Saudi Arabia highway shields

[1ec5]

Saudi Arabia has a three-tiered system of highway shields. It will be an interesting design challenge to make the shield icons legible and spacious enough for the route number.

One national highway has been one mapped so far, tagged network=SA:national. (There’s also a redundant partial route tagged network=SA:highway.) No city or branch roads have been mapped as relations yet.

So far, routes have only been tagged with the route number in Western European numerals, even though the signs only bear Eastern Arabic numerals.

[ZeLonewolf]

I'm inclined to think that the use of western numerals in the ref tag is a tagging error when the signage on the ground is using eastern numerals.

[michaelblyons]

I'm inclined to think that the use of western numerals in the ref tag is a tagging error when the signage on the ground is using eastern numerals.

I concur. If both are present and either is sufficient to mark the sign (like Iran, it sounds like?) I'm fine with Western numbers. But in the examples above from Saudi Arabia, I imagine it's more helpful to mark the text that appears on the sign. Even if my brain reads the tertiary highway above as "gamma gamma gamma gamma,"* it'll be more helpful than having to look up which digits correspond to the ones on the sign.

* Chamelion

[1ec5]

465 would be a good place to explore whether it’s more important to match the digits on the sign or use digits that are readable to users. (There could even be a user preference about that, just like many maps have about local language versus user language.)

But as for the tagging, I agree that ref might as well be in the local writing system. Apart from use cases like #21, there isn’t a strong need for the ref to be parseable as a numeric value in general. It does raise the question of which key to use for the Western Arabic numerals, should mappers want to indicate that somehow. int_ref? ref:mul-Latn? ref:posix? 🙂

If the refs are retagged to use Eastern Arabic numerals, they’ll probably disappear from this style initially, because the font we’re using doesn’t have a very large character repertoire: #393.

[1ec5]

If the refs are retagged to use Eastern Arabic numerals, they’ll probably disappear from this style initially, because the font we’re using doesn’t have a very large character repertoire: #393.

584 and #590 fixed this issue, so if we were to implement this shield today and later on the relations were retagged with Eastern Arabic numerals, the map would automatically reflect the change once the vector tiles are updated.