[Tagging] Rio de la Plata edit war

Adam Franco adamfranco at gmail.com
Tue Aug 4 14:17:13 UTC 2020

It seems to me that the main underlying conflict is that (at least in the
default Carto rendering on openstreetmap.org a few years ago) the Rio Plata
was getting rendered as land at low-zooms and South America simply looks
wrong when such a large water area is rendered as land.

https://github.com/gravitystorm/openstreetmap-carto/issues/1604 is a very
long issue thread that addresses this land/water rendering problem,
starting first with some of the Great Lakes not rendering, then expanding
to cover other water areas. I'm not familiar enough with the Carto code to
know the current state of things, but would this rendering issue still be
present today if the coastline was drawn from Punta del Este to Punta Rasa
as muralito suggests? If this rendering issue has been addressed then it
feels to me like the stakes of the "coastline" tagging placement become
much lower.

My impression is that low-zoom renderers seem to often interpret
"coastline" as "the junction between land and 'big water'" whereas people
standing on land looking at water might often interpret "coastline" as
solely "the junction between land and ocean/sea" and may not want to
include rivers, estuaries, canals, navigation channels, salt-marshes,
protected bays, and other features that sit a little bit away from where
ocean waves are crashing on the beach. While usually these features are
small enough to ignore at low-zoom, Rio Plata, the Saint Lawrence estuary
<https://www.openstreetmap.org/node/368389866#map=6/48.254/-66.270>, Long
Island Sound,
<https://www.openstreetmap.org/node/368389866#map=8/40.393/-72.103> Pamlico
Sound <https://www.openstreetmap.org/relation/11190230#map=6/32.510/-72.751>,
and others are big enough to be visible at low-zooms qualify as "big water"
connected to the ocean even if they aren't fully ocean themselves.
Similarly does the "coast" shift out to include barrier islands and enclose
the protected waters behind them? Looking at current mapping it does
sometimes and doesn't other times.

All that said, there are probably many other data consumers who aren't able
to leverage the techniques used by Carto and would like to make their own
decisions about how to render the distinctions between land, fresh water,
and big salty water. Being able to tag an estuarine environment with its
own tags could allow data consumers to make their own choices about
rendering and placement. Some potential use cases:

   - A general purpose renderer like Carto would probably want to display
   estuaries as water to make the land shape more closely match low-zoom
   satellite imagery. Most traditional general-purpose maps care more about
   the distinction between land and water rather than what the properties of
   the water are

   - A marine ecology map might wish to render oceans and estuaries, but
   hide rivers and land-based features from display.

   - A map focused on rivers might want to highlight the distinction
   between purely fresh-water rivers and estuaries in a more pronounced way
   than a general-purpose map would want to.

Joseph previously suggested dedicated estuary tagging:

On Tue, 4 Aug 2020 at 06:43, Joseph Eisenberg <joseph.eisenberg at gmail.com>

> I have previously proposed that estuaries should be mapped by extending
> the coastline upstream to the limit of the estuary, and also mapping the
> area of the estuary as water with water=estuary

Unfortunately, only a waterfall or dam is going to provide the kind of
hard-lined transition point between ocean and river that would allow a
"coastline" way to cut across the water without ambiguity as to its
placement. All other transitions points on every river-mouth (estuary or
otherwise) are going to be ever changing with rainfall and tides and at
best will only ever be an agreed-upon average or culturally defined
placement. It is simply impossible to verifiably map an ever-shifting
gradient with a single line across the water.

I'd suggest that estuaries get their own tagging as areas (maybe with
sub-sections covering differing properties like tidal influence, salinity
thresholds, etc) and let the "coastline" cross it wherever local mappers
agree a good average threshold is met. The estuary areas should have their
extent based on physically defined thresholds like "average annual salinity
greater than x%" or "average current dominated by ocean-flow vs
river-flow".  Data consumers can then be free to use the mapped "coastline"
line or ignore it where it crosses an estuary, including the greatest
extent of the estuary in their interpretation of "coast" or the most
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/tagging/attachments/20200804/78260c86/attachment.htm>

More information about the Tagging mailing list