[talk-au] Aligning admin boundaries (including those resulting from ABS 2006 imported data) to coastline..

[Ian Sergeant]

On 20 January 2011 18:22, <Pheasant.Coucal at gmail.com> wrote:

> Victorian Coastline notwithstanding guys, please remember the data under
> discussion is the ABS2006 set. It seems to have largely followed
> physical boundaries; but not quite always. It seems to have followed
> suburb boundaries; but not quite always. See the pattern yet? I have
> found cases of the boundaries following roads which closed prior to the
> 1920s. And rivers which since have flooded and cut across old loops in
> the 1970s. And will be wildly wrong next year after the recent floods.

Can we can just confine the discussion to coastline then?  As you say,
there is unlikely to be a definitive answer for other boundaries, but
the coast is the coast, yes?

Steve Bennett proposed the five options..

1) Administrative boundaries are as imported, and will randomly
criss-cross the coastline (current situation)
2) Admin boundaries are on the same way as the coastline
(natural=coastline;boundary=administrative;admin_level=8...)
3) Admin boundaries are colinear with the coastline, but on a
different way (natural=coastline, than a separate, colinear way with
boundary=administrative;admin_level=8...)
4) Admin boundaries are parallel with the coastline, inside it. (How far?)
5) Admin boundaries are parallel with the coastline, outside it. (How far?)

> Renders be buggered. They can/should be
> fixed, our aim is data capture not pretty-printing, YES?

So you would discount options 4 and 5 on that basis?

Would you then support option 3?

> Special note to Ian Sergeant: Stop arguing with John Smith. You will not
> outlast him. You are making a fool of yourself. You are starting to look
> like an immature jerk. At least keep a minimum of dignity.

Apologies.  I tried to focus the discussion, but obviously failed to do so.

One of my pet-hates in OSM is the way the admin boundaries currently
mis-align to the coastline.  I often realign them, but it would be
nice to have a consistent and documented approach.  That really is the
extent of my interest here.

Ian.