[OSM-talk] SteveC; C = Cool

[Matt Amos]

On Wed, Jul 8, 2009 at 12:05 PM, Stefan de Konink<stefan at konink.de> wrote:
> On Wed, 8 Jul 2009, Matt Amos wrote:
>
>> On Tue, Jul 7, 2009 at 10:26 PM, Stefan de Konink<stefan at konink.de> wrote:
>> > SteveC wrote:
>> >> inventing nodes, ways, segments (remember them?)
>> >
>> > You *did not* invent the spaghetti model, please give credit to the
>> > original inventor Stan Aronoff, in Geographic information systems: A
>> > management perspective (1989).
>>
>> actually, OSM doesn't use the spaghetti model. according to [1,2],
>> Aronoff's spaghetti model treats points as coordinates and lines as
>> lists of coordinates
>
> Isn't this exactly how segments and ways are stored within OSM? An XML
> subtree referencing to points (thus lower diminensional objects)?

from those references, it seems that the spaghetti model includes
coordinates directly, rather than referencing a lower dimensional
object by ID. apparently it's called the spaghetti model because each
way is independent, like spaghetti strands (presumably as opposed to
potato waffles or something "joined").

spaghetti model - coordinates are directly included, topology is implicit.
<node lat=y lon=x/>
<way><node lat=y1 lon=x1/><node lat=y2 lon=x2/><node lat=y3 lon=y3/></way>

graph theory model - coordinates are logically referenced, topology is explicit.
<node id=1 lat=y lon=x/>
<way id=1><node ref=1/><node ref=2/><node ref=3/></way>

OSM uses the latter.

>> - basically what the OGC's "simple features
>> architecture" [3] uses - and there's no explicit connectivity. OSM, on
>> the other hand, uses a topological model which comes from a graph
>> theory background, so really we should be crediting Leonhard Euler.
>
> Always good to credit him :)

yep. he was a total genius - invented a whole new branch of
mathematics without which we wouldn't have amazon/netflix
recommendations ;-)

cheers,

matt