- From: Andrew Daviel <andrew@andrew.triumf.ca>
- Date: Mon, 17 Mar 1997 18:53:40 -0800 (PST)
- Cc: meta2@mrrl.lut.ac.uk, www-talk@w3.org
The Web needs a Metadata registry, IMO. In HTML, many organisations and individuals are starting to generate Metadata using the META tag without any agreement as to what the data means. While I applaud the effort to generate useful metadata, a proliferation of unregistered types is going to dilute its utility greatly. Given that many organisations will want to use private metadata, and that particular disciplines have their own metadata, a single global repository of all metadata types is clearly unworkable. Proposal: That a top-level registry be set up at an institution such as W3.org, which would maintain a list of top-level schemas. Examples of metadata with schemas and subschemas: <META NAME="DC.Author" CONTENT="Oscar Wilde"> <META NAME="MCF.versionNumber"> CONTENT="1.3b"> <LINK REL="DC.Author" HREF="http://andrew.triumf.ca/~andrew/"> <A REL="MCF.helpPage" HREF="http://some.org/gxt/help.html">Help on GXT</A> The schemas DC (Dublin Core), MCF (Meta Content Format) etc. would be registered with the top-level registry. The organisation registering the schema would be responsible for documenting it. Anyone using a registered schema would hopefully have read the documentation first. Anyone creating a new schema-less metadata type is on their own - subject to misinterpretation. Some existing ones are pretty well defined, though: HTTP-EQUIV="Refresh" - Document reload with optional URL NAME="Description" - Document description for search engines NAME="Keywords" - Additional keywords for search engines NAME="Robots" - automated agent control NAME="Generator" - Added automatically by publishing tools others aren't so well defined, e.g. "Classification" Andrew Daviel TRIUMF, Vancouver Webpages
Received on Monday, 17 March 1997 21:54:21 UTC