[talk-au] Suburb boundaries - getting close

Darrin Smith beldin at beldin.org
Mon Feb 16 19:39:40 GMT 2009

On Mon, 16 Feb 2009 22:09:15 +1100
Franc Carter <franc.carter at gmail.com> wrote:

> Ok, it seems my conversion script is now producing sane results so
> it's time to work out what the final output should look like.
> The first question that I think we need to answer is, how do we
> represent the
> data in OSM, there appears to be 3 options:-
>    1. Closed ways
>    2. Relations
>    3. Borders with a left/right tag

My vote is for #2, and I'd be strongly against the use of #3 since it's
essentially the system #2 set out to replace and is so dependant on way
direction and making adjoining suburbs all match directions vs
left/right will be painful. #1 is a fine choice in city regions but I
think it will cause ways to be too large in country regions, it also
prevents someone telling which suburbs a boundary way lies in.

> Then we need to decide on what tags to apply to the data. The raw
> data has three fields
>   * STATE_2006     A numerical identifier for the state the suburb is
> in
>   * SSC_2006        An identifier provided by the ABS
>   * NAME_2006      The name of the suburb, which may have the old
> name in '()' after it.
> So, my initial proposal for tags is:-
>   * name=?
>                                                         (with any old
> name removed)
>   * source=Based_on_Australian_Bureau_of_Statistics _data     (ABS
> ask for this)
>   * ABS:reviewed=no
>   * ABS:STATE_2006=?
>   * ABS:NAME_2006=?
>   * ABS:SSC_2006=?
> The 'ABS' part is just a suggestion - It's a bit short for my liking

My thought: Make it  au:ABS:...  that way it flags it as an Australian
thing, and within Australia I don't think there's too many multiple
uses of 'ABS' in this context :)

> We also need to decide where these tags go - nodes, ways, relations.
> And if we go for the left/right approach a decision on how to

I think how far 'down' the tagging goes depends on how we want to
handle the update every 4 years. 

- If we plan to do a point by point check each time then we probably
need to tag each node with a unique ID number to detect changes.

- If we plan to do more of a diffing of the 2 data sets and updating
changes only then we can probably get away with just tagging the data
to the ways.

I think the 2nd option is going to work better for us in the long run
(given how much adjusting the boundaries are looking to need anyway).

Of course if we choose option #2 above then I think both ways and
relations will need to be tagged, although the ways will only
need the source= tag and the unique ID #.



More information about the Talk-au mailing list