HTTP 'Accept' header indicating preferred MIME type (i.e. 'application/atom+xml'), or alternatively you could pass that MIME type explicitly in the SRU request using the SRU 2.0 parmeter 'httpAccept'.

So, yes, we do serve up ATOM as well as RSS.

For your information we also use ATOM as the basis for our JSON profile, i.e. the JSON is a fairly direct mapping of an ATOM document with some obvious changes in respect of JSON not allowing repeated object keys (so theese are mapped to single key with an array value type).

Do feel free to ask for any further information or clarification on our response types.

Cheers,

Tony


Post a comment

(If you haven't left a comment here before, you may need to be approved by the site owner before your comment will appear. Until then, it won't appear on the entry. Thanks for waiting.)