[GeoJSON] Fwd: [Mass-Market-GEO] Paul Ford on GeoJSON

[Howard Butler]

On Nov 7, 2013, at 1:53 PM, Raj Singh <rsingh at opengeospatial.org> wrote:

> Interesting mention of GeoJSON in this Businessweek article on the Twitter encoding specification. 
> 
> discussion topics:
> - since GeoJSON is used by Twitter, it probably surpasses KML as the most prevalent geodata format in usage, with 600,000 geo-referenced tweets per day [1]. 
> - the POI SWG is considering GeoJSON as the default encoding format
> - OWS-10 is considering GeoJSON as the encoding format for geodata change logs
> 
> Is it time for OGC to formalize usage of GeoJSON in an international standard?

Haha serious, but are you trolling us?

Please sell us on the benefits of formalizing recognizing GeoJSON as an OGC "international standard" that it doesn't already enjoy as a defacto standard by already winning over the javascript market as the geo-geometry transport for the current generation of software.

Please recognize that the process that produced GeoJSON is an intentional and explicit reaction to the broken standards that OGC was producing at the time (and maybe still is, but maybe that is starting to correct itself, I don't know). OGC's process was broken and its product was broken as a result. GeoJSON was simply an agreement by some developers reached in a public sphere. It took a while to get to an agreement, but it was successful because:

1) The scope stayed (mostly) limited.
2) Each participant had goals that were mostly aligned.
3) Each initial participant turned around and produced open source software implementing the specification after its finalization.
4) It is a well written document, with great examples, in an open format behind no ToS
5) It hasn't changed in ~5 years.

OGC has trouble with #1 because it needs to keep producing standards to stay relevant. It struggles with #2 because it serves the organizations that financially support it rather than the organizations that implement its standards. It has historically not cared about #3 even though the open source projects were often the backbone of interoperability purportedly afforded by OGC standards. From what I can tell, OGC is working hard on #4, but this has historically been a big problem. With respect to #5, WMS 1.3 is my canonical response to changing a successful document for the sake of correctness -- WMS 1.3 totally screwed up what was a burgeoning ecosystem of software implementations.

If OGC were to try to fast-track GeoJSON as a recognized standard, wouldn't the same reaction to the fast-tracking of the REST API happen? Wouldn't GeoJSON have to be "aligned" with existing OGC standards (notably coordinate system stuff which has been well-tread in this mailing list)? What does OGC standardization of GeoJSON do other than muck up the ownership of the specification (currently this email list, funny as that is)?

Make a better GeoJSON that corrects its sins but builds on its successes (with a different name please). In my opinion, GeoJSON is successful because of the five points above rather than being a geotechnological wonder of software engineering. I can't speak for others, but I think GeoJSON's standardization story, if there ever is one, is to follow whatever JSON itself does (IETF, yeah?). 

Or do nothing, which has worked out quite well for the past five years :)

Howard