Word documents and content_type_norm

[tokee]

If we do a search for content_type_ext:doc AND content_type:"application/msword" in the Danish Netarchive Search, we get the facet for content_type_norm:

• other : 3577875
• word : 17606

There seems to be a problem with deriving the normalised content type with Word documents?

Maybe a more overall issue would be to search for all records that has other as nrmalised content type and facet on the different content type fields to see if there are more heavy hitters that are not handled?

[anjackson]

This may be related to #289 where at least part of the problem is that the content type does not fall back on the content_type_served when format identification via Tika/DROID fails. EDIT: Hmm probably not.