Sascha Brawer

Next step: https://github.com/brawer/wikidata-qrank/issues/43

While #43 seems to compute the correct results, the approach taken turned out to be too slow, especially for large wikis such as `enwiki`, `dewiki`, `commonswiki` and `wikidatawiki`. https://github.com/brawer/wikidata-qrank/issues/45

For a Python example, have a look at the data pipeline for [this map of castles in Switzerland](https://castle-map.infs.ch/#46.805,8.205,8z); it uses QRank for choosing which castles get displayed at what zoom...

For use by Wikimedia, [CC0-1.0](https://spdx.org/licenses/CC0-1.0.html) would actually be more useful than [CC-BY-4.0](https://spdx.org/licenses/CC-BY-4.0.html). In particular, [Wikidata Lexemes](https://www.wikidata.org/wiki/Wikidata:Lexicographical_data/Documentation) are licensed under CC0-1.0, not CC-BY-4.0. But that’s a minor detail. @macchiati, what do...

We’re working on it. Let’s keep the bug open until either the license has been changed, or the request has officially been declined.

If you don’t mind, could you send a pull request to fix the problem?

Once fontdiff gets updated to use the most recent HarfBuzz (with AAT support enabled), you could derive two test fonts from the original input font, eg. with fonttools. One font...

There’s a website at https://summerofcode.withgoogle.com/ but I’ve left Google so I wouldn’t know anything beyond that pointer. Good luck, though!

Should we always use hinting, so that the tool stays simple to use?

Cool idea. Send a pull request?