Re: Media type recognition by HTML user agents

I think we are OK, here.

to follow up on what Jason White said:

> In the minutes of the recent HTML WG audio call, there was said to be a
> need to "partition the name space" of media types. As I understand it, the
> intention is to reserve a media type called "other" (or an equivalent
> name), and then to limit HTML user agents to the list of base types which
> has thereby been established.
> 

This was a comment from a participant.

The action taken was that this is on the list of topics where the
next step is for the editors to put our recommendations into the
document.  HTML Working Group members can challenge this on the
basis of what they read in our report, and other can challenge them
once they go public in the Proposed Recommendation.  But the
HTML WG policy is that what we said goes in unless someone takes
the initiative and can prove that there is a good reason not to.

The requirement of record that the editors are working from
states that we recommend that HTML user agents are _not_ limited
to the list of base types published in the HTML spec.

> I do not understand the rationale behind this proposal, and nor is it
> necessary. User agents should simply ignore any base type that (1) is not
> one of the predefined base types in the HTML specification; and (2) has
> not been defined by the style language which is currently in effect.
> 
> A minimally conforming HTML user agent would ignore any media type that is
> not exactly one of the predefined types mentioned in the specification.
> 

Almost.  

Replace "exactly" with "in its base type, extracted as per the
algorithm in the spec," and you are OK.

> If any parameters are added to the base type which the user agent can not
> interpret, then the resource to which the media type relates (E.G. a style
> sheet) should not be accessed.
> 

It is probably better that the user agent ignore the unknown
parameters, not the stylesheet, if it is for a known base type.
This is more in line with Internet and HTML tradition.

I believe that is the way what is headed for the editors' output
would work.

We get a cut at clarifying this in the Browser Guidelines.

-- Al

Received on Friday, 24 October 1997 10:01:22 UTC