New typeface character extension

Andrea T.

Hi everyone!
Lately I've been trying to figure out how foundries or individual type designers decide which charset extension makes sense to develop during the design process for a first release.
Let me explain better: when you're working on a typeface for the retail market, there’s no guarantee of a financial return, so every hour you put in matters. If I go for a charset that's too limited, I risk narrowing down the potential audience for my typeface. But if I develop too much, I might just be wasting valuable time.
In your experience, what’s a good balance between these two? Do you usually set a base charset when launching a new typeface? Do you go by Unicode blocks or focus on language coverage?
Of course, it probably also depends on the typology of the design, a more experimental display font might not need much, while something more functional and versatile would probably require broader coverage.

I’d be happy to hear your thoughts

PS: I shared it in Type Business because I feel it’s mostly an economic matter. Though I might be wrong 

Andreas Stötzner

You are right in that a text face is likely to go better with a well-supplemented character set, whereas a playful display font may be sufficiently equipped with a more slim repertoire. It’s always a new game with every font. But also a display font may deserve the attention for, e.g., providing a full character set for (nearly) every Euro-Latin orthography.

Above all, it depends on the field of usage scenarios a given font is intended to cater for.

Thomas Phinney

For Latin language fonts (and probably all fonts), do not go too much by Unicode blocks. They are not a very useful thing in conceptualizing your character set. (And I say this as somebody who is a fan of Unicode, and has done entire presentations on Unicode!) Rather, do it by languages or groups of languages. Things like punctuation marks, currencies, and various other important symbols have their own separate Unicode blocks, so there are lots of Unicode blocks you will likely only cover portions of.

There are existing character set standards that might help you think it through. But for a first-ever release, of a Latin typeface, I might go for something like Adobe Latin 2 (250 glyphs) or if more ambitious Adobe Latin 3 (331 glyphs). (Yes, there is an Adobe Latin 1, but it is a legacy character set, with 229 characters.) Basically the difference between those two character sets is, the first covers western European languages (all of WinANSI and MacRoman), the second adds all the major central/eastern European (Latin) languages. Maybe drop a few of the less important symbols like the litre symbol.

Other folks have their own versions of these, but groupings quite similar to these two are quite common.

Mads Wildgaard

Some very good advice here already, as to be expected. My personal take is; extend latin languages fully— there's no excuse for a macron combination not working in some specific name. It's so easy to make. 

Thomas is very right that it's a good idea to take a pragmatic look at what unicode blocks consist of and then take a step back and think if needed or not. Personally I've found that primes (double and single) has often been requested by clients, so I try to always include them now. Just take a look at general visual language around you and see what it consists of.