Should ₹ be part of the Latin font subset? (2020) | Hacker News Reader

Should ₹ be part of the Latin font subset? (2020)

shkspr.mobi

10 pointsrobin_reala6 comments

> Roboto's Latin subset contains Old English characters like þ (Thorn) and Ð (Eth). Are obsolete characters used only in mediaeval text really more important to include than the currency for a billion people?

Aren't these characters used in today's standard Icelandic?

I might be misreading this but is this article advocating for moving the character in the Unicode codespace? If there's room in Unicode, then sure. If there's not, it seems to me that rearranging Unicode for societal changes like this can only lead to headaches when code points map to different characters depending on what version of Unicode the font uses (and what version of Unicode the website expects). But I'm not an expert on this sort of thing so I might be understanding it wrong.

Although I do understand that Unicode makes more sense for Americans like me, and if I were Indian I might feel different.

No, that’s not the proposal. The problem that Google has is that if it ships every possible Unicode character in a font, then that font will weigh many megabytes and the vast majority of that will not be used. So they do “font chunking”: they split the font into smaller files that contain a subset of the characters, and serve it to users. If you look at Google’s stylesheet for Roboto you can see that the font is split into 7 chunks:

https://fonts.googleapis.com/css2?family=Roboto&display=swap

Each chunk has a unicode-range hint associated with it, which tells the browser that it doesn’t need to download the font unless those Unicode characters are found in the markup it’s processing.

The article’s assertion was that ₹ is a a common enough character among English speakers / writers that Google should move it into the base Latin chunk instead of putting it into the extended Latin chunk, which would save a bunch of font downloads.

This is all a balancing act though. Every character you move into the “default” chunk makes that chunk (and the unicode-range definition) larger, so you have to be careful and work with data. And one imagines that Google has some pretty good data in this area.

Why not include all the currency characters by default then?

Yes, if it's regularly used in Latin character text.

Reply on news.ycombinator.com