Re: What to call the RSS file

[mike@datachannel.com]

> 
> Dispatching and versioning in XML, with the possibility of multiple
> formats, namespaces and versions, is said to be too complex for the
> internet media type (MIME) system. text/xml lets you know that you
> should point an XML Parser at it, but what about beyond that? 
RFC3023
> makes recommendations about further gradations of information (e.g.,
> application/soap+xml), but it still isn't adequate for some things,
> depending on who you talk to / are yelled at by.

I use "Content-Type" to specify the low-level format 
(like "text/xml"), but like you said, that isn't enough.

So I've used "Content-Model" to identify the 'meaning' or 'structure' 
of the data. Ideally it would be multi-valued to allow for different 
meanings of the same bag of bytes - for  
example "ldap.user,siebel.customer" - but some people have a hard 
time with that, so I often use just a single value.

This can also be used in a corresponding 'Accept-Model' header when 
asking for information - like when a browser says it accepts both MS-
Word and HTML, and it lets the server decide what to send.

Some issues are:
 - who decides the values?
 - what is the syntax? (URI or 'dot' notation? either can be rooted 
in the DNS namespace).
 - will people understand that this isn't 'xml-namespaces'?


Mike