Summary of mapping in Alameda, Petaluma, Elfland, and some code in iD.
I saw an OpenStreetMap diary recently “What I did in OpenStreetMap in March 2022”, and Amanda is on the right track. This is such a good idea to post about OpenStreetMap on a regular basis - I know that a monthly book blog has helped me read more often. I don’t know how to borrow this idea without outright stealing. For now I can maybe post a quarterly update?
The map has building footprints, but almost none are labeled. I added restaurants, gas stations, a Whole Foods, and sculptures at the train station.
I walked from the ferry dock to my place, and looking back added some crosswalks (or converting footpath to crosswalk). A school in the neighborhood added a one-way pick-up/drop-off area and rearranged their sports fields to fit the remaining space, so that was a more complex edit.
Elfland, Somerville, MA
A colleague looked up Alameda on Google Maps and we started talking about which businesses and parks show up. She told me about “Elfland”, a collection of tiny buildings which appeared on a vacant lot in late 2021. After looking on Twitter to confirm some details, I removed the old gas station from OSM and added a sculpture marker.
Code in iD
In 2016 I added some code to iD for right-to-left language layout. In 2017 we discussed a notable weirdness to how street labels in many scripts showed up in Chrome. SVG textPath is really obscure so the Webkit bug goes back a long way. I was able to use Unicode presentation forms and JS to improve how Arabic and Hebrew appear in Chrome. Unfortunately other South Asian languages have multiple combining characters which can’t be faked like this.
In October 2021, a fix to the Webkit bug landed in Chromium v96. I made a PR to turn off my changes, and promised to return and clear out my patch 6 months later. Here’s the problem - Safari Webkit still has this issue. A small % of people still use Safari, including every iOS user (other iOS browsers are a wrapper around Webkit). New and non-English-speaking users may be more likely to be on Safari. So I think this code ought to stay in. I’d like to make my alif-toolkit module a bit lighter and a bit better, but that’s what it is for now.