[GeoJSON] Content-type

[Erik Wilde]

hello billy.

On 2013-11-03, 17:03 , Billy Newman wrote:
> To be 100% honest I am not sure why application/geojson+json would break
> compatibility.  If you are serving GeoJSON and using application/json
> and change to server application/geojson+json you are still returning
> the same thing, in this case GeoJSON.  Isn't application/geojson+json a
> sub-set of application/json?

logically it is and the structured suffix makes it look a little bit as 
if it were, but strictly speaking, media types have not hierarchy other 
than the type/subtype structure, so application/json and 
application/geojson+json are completely different things. whether over 
time the structured suffixes will create a defacto structure of 
registered media types that might eventually be codified is hard to 
tell, but as it stands now, when you ask for application/geojson+json, 
an old server would just say "no, i only have application/json", and 
there would be no matching process to connect those two in any way.

thinking of it, maybe something like http://tools.ietf.org/html/rfc4647 
should exist for media types as well, and then could take structured 
suffixes into account. but the problem also is that different 
communities have different practices around how they use media types. in 
RDF, for example, media types play a minimal role, just distinguishing 
between various serializations of the same generic metamodel, whereas 
other communities have adopted different practices, making media types 
more meaningful, semantically speaking. it might actually be pretty hard 
to unify all of those practices. which is not to say that it might not 
be a valuable exercise as an I-D, so that it's captured in one place. 
but it might turn to be an exercise where the end result is that there 
is no single simple unified model that could be used across all domains, 
so what remains would be the simple matching process as defined by HTTP.

http://tools.ietf.org/html/draft-ietf-httpbis-p2-semantics-22#section-5.3.2

cheers,

dret.

-- 
erik wilde | mailto:dret at berkeley.edu  -  tel:+1-510-2061079 |
            | UC Berkeley  -  School of Information (ISchool) |
            | http://dret.net/netdret http://twitter.com/dret |