[Tagging] Feature Proposal – RFC – natural=peninsula (Was: Feature Proposal – RFC – place=peninsula)
kevin.b.kenny at gmail.com
Thu Jan 10 21:53:54 UTC 2019
Let me just add a general note to this discussion. I continue to be
interested in studying how to do better label rendering for elongated
features such as certain seas (e.g. the Red Sea), gulfs (the Gulf of
Bothnia or the Gulf of Aqaba), bays (Chesapeake Bay), peninsulas (Cape
Cod), isthmuses, islands (Jura), countries (Norway, Sweden), lakes
(Lake Michigan), channels (Skagerrak), sounds (Long Island Sound),
straits (Queen Charlotte Strait), etc.
The idea would be to examine techniques like those presented in
The figures of the paper (from about figure 10 on) show some of the
sort of results that can be achieved. (It's unfortunate that the scan
quality is so poor that the labels on the shaded areas are almost
impossible to make out.)
Those techniques absolutely require that the features to which they
apply be represented as areas.
I can certainly pilot the project on relatively noncontroversial area
features such as the aforementioned countries, lakes and islands, but
other area features will eventually require tackling both the
technical difficulties in dealing with enormous areas and the
political difficulties in dealing with areas for which part of the
boundary is indefinite.
I surely won't upload any controversial multipolygons at least until a
pilot project is done, and I've no fixed time frame in which I plan to
proceed. (So many projects, so little time!) But I think we can't
afford to forget that high-quality labeling may eventually depend on
resolving these questions.
I know that there will continue to be a healthy controversy, but I
remain confident that if there is a tangible benefit to be achieved
from tagging indefinite objects, eventually the community will accept
it as a necessary evil, and we'll work our way over the technical
More information about the Tagging