Re: [dita] MIME type for DITA - fragment identifier issue

[Erik Hennum <ehennum@us.ibm.com>]

Hi, Deborah, Gershon, and Paul:

I'm at the Semantic Technologies conference and will have to miss the call today.

I think it's acceptable to be explicit or silent about the DITA fragment scheme in the MIME type definition.

If a content dispatcher doesn't implement the DITA fragment scheme, it means that they ignore the part of the request that addresses content within a topic. For instance, if a topic has a conref to a paragraph, the entire topic containing the target topic would be fetched and routed to a DITA handler instead of just the referenced fragment. Given that a fragment of content is not a complete DITA object anyway, it could even be appropriate to supply the full topic and extract the fragment locally. Whether the fragment is extracted remotely or locally, the dispatcher still wouldn't have to open the topic, which is the rationale for the MIME type.


Hoping that's useful,


Erik Hennum
ehennum@us.ibm.com