Better all-round 3D support

[phayes]

No description provided.

[phayes]

To Do:

1. provide a "flatten" method.
2. Decide how to want to handle 3D geoms for methods such as length, distance etc.

[mkdgs]

i think distance and length is trivial
http://en.wikipedia.org/wiki/Euclidean_distance

and intersect and other a little bit harder ...

postgis has a separate some 3D function (like distance3D)
http://postgis.refractions.net/docs/reference.html#Geometry_Constructors

i don't know how is work in GEOS, but it may be more "backward compatible" to separate 3d function

[phayes]

I think it's definitely a good idea to separate out the 3D stuff into different functions. I checked with GEOS people on IRC, looks like it only support 2D.

[phayes]

The start of the PHP-native distance calculations are in place. "distance is trivial" my ass! :-P

I'm still running into some problems with divide-by-zero errors that I need to investigate.

[phayes]

Now that I've properly implemented distance in 2D, I'm beginning to realize how out of my league I am for implementing the same in 3D.

Problems for 3D distance:

1. line-segments could be closest anywhere along the curve (not just at an endpoints as in 2D).
2. We would need a 3D intersects method, including detecting if two polygons intersect in 3D space.

[phayes]

Hi @mkdgs,

I definitely appreciate where your coming from. The problem is that if we were to implement it this way and removing all these non-universal methods from the parent classes will involve a lot of code duplication in child classes.

I wish there was a way you could have a child-class "undeclare" a method.

I wonder if there is a way we can both be happy by using traits instead of direct inheritance?

In either case, can we not deal with these issues in this pull-request? Let's keep to the existing pattern (with all it warts) for the 3D stuff, and deal with how we do method inheritance in a separate issue?

[mkdgs]

thanks @phayes,

it may be for the future of geoPHP. for now just i have only moved the methods of Point.

take a look into Simple Feature Access at http://www.opengeospatial.org/standards/sfa
there is a lot of good things inside :-)

[phayes]

I'll check that out @mkdgs . Thanks for the link.

[phayes]

I've started testing this and I'm running into all sorts of errors when I run the tests.

1. GeoRSS adapter: "Invalid GeoRSS".
2. WKB adapter: "truncated data"

I'm going to try to correct these errors.

[mkdgs]

hello phayes,
take a look on my repo, i have made many fix/change