OpenSearch Description
Dan Tobias (Talk | contribs) (Created page with "{{FormatInfo |formattype=electronic |subcat=Web |released=2005 }} An '''OpenSearch Description''' is an XML file placed in a website and referenced in a link tag in a web ...") |
Dan Tobias (Talk | contribs) |
||
| Line 6: | Line 6: | ||
An '''OpenSearch Description''' is an [[XML]] file placed in a website and referenced in a link tag in a web page header in order to instruct software on how to access the site's search feature. It was designed to be used by search aggregators which search multiple sites at once, and is also used in programs such as the Firefox browser that give choices of which site to use to handle searches in its search bar. | An '''OpenSearch Description''' is an [[XML]] file placed in a website and referenced in a link tag in a web page header in order to instruct software on how to access the site's search feature. It was designed to be used by search aggregators which search multiple sites at once, and is also used in programs such as the Firefox browser that give choices of which site to use to handle searches in its search bar. | ||
| − | The format specs call for use of a namespacee | + | The format specs call for use of a namespacee [[URI]] <code>http://a9.com/-/spec/opensearch/1.1/</code> which has for many years given a "404 Not Found" error, as this project has moved from Amazon's A9 server to opensearch.org, and then to Github as the opensearch.org domain was transfered to an open-source project to build searching software. This makes finding the actual spec rather difficult now, and is grist to the mill of the endless debate about whether using "http" URIs as namespace identifiers is really such a good idea; it awkwardly mixes together the concept of having a permanent identification string to signal adherence to a particular specification (which is required by its very nature never to change) and the concept of giving information on where to locate the actual specification (which is subject to constant change as organizations change; they're only as stable as the domain name owner wants them to be). The result is endless forum arguments: "That namespace URI is '404 Not Found'!" "Well, that's not actualy intended to be an actually-dereferenceable address; it's only there to identify the namespace!" "Then why not use some other URI scheme that's clearly not dereferenceable, like 'info:'?" "Because we'd like to have the possibility of providing information on the spec at the URI given." "So why don't you in this case?" "Because the owner of the domain doesn't want to." |
Anyway, all of that makes it hard to find the actual spec; by using the archive.org Wayback Machine on the original URI and browsing around a bit you can eventually find it, but you can skip all that fun because it's linked below. | Anyway, all of that makes it hard to find the actual spec; by using the archive.org Wayback Machine on the original URI and browsing around a bit you can eventually find it, but you can skip all that fun because it's linked below. | ||
Latest revision as of 21:07, 6 June 2021
An OpenSearch Description is an XML file placed in a website and referenced in a link tag in a web page header in order to instruct software on how to access the site's search feature. It was designed to be used by search aggregators which search multiple sites at once, and is also used in programs such as the Firefox browser that give choices of which site to use to handle searches in its search bar.
The format specs call for use of a namespacee URI http://a9.com/-/spec/opensearch/1.1/ which has for many years given a "404 Not Found" error, as this project has moved from Amazon's A9 server to opensearch.org, and then to Github as the opensearch.org domain was transfered to an open-source project to build searching software. This makes finding the actual spec rather difficult now, and is grist to the mill of the endless debate about whether using "http" URIs as namespace identifiers is really such a good idea; it awkwardly mixes together the concept of having a permanent identification string to signal adherence to a particular specification (which is required by its very nature never to change) and the concept of giving information on where to locate the actual specification (which is subject to constant change as organizations change; they're only as stable as the domain name owner wants them to be). The result is endless forum arguments: "That namespace URI is '404 Not Found'!" "Well, that's not actualy intended to be an actually-dereferenceable address; it's only there to identify the namespace!" "Then why not use some other URI scheme that's clearly not dereferenceable, like 'info:'?" "Because we'd like to have the possibility of providing information on the spec at the URI given." "So why don't you in this case?" "Because the owner of the domain doesn't want to."
Anyway, all of that makes it hard to find the actual spec; by using the archive.org Wayback Machine on the original URI and browsing around a bit you can eventually find it, but you can skip all that fun because it's linked below.
MediaWiki adds by default an OpenSearch Description to sites using that software, in a page name of opensearch_desc.php in the site root.
