- From: Roy T. Fielding <fielding@apache.org>
- Date: Tue, 6 May 2003 17:31:31 -0700
- To: Martin Duerst <duerst@w3.org>
- Cc: "Julian Reschke" <julian.reschke@gmx.de>, <www-tag@w3.org>
> Are you saying that if a file foo.html is put as > text/html; charset=UTF-8 > but the server has a configuration for .html to > use text/html; charset=iso-8859-1, then the correct thing for the > server is to serve that file as text/html; charset=iso-8859-1, even > if that is observably wrong? No, I said: >> What the server should do is reject the PUT if the metadata is >> inconsistent with its configuration. > > Why shouldn't the server just update the configuration? It can if that is how it has been configured, assuming that the requesting user has authorization. For example, if the metadata is stored in resource forks, databases, or some other mechanism, rather than derived from the storage name. There is nothing preventing someone from using a metadata storage mechanism for determining media type in response to GET rather than the default filesystem naming configuration. Likewise, there is nothing forcing folks to use PUT in order to create website content. Personally, I hate silent overrides because they are usually due to a client mistake rather than intentions of the user. I want the server software to warn me when the client has asked to do something that doesn't match the configuration. > Is there anything in the HTTP spec that says that server > configuration is more important than the metadata attached > to a document? Only the security considerations. ....Roy
Received on Tuesday, 6 May 2003 20:29:04 UTC