On 27/01/11 20:21, 4x4falcon wrote: > On 27/01/11 14:12, Steve Bennett wrote: >> On Thu, Jan 27, 2011 at 3:10 PM, 4x4falcon<[email protected]> wrote: >>> Physical things can change (eg road or railway realignment), non >>> physical >>> don't necessarily change. In the case of boundaries do we definitely >>> know >>> that when a road is realigned does the boundary change with it. This >>> has >>> been discussed many times since the import of the ABS boundary data. >> >> My guess would be that the boundaries generally do change, but I could >> be totally wrong. And I think high quality traces are likely to be >> more accurate than the ABS's data. > > That's the problem it's your guess. There has been no definitive answer > since the original import. > I really don't know why this knotty problem keeps cropping up. I can readily find examples locally for each of these cases:
* ABS2006 boundaries don't follow roads when road alignment changes. * ABS2006 boundaries don't follow rivers when river channels change. * ABS2006 boundaries don't follow coastlines when coastlines change. * ABS2006 boundaries do not always correspond to suburb boundaries. * ABS2006 boundaries do not always correspond to postcode boundaries. * ABS2006 boundaries do not always correspond to shire boundaries. * ABS2006 boundaries do not always correspond to parish boundaries. I couldn't find an example rail line easily; but I think the pattern ought to be clear. (Simple logic says a single exception disproves the rule.) As I've stated elsewhere in this very thread, just because the boundaries *look* like something we want; does not necessarily *mean* they represent that thing. And bending them to fit a fiction just damages what worth they originally had... statistical boundaries for a past census? This is the one immutable fact about them. Their casual correspondence to a feature we would like to map helps locate and identify that feature (even if the drawing happens to be co-linear with the ABS boundary), and they should be applauded for providing those few hints, especially in areas with few other mapped features. However, if somebody has an accurate survey of (to drag the conversation back to it's Subject), say the Victorian Coastline, why aren't they drawing/updating the thing they have surveyed. Who cares if it happens to follow another line on the map? This is not a comparison of like with like. The real problem is: How to explain this clearly and simply to people who don't read/understand/follow this argument? My 2c. _______________________________________________ Talk-au mailing list [email protected] http://lists.openstreetmap.org/listinfo/talk-au
