Jones Magloire

I also did some quick perf (Node v8.16.0) testing and it found that : - Memory print : `Set` uses less memory than `Object` (for localities Set = 59.9MB and...

I tried your branch, it use much more memory (415MB) and seems to be slower than Set (4 times slower than Set) :thinking: Here is the test : https://gist.github.com/Joxit/32ccd5b5f63b474f30e707d804cbda25 Maybe...

Is this normalizer necessary? The API is already doing the lowercase transformation for `layer` and `source` (`category` too ?) and there are also some check for ids :thinking: The code...

Ha ha ha, it's ok, 4 years ago, the statute of limitations has passed :p

:+1: we should not use `ids` for concordance. A feature like this would be very interesting, especially with the OSM data :+1:

I think Travis agrees :smile:

In [OSM Wiki](https://wiki.openstreetmap.org/wiki/Key:addr:place) they say: > Use addr:place=* instead of addr:street=* for buildings whose number belongs not to a street, but to some other object. It is okay to have...

Unfortunately this is a data issue from Who's on First. Pelias uses polygons from WOF in order to get the full administrative hierarchy. We faced the same issue at @jawg...

I agree! So you think we should only update the schema (two apostrophes are already in the list)? And the label will still use the Right Single Quotation Mark?

When I check pelias compare, the data uses old data and the common apostrophe https://pelias.github.io/compare/#/v1/autocomplete?text=1122+Emmy%27s+Canyon+Lane I have some other examples in France such as `40 Rue De l'Arsenal` which moved...