[Talk-ca] Canvec.osm sample files are ready

Richard Weait richard at weait.com
Wed Jun 2 05:55:02 BST 2010


On Tue, Jun 1, 2010 at 11:03 PM, Bégin, Daniel
<Daniel.Begin at rncan-nrcan.gc.ca> wrote:
> Bonsoir tout le monde!
>
[ ... ]
>
> I have added Halifax and Peterborough.  I also replace Waterton Park area with a new file that includes addressing.
>
> These are sample, not the final product.  I am waiting on comments from the community to finish the Canvec conversion programs and start conversion process.  I would like to know if the Karlsruhe schema is Ok.  It sometime needs editing around access ramps  buildings. Only local mappers can do that.  Does the distance between adjacent interpolation lines is Ok? - where addresses show up.

Dear Daniel,

I'm having a first look at 031D08_Sample.osm
Please excuse my ignorance of local details.

Overall the presentation of the data looks very good.  In specific the
implementation of Karslruhe addressing schema geometry looks
wonderful.

I see a few things that could be improved:

0) This file marked 031D08 appears to be Halifax.  I was expecting
Peterborough.

1) Junction of highway 111 and highway 118
a) Both are dual carriageways.  Both show both carriageways traveling
in the same direction.  - similar problems appear on other divided
highways.
b) The motorway_links have directional trouble too. See the
motorway_link ways at the north end of the map.
c) carriageways that cross with a bridge often have a two nodes near
the overlap.  These nodes are within ~25cm of each other but are not
co-located. They do not appear to mark a change in tagging.  They do
not appear to be required by geometry.  They are confusing, because
d) some bridges share a node with the road below creating an incorrect
junction.

2) Junction of Highway 111 and Victoria Road
Has 4 identical issues a) through d) as 1) above.
a) carriageways are same direction rather than opposite directions.
b) some ramps are the wrong direction.
c) 3 of four cross overs have double nodes.
d) the fourth cross over is joined by  a node.

3) Inspector Court
a) two buildings are drawn overlapping the road.  It is possible that
either the buildings or the road has changed.  Other layers show
similar problems in other areas.  For example see Drysdale Road
overlapping natural=wood.

4) North West Arm Drive and Osborne Street
a) The divided road shows both carriageways traveling in the same direction.

5) Connaught Avenue - divided residential road
a) both carriageways are drawn in the same direction
b) one carriageway is missing the name=Connaught Avenue
c) junctions are drawn simplified to a single node.  - Can that angle
be changed?
d) Karslruhe address schema ways use interpolation=yes.  I was
expecting to see interpolation=even/odd

6) Junction of  Bedford Highway, Lady Hammond Road and Windsor Street
with unnamed secondary link
a) simplified junction at a single node.  Can Bedford and Lady Hammond
use same narrow angle used by Windsor Street?
b) Dual carriageway missing name on one carriageway
c) dual carriageway of mis-matched road types? Secondary and residential?
d) dual carriageways with same direction on both carriageways.

7) Cannon Terrace - A gated community. Nice touch!
a) the turning circle marked fixme=feature may not exist.  Probably
best to drop the feature if we have no data for it.  Local mappers can
add them where they exist.

Again, overall I think the data looks really good.  Just a few things
to sort out.  Let me know when I should look again.

Best regards,
Richard




More information about the Talk-ca mailing list