OK, so the Trackback specification includes the ability for a pinging page to specify its article title and blog name when it sends a trackback ping. However, these fields are optional. When not present, typically the recipient of the ping will then make an HTTP request back to the pinger to determine the article title from the page's <title> element. But there is no standardised way of determining the blog name -- or more generically, because Trackback can be used by sites other than blogs -- the name of the "collection of web pages".

The following informal specification suggests a method of determining the name of the collection of documents to which a given HTML document belongs. It also provides a clear way for you to mark your own HTML documents as being part of a particular collection.

Determining the Collection for a Given Document

Attempt to find the name of the collection in the following places. The first successful result should be considered the name of the collection.

  1. Check for the presence of an HTTP Header 'X-Blog-Name'. If this exists, the value of the header is the name of the collection.
  2. Check for the presence of an HTTP Header 'Link', formatted to the specification for HTTP link headers, with a rel value of "blog-name" and with a non-empty "title" value. The name of the collection is the "title" value. Additionally the URL of the main page of the collection can be found in the "href" value.
  3. Check for the presence of an HTML <meta> element with the value of the "name" element of 'Blog-Name'. If found, the name of the collection can be found in the "content" attribute.
  4. Treating HTML <meta> elements with an "http-equiv" attribute as being equivalent to real HTTP headers, repeat steps #1 and #2.
  5. Check for the presence of an HTML <link> element with a "rel" value of "blog-name" and if found, use the same semantics as #2.
  6. Repeat steps #1 to #5 looking for 'index' instead of 'blog-name'. 'x-index' SHOULD NOT be accepted.
  7. Repeat steps #1 to #5 looking for 'top' instead of 'blog-name'. 'x-top' SHOULD NOT be accepted.

String comparisons for HTTP header names, HTML <meta> name and http-equiv values and Link rel values MUST be case-insensitive.

Marking the Collection on your Documents

Mark your document with the collection in a manner that can be found using the algorithm described above. Multiple methods MAY be employed, but in this case, note that only the first matching collection name will be used.

It is advised that producers take advantage of the HTTP header methods to reduce bandwidth and parsing requirements on the part of the client.

Although implementations are required to make string comparisons case-insensitive, producers SHOULD use the case specified in this document for output.