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

[Howard Butler]

On Nov 11, 2013, at 1:00 PM, Raj Singh <raj at rajsingh.org> wrote:

> In general I agree that, "if it ain't broke, don't fix it". Changing a specification after years of implementation can do more harm than good, even when the proposed changes are well-meaning. 

That's comforting, but it wasn't Adrian's posture when was talking about alignment with OGC and the SRS changes required to do so. OGC standardization talk from that point was bound to get our heckles up given the history of this group and the fact that the horse is well out of the barn as far as industry-wide implementation was concerned.

> 
> Formal standardization within OGC can sometimes be of benefit when there's a concern about intellectual property since part of OGC's role is to  defend its work against patent trolls, mainly by identifying prior art. However, it's highly doubtful someone would claim GeoJSON violates any patent since it is so clearly an expression of the simple features model.

GeoJSON is probably a reaction to OGC kind of abandoning SF in favor of GML and imposing a ton of complexity on implementors in the process. GeoJSON is indeed quite clearly SF's geometry model + JSON objects.

> 
> And yes I agree the SRS part has warts, but what SRS encoding doesn't?
> 
> I do wish that the web site had better (or easier to find?) examples that were richer than basic geometry encodings. 

Might it be useful/helpful for us to move the website to github pages so that improvements like this could be easily incorporated? The website's current status predates easy availability of static website publishing platforms and is a bit of a time capsule in that regard.

> 
> Putting my money where my mouth is, I spent the weekend developing GeoJSON output for my OpenPOIs service. Here's an example. Feedback and testing is very welcome. I've only tried it with Geolint so far.
> 
> http://openpois.net/pois/a75da8b9-5e85-46f2-9546-35c87ef0dda3.geojson

Neato.