- From: Tim Berners-Lee <timbl@w3.org>
- Date: Tue, 10 Aug 2004 12:38:29 -0400
- To: 'www-tag@w3.org' <www-tag@w3.org>
The accuracy of metadata is relies on the server administrator and the author of documents, and the software that they use. Practically, the capabilities of the tools and the social relationships may be the limiting factors. For example, Expiry Date: It is important that Expiry Date on documents be set accurately to allow caches to refresh, not wasting bandwidth, but avoiding stale data. The author (rather than the webmaster) will often be the authority on what the actual expiry date, or likely lifetime, of a document is. However, if the author has does not go to the trouble of thinking about and setting the expiry date explicitly, then the server may be in a position to guess as a function of the time since the document was last edited. Character encoding: The client editing software is in the best position to know what character encoding was used for a document. In the case of XML documents, the encoding can be deduced by reading the document, so the server software can ensure that the correct metadata is sent by sensing it (dynamically or statically). Other techniques vary from insisting on a particular encoding (and checking validity) for all files; or allowing a user interface to specify the encoding, or using a filename convention. Content type: This is often managed by a site-wide filename convention, such as mapping "*.html" to application/html+xml. This can be a problem as it makes it difficult to introduce new types, and users may have other constraints on filenames from other tools. Webmasters should allow local user control of content-type generation, so that new Internet Media Types can be introduced. A common problem is that new forms of data (e.g. Java programs or Speech dialogs) end up being served as text/plain by default, and this prevents client systems to treat the data appropriately. This also prevents future migration of the information from the existing content type to a new type in the future. Solutions include local metadata files, and user metadata control forms. In all cases the accuracy of these and other metadata is just as important for dynamic web resources, where a little bit of thought and programming can ensure the correct metadata for a huge number of resources. __________________
Received on Tuesday, 10 August 2004 16:38:34 UTC