[OSM-talk] 12nm territorial borders - useful or rubbish?

[andrzej zaborowski]

Hi,

On 14 February 2011 03:22, Frederik Ramm <frederik en remote.org> wrote:
>   I've been thinking about the 12nm territorial borders on sea that we have
> in many places, notably in Europe. Many of them seem to have been
> auto-generated by simply placing a buffer around the coastline.
>
> My first question is, do they really have legal significance? They certainly
> give the impression of high precision, hugging every protruding bit of
> coastline in a safe distance.
>
> For example, if I am inside this triangle between Scotland and Ireland, will
> my legal status (concerning, say, fishing quotas, or whom I can marry on
> board of my vessel, or whatever funny things influcenced by international
> borders) be really any different from the status I had if I moved my vessel
> 2 miles in either direction?
>
>  http://www.openstreetmap.org/?lat=55.065&lon=-5.567&zoom=10&layers=M
>
> Or are we, by using these auto-generated (and perhaps not human-reviewed?)
> borders, suggesting a precision that isn't there? Would the UK coastguard
> have a good laugh when I claim to be in international waters at that
> location?
>
> My second question is, assuming that indeed there is significance to the 12
> nm boundary - should such auto-generated data be in OSM at all? If you're
> out on the sea, should whatever navigational aid you carry not compute by
> itself how far you are from the coast, rather than telling you whether
> you're to the left or to the right of a previously computed 12nm line?
>
> And my third question is, assuming that there are really good reasons for
> having these lines in OSM - who takes care of updating them once the
> coastline is modified by a mapper? I think it is a rather unique situation
> to have that kind of data-derived-from-other-OSM-data in OSM itself, and
> thus this has many of the same problems that an import would have (i.e. the
> source data has changed, what now).
>
> I'm not saying we should delete them; but whenever I see them on the map I
> tend to shrug and say "well, seems like someone was trying out his PostGIS
> skillz", and somehow I have the suspicion that the 12nm line as depicted on
> our maps may be little more than "that's what computer geeks do if you tell
> them the border is 12 miles out...".

I don't know how accurate these borders are, but I would prefer that
they stay because closed polygon borders are useful for many types of
spatial tasks.  If you remove those pieces of border, where are you
going to draw the border and why.  You might say tools should generate
the border in their pre-processing but unfortunately the 12 nm rules
is not universal.  It would also rise the entry level effort for tool
authors (for example in a render, should the implementation be part of
the code or part of the stylesheet?  Placement of borders is not
exactly a style issue)

I believe borders are no exception in OSM.  All data is best-effort,
no warranties.  They are improved when better data becomes available
and there are mappers with bandwidth to do the update.

Cheers