spec for vector data container - devs only

[Nieko]

In my implementation I only check the sx/sy coordinates to see if it belongs to a chunk... this way there should not be any duplicates... even if there are duplicates - it should not create any major rendering problems...

How would the routing algorithm work in that case? Or: when do you start taking roads from other chunks into account? It might be the case that the start of the road that's best suitable in the routing algorithm lies in another chunk.

I wonder if you would then always have to check the 8 surrounding chunks (assuming roads don't span more than one chunk).

[Codaz]

Nice Deniska,

i was wondering if you coun't lend some of the geocoding of opentom.

www.opentom.org

It's Tomtom, but then written in Linux.

The toolchain is in the wiki.

Maybe then the programming process would speed up.

Nieko i will talk to you in dutch on this.

[deniska]

Nice Deniska,

i was wondering if you coun't lend some of the geocoding of opentom.

www.opentom.org

It's Tomtom, but then written in Linux.

The toolchain is in the wiki.

Maybe then the programming process would speed up.

Nieko i will talk to you in dutch on this.

Although, I did not spend much time reading all the stuff that they have, it seems that most of their code is based on closed-source tomtom api. Which means it cannot be easily ported to any other platform... Correct me if I am wrong..

[deniska]

How would the routing algorithm work in that case? Or: when do you start taking roads from other chunks into account? It might be the case that the start of the road that's best suitable in the routing algorithm lies in another chunk.

I wonder if you would then always have to check the 8 surrounding chunks (assuming roads don't span more than one chunk).

The easiest (perhaps not most practical) solution would be to load all the data chunks that make up a square, spanning over both source and destination points in to a memory and then run one of the famous algorithms on that set..