As part of T323113, we've decided that map data from Commons has sufficiently different characteristics that we should not include it in the first phase of expanding map data, and will discuss whether it makes sense to expand at all.
Informational points:
- One use case for Commons map data is to reuse shapes and points across pages, and even across wikis. This is similar to how templates are used.
- The other use case is to provide translations for labels and popup text, via the JsonConfig mechanism. This use is similar to i18n messages.
- Our investigation into average geoshape payload did not cover Commons map data, it was not included.
What we want to know:
- How prevalent is Commons map data in maps?
- What is the average additional payload which would be added if maps are expanded?
- What is the average latency of the call to expand map data?
With this, we can estimate the impact of expanding Commons map data at parse-time.