[Talk-ca] Internal CanVec conflicts

[Paul Norman]

> From: Bryan Crosby [mailto:azubryan at gmail.com]
> Sent: Saturday, November 10, 2012 8:33 AM
> To: talk-ca at openstreetmap.org
> Subject: Re: [Talk-ca] Internal CanVec conflicts

> This will be my first and only response to this.  If the Canadian OSM
> community feels we need to check every stream, island, oxbow lake, piece
> of 'forest', esker, soil type, muskeg, swamp for accuracy, so be it.
> Keep in mind that in doing so, 80% of the country will receive no data,
> and the community will be down one mapper.

If you don't want to engage in a discussion that is of course up to you.

You appear to be mixing accuracy with consistency. My message was only about
consistency.

Accuracy is about if something is right. Consistency is about if something
makes sense.

Accuracy problems with CanVec are normally limited to old data and the
limits of resolution.

Consistency problems vary more. Generally they're caused because CanVec is
generated from multiple data sets gathered independently without reference
to each other. I had many examples locally of water data indicating a stream
with more recent data indicating the area had been built over. I've fixed
what I've found locally so I don't have any of those examples, but
http://www.openstreetmap.org/?lat=45.695&lon=-73.905&zoom=17 is a good
example from Quebec of the type of result. You don't need to look at Bing to
see that there's something odd there.