Projects Google Fonts ought to fund in 2023?

I'm the font development programme manager, and I'm putting together a 2023 plan for programmes to fund - typeface projects, code/tool projects to support font development, event sponsorships, etc.

I want to make a more public call for project proposals generally, so starting with a thread here first - although I guess most people will want to get in touch with my privately.

Type Drawers private message, dm on Twitter, email to fonts@google.com, etc all go to the same inbox :)
«13

Comments

  • Thanks for posting this, David. Are there any specific focus areas for Google to pursue with priority?
  • Thanks for posting this, David. Are there any specific focus areas for Google to pursue with priority?
    I agree with Andreas. This post is very short in details. Could you be more specific, please?
  • As long as we’re talking about Google money:

    Buy the entire Berthold library from its current “owner,” pay back royalties to its surviving designers, and make it open source.
  • Aiming a bit lower:

    Add (or pay to have added) IPA glyphs to Source Serif Pro, Alegreya, Bodoni Moda, Vollkorn, Cormorant, EB Garamond, Amiri, Spectral, Tiro Indic, Neuton... basically all the modulated-weight serif Latin designs.

    It’s hard to tell whether some of them already contain IPA glyphs without cracking them open in FontLab—for example I could swear Gentium has IPA, but you can’t tell from Google Fonts’ glyph dump default view.
  • @John Butler it should not require using a font editor of any sort, if you merely want to know whether some unicode range have any/some glyphs or not. I don't think the IPA range is part of font validator's os/2 tests (which shows coverages of some unicode ranges if some unicode bits are set in the os/2 table, i.e. the font claiming to support those languages), but current ftdump (part of freetype's demo tools) has a -c and -C options, the former shows summary of unicode range coverage, the latter actual mappings between unicode points with glyphs. You can look for, or programmatically extract the IPA range from either -c or -C outputs. Some tips involving using ttx/fonttools to extract cmap should work too.
  • John HudsonJohn Hudson Posts: 2,955
    @John Butler

    Add (or pay to have added) IPA glyphs to ... Tiro Indic
    Best target would be Castoro, the spin-off of the Latin subset of the Tiro Indic fonts, which has already been extended to more European language coverage than the Indic.

    I’d be happy to provide Google with a project proposal and pricing for a fully implemented phonetic notation extension to Castoro (minimum IPA, but also options for Americanist and Uralicist).


  • I'd love to see more fonts with optical sizing; there are plenty of overly stylized and plenty of plain workhorse, but something in the middle would be great. Full IPA support would be awesome, as well.
  • KP MawhoodKP Mawhood Posts: 294
    Increase multilingual support for mega corp branding:

    For instance:
    – Support for global markets that harmonises with popular types for branding
    – Educational resources on how to "font pair" across writing systems/scripts
    – Publicise as Noto alternatives for branding applications
  • What about funding for Hind to have true italics? o:)
  • Here’s an idea to spread the Google wealth: create an “Exquisite Corpse” font project. Set up a font template and hire 75 type designers to each design a single, assigned glyph. No designer may see what the other designers are doing. If the resulting font becomes a hit, you’ve made a work of art! If the font stinks, hire another 75 designers and start again.

  • John HudsonJohn Hudson Posts: 2,955
     “Exquisite Corpse” font project
    Or ‘Chinese Whispers’ font project: each designer can only see the glyph that the previous designer has created. So someone designs A who shows it to someone who designs B who shows that to someone who designs C, etc.
  • James PuckettJames Puckett Posts: 1,969
    Get everybody to design one glyph based on TypeCooker results.
  • Dave CrosslandDave Crossland Posts: 1,389
    edited August 2022
    Katy, John, Michael, Cristóbal - thanks for the suggestions, that's the kind of thing I am soliciting. Yves, I am intentionally not prescribing any particular direction as I'm curious what suggestions people might have, and don't want to foreclose anyone's ideas. I'm hoping to hear ideas I wouldn't have had myself :D

    Hin-Tak, maybe we can finally sort out the remaining edges of FontVal and wrap it up nicely. Missing italics, missing latin script language support in popular Latin fonts, more fonts, these are all clear needs of the GF library to me :) 

    It’s hard to tell whether some of them already contain IPA glyphs without cracking them open in FontLab—for example I could swear Gentium has IPA, but you can’t tell from Google Fonts’ glyph dump default view.
    Right, for like a decade the subset definitions didn't include a lot of stuff that was in a few mega-glyph-set families, so they weren't possible to use via the web API, but with the recent addition of all Noto fonts in fonts.google.com/noto then subsets for everything in Noto were developed - and now we can add API serving support for them. 

    The issue to solve this is here, https://github.com/googlefonts/gftools/issues/602

     “Exquisite Corpse” font project
    Or ‘Chinese Whispers’ font project: each designer can only see the glyph that the previous designer has created. So someone designs A who shows it to someone who designs B who shows that to someone who designs C, etc.
    That's more or less the basis of my collaborative typeface design sketching workshop:

    You have a class of say 30 people draw 30 sketches with pencil on A4 paper in say 10 minutes, then bring them all to the front/hallway and lay them all down, and then the instructor (or by a more democratic method) picks a 'survivor' and then ahem selection pressure is applied to kill off the others. The survivor is snapped from a phone and projected on the classroom projector, and then they all have another 10 minutes to draw a new generation based on the survivor's DNA. Next iteration, they add more glyphs. Next iteration, instead of higher mutation due to redrawing by eye, don't project to the front, but photocopy the survivor for every participant to trace from and extend the existing design with additional glyphs, possibly upgrading to A3 size paper or using the scaling feature of the photocopier. Loop the process.

    I did this at a few of the Understanding Fonts and then Crafting Type workshops, and I initially proposed it to the OSP.kitchen folks in Brussels in 2007 or so.

    I then developed this process in response to the putrid smell that arises from direct collaboration on typeface designs, some decade ago when I was directing development of such real-time collab features in FontForge - 

    https://www.youtube.com/watch?v=rMR7H8IWlTI

    https://www.youtube.com/watch?v=BqwTim6gMHA

    https://www.youtube.com/watch?v=xZY08SnS9oo

    - but FontForge was a creaky foundation and the feature bitrot quickly. Today the most relevant implementation is Simon Cozens' GlyphsApp CritParty Plugin (repo on Github):

    https://www.youtube.com/watch?v=_Vx2f_Du1LI

    Those 2 experiences led to the discussions about 5 years with Jeremie Hornus at Black[Foundry] about CJK font development, that led to RoboCJK. RoboCJK AFAIK has never been presented at any conference, at least not one with a video online, but its recent successor, Fontra, was presented this year at the NYC Typographics TypeLab:

    https://www.youtube.com/watch?v=Xhw5iCTK1XI

    In RoboCJK there is no real-time collab, rather it is "turn based" - like an 80s/90s RPG - where you can "lock and check out" some glyph set and then check in your improvements, possibly subject to view. This makes it more straightforward to organize the "evolutionary peer production" method of the workshop, I think.

    Hin-Tak Leung said:
    @John Butler it should not require using a font editor of any sort, if you merely want to know whether some unicode range have any/some glyphs or not.

    FontVal is cool... but an instant checker like the "local font" feature of Rosetta Type's Hyperglot or Roel Nieskens' WakamaiFondue are nice for a really quick inspection of a font's language, character and OpenType Feature/Axes support. 

    I designed PyFontaine to be a "swiss army knife" for rapidly developing unicode character set checker based profiles, but the problem with that, hyperglot, CLDR, and most other language checkers is that they go off unicodes only. 

    In response to that critique, Simon Cozens recently developed ShaperGlot (repo on Github) which does OpenType feature code analysis too, so that for example you can tell if a Latin font containing unicodes for Dutch or Turkish actually supports those languages with the required features, which are auto-generated by editors like GlyphsApp.
  • Christian ThalmannChristian Thalmann Posts: 1,938
    edited August 2022
    I second John's IPA motion; I've suggested that in the past as well... Cormorant probably doesn't need it since IPA is usually not used in display settings. It would be a great fit for Ysabeau, though!
    (Of course, that's only sensible if you ever figure out how to publish Ysabeau. :wink:)
  • I have some ideas regarding fontval (mostly with what is known as "FontVal-RX" or the RX branch: https://github.com/FontVal-extras/FontVal-RX ) , its relationship with VTT and FreeType version 3 . Will write separately /privately.

    Regarding VTT etc... completely wild suggestion: pay @Georg Seifert enough money to open-source glyphapp :smile:
  • When my new next-door neighbor found out I did fonts, she told me that she uses IPA fonts and would be thrilled if Google offered more fonts with IPA support.

    Specifically, she uses them in annotating choral sheet music. Of typefaces with IPA support currently available, I was going to say Montserrat is her favorite… but in fact she is getting fallback and not realizing it!   :#
  • As long as we’re talking about Google money:

    Buy the entire Berthold library from its current “owner,” pay back royalties to its surviving designers, and make it open source.
    Monotype announced this today.
    https://www.monotype.com/company/news/monotype-acquires-bertholds-renowned-typeface-inventory


  • that is a BIG BANG. I’m inclined to record this development under ‘good news’, but I’m not entirely sure about this.

  • Dave CrosslandDave Crossland Posts: 1,389
    edited August 2022

  • They saw this thread and decided to get ahead :-))

  • Bertholds' crazy "you can't use this font for anything" EULA is still up!
  • Of course.

    Just because the business deal has closed does not mean they are ready to pivot on such things. It may well be months—perhaps MANY months—before Monotype changes the EULA or brings them under the Monotype EULA.

    Perhaps the wackiness of the Berthold EULA will make this a higher priority, but don’t expect it to be in place tomorrow.
  • I would like to see a good serif font "for everyday", namely a font like Times New Roman and Century Old style. Also wish there was a centerline adjusted version of it, which would be IMO better for source code etc. 
  • John HudsonJohn Hudson Posts: 2,955
    I would like to see a good serif font "for everyday", namely a font like Times New Roman

    STIX Two Text: based on 10pt metal Times.
  • John SavardJohn Savard Posts: 1,088
    edited August 2022
    DAMN THEY MOVE QUICK BRO

    This tactic may prove counterproductive. If Monotype becomes for type what Hasbro is for boardgames, politicians may decide that giving strong legal protection to type designs is against the public interest.

    STIX Two Text: based on 10pt metal Times.

    I remember installing the original STIX fonts on my computer, and having to remove them, because there was some atrocious bounding box issue that meant that, at least in programs like WordPad, the font simply didn't work properly. I am glad to see this is corrected.

  • John HudsonJohn Hudson Posts: 2,955
    The old STIX PS Type 1 fonts and their direct derivatives had lots of problems. STIX Two is a completely different typeface: there’s barely any DNA of the old STIX left in it.
  • Also Crimson Pro, Roboto Serif, EB Garamond, etc. 

    What is "centerline version"?

  • What is "centerline version"?
    Sorry I should have elaborated on this. So usually fonts are "baseline", i.e. capitals and digits are aligned along its baseline. "Centerline" font I call a font which has its capitals and digits  aligned along its centerline (midline of lowercase text).
    In my experience it is more adequate solution for a medium with a lot of "mixed" 
    information like source code etc. Not that I have seen a lot of such fonts but maybe 
    someone makes such version of Times.

       Also Crimson Pro, Roboto Serif, EB Garamond, etc. 

    These are good of course.
    Still Roboto IMO not even on the same readability level as Times. Crimson is better. Garamond still a very good font but huge ascenders and some characters are a bit distracting  to be honest. 
     
    For work I choose what reads best, and in all my life I haven't seen better option than Times and Century Old Style based fonts.
    And on Windows particularly Times New Roman reads especially good, that is due to some "secret sauce" I beleive that makes it behave well in Windows apps. And yes I think there is room for improvements  so I d like to see more work around these excellent fonts.

  • @Mikhail Vasilev you might also be interested in the Libertinus fonts. They don’t have a center-line cut (as far as I can tell) but there is a monospaced version which might fulfill the same role.
Sign In or Register to comment.