Re: MathML - missing Media Type Registration template

Le 30-mars-09 à 22:39, Chris Lilley a écrit :
> PL> Without being certain the three mime-types we would propose  
> could be:
> PL> - application/mathml+xml (this one for sure)
> PL> - application/presentation+mathml+xml
> PL> - application/content+mathml+xml
>
> I believe that the "+" is reserved, so I would suggest
> - application/mathml+xml (this one for sure)
> - application/presentation-mathml+xml
> - application/content-mathml+xml

I guess I should not argue with you but the usage of + in RFC 4288 is  
explained as:
> In accordance with the rules specified in [RFC3023], media subtypes  
> that do not represent XML entities MUST NOT be given a name that  
> ends with the "+xml" suffix. More generally, "+suffix" constructs  
> should be used with care, given the possibility of conflicts with  
> future suffix definitions.

And in this case, I believe the care for future suffix definition is  
cared for:
we really expect any mime-type following the glob application/*+mathml 
+xml to be related to MathML!

> PL> The hot debate is whether 3 mime-types run a risk of being  
> refused by
> PL> IETF or W3C liaisons at a relatively late stage.
>
> Yes, there is that risk. In particular is something that accepts the  
> first one expected to accept the other two (subset?) ones as well?

I will try to discuss this on ietf-types mailing-list if it's ok with  
everyone.

I strongly believe that specialized mime-types will be useful in the  
future!

> The only analogous situation I can think of is X3D which has three  
> media types for three different encodings:
>  model/x3d+xml [...]
> While yours is more a supertype plus two subtypes

Right, quite different. It may be we're the first to try that, I'd ask  
this on the ietf-types list.

paul