BROKKEN

Creating a crossover between a broken brutalist and an ‘esoteric’/fantasy font. You can find the current state (untested) as a usable Brokken.otf (OS X) and Brokken.ttf (WIN) in the fonts folder. If you want to use the font in a semi-tested version please use one of the stable releases. Brokken is developed in Glyphs 3 and doesn't use any Bézier curves at all. More info about its design and features can be found here.

Status Quo

• The font is in an early beta state. Use at your own risk. All central European letters (including all diacritics) are supported. A lot of testing is still neccessary.
• The bold, semibold, regular, light, and thin versions can be considered usable, an italic version is in development (currently only untested builds can be found in fonts)
• The font features two masters (Brokken bold and Brokken thin) where the other versions (Brokken light, Brokken regular and Brokken semibold) are interpolated from.
• Currently there are 387 glyphs in each master, that feature a total of 359 (master bold) + 567 (master thin) kerning pairs (excluding group kerning).

Known Issues & Open Questions

• I can't test if the Dutch ij/IJ ligature works properly since I don't know the language. UPDATE: Thanks to Henrike I understood the problem a lot better: The key issue here is that a íj́ is difficult to input on a standard keyboard, therefore it is neccessary that once a í following a j is detected the j needs to be replaced by j́. This is also true for captial letters. (see here). Although I think I followed all recommendations from here there is still an error report once I run fontbakery check-googlefonts.

  Language          FAIL messages
  ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
  nl_Latn (Dutch)   Shaper didn't attach acutecomb to uni0237
  ^                 Shaper didn't attach acutecomb to J
  

  This error does not occur when the font is exported by the Glyphsapp export function. Although a total different error appears:

  FAIL The following glyphs have components which themselves are component glyphs: * uni01F4 * Gcircumflex * uni004A0301 * Jcircumflex * Ubreve * Aacute.ss01 * Abreve.ss01 * Acircumflex.ss01 * Adieresis.ss01 * Agrave.ss01 and 29 more.

  I have a subtle suspicion that these errors appear due to my somewhat convoluted setup: the component characters (such as acutecomb) consist of components on its own. Maybe there is a mismatch with this configuration. After testing I can confirm this -- when converting the components within the acutecomb glyph into paths the font happily passes the Dutch language check in fontbakery. It is strange that this error does not occur with any other comb-characters. UPDATE: Issue seems to be resolved.

• Spacing and kerning is rudimentary at best, but it is in constant development.

• Certain group kerning features apparently do not work under Windows – I can't figure out why. (Example string: können)

• The font is primarily tested on OS X, therefore I don't know if the automatic hinting works 'good enough' for all letter combinations on Windows or Linux. I mitigated some issues by adding more sidebearing (+5 on each side) for each glyph in the thin-master.

• Critical: The horizontal scaling of the letters/glyphs needs to be adjusted, the font is too tall at the moment UPDATE: Did some rescaling, metrics still need to be adjusted since the descenders are now cut off. UPDATE 2: Metrics seem to be OK for now (tested with OS X and Windows 10). TODO: Metrics need to be adjusted to comply with this.

• I can't estimate if the shapes of the Icelandic letters (such as Þ, þ or ð) are recognizable for native speakers/readers since I don't know the language

• the number glyphs need some rework/finetuning

• all punctuations need some testing

• the ogonek diacritic instance uses currently the bottom-anchor point, not the ogonek-anchor point

• the lcaron, dcaron, Lcaron and Dcaron are composed from the standard caroncomb-component, but need their own caroncomb.alt-component instead, otherwise the shape of the characters are wrong/off. See here.

• missing characters for Catalan (therefore failing the GF_Latin_Core-test by fontbakery)

Useful Commands

• The builds are created using fontmake (fontmake -i -g ../sources/brokken.glyphs -a) which needs to be run inside the fonts directory. (The argument -i is given in order to interpolate between the masters, otherwise only the thin and bold masters are exported to ttf/otf. The argument -a is given to apply basic autohinting to the ttf version.) See here.

• activate Python3 virtual environment: source myenv/bin/activate

• run QA-tool fontbakery with Google Fonts template: fontbakery check-googlefonts Brokken-Bold.ttf --full-lists (reports can be found in the reports directory).

• building the *.ttf font files with fontmake and applied autohinting will result in an error message from fontbakery:

  FAIL This is a hinted font, so it must have bit 3 set on the flags of the head table, so that PPEM values will becrounded into an integer value.

  As far as I can see this flag can not be set inside Glyphsapp (see here and here). Therefore it seems necessary that this issue needs to be adressed by running gftools fix-hinting on each generated ttf. Useful bash-scripts can be found here. To streamline this process I condensed the whole building and testing process into one bash script:

  #!/bin/bash
  
  # initialize python3 virtual environment
  source myenv/bin/activate
  
  # clean the export folder
  echo "deleting all contents of fonts folder"
  rm -rf $HOME/Documents/GitHub/brokken/fonts/*
  
  # build the font using fontmake
  echo "building font with fontmake"
  cd $HOME/Documents/GitHub/brokken/fonts
  fontmake -i -g ../sources/brokken.glyphs -a
  
  # correct ttf version with gftools
  echo "gftools"
  cd $HOME/Documents/GitHub/brokken/fonts/autohinted/instance_ttf
  for font in *.ttf
  do
      gftools fix-hinting $font;
      if [ -f "$font.fix" ]; then mv "$font.fix" $font; fi
  done
  
  # check font with fontbakery
  echo "checking font with fontbakery"
  fontbakery check-googlefonts Brokken-Bold.ttf --full-lists
  
  # cleanup
  echo "moving autohinted contents to default instance_ttf folder"
  mkdir -p $HOME/Documents/GitHub/brokken/fonts/instance_ttf
  cp -r $HOME/Documents/GitHub/brokken/fonts/autohinted/instance_ttf/*.ttf $HOME/Documents/GitHub/brokken/fonts/instance_ttf
  rm -rf $HOME/Documents/GitHub/brokken/fonts/autohinted
  
  echo "renaming instance_ttf and instance_otf to ttf and otf"
  mv $HOME/Documents/GitHub/brokken/fonts/instance_ttf $HOME/Documents/GitHub/brokken/fonts/ttf 
  mv $HOME/Documents/GitHub/brokken/fonts/instance_otf $HOME/Documents/GitHub/brokken/fonts/otf 
  
  # make a copy of the otf files in the document directory for testing in InDesign
  echo "copying the contents of otf folder to the InDesign test folder"
  cp -r $HOME/Documents/GitHub/brokken/fonts/otf/*.otf $HOME/Documents/GitHub/brokken/test/"brokken-test Ordner"/"Document fonts"
  

Design/Approach

The font attempts a hybrid between a brutalist aesthetic, at times reminiscent of a Textura, and a design that evokes the aesthetics of 1980s pen-and-paper role-playing games and fantasy novels. As a ‘creative constraint’, the font does not use any Bézier curves. The name ‘Brokken’ is derived from its rock-like appearance (German ‘Brocken’ = chunk, bolder), the approach that all letters are created by combining chunks, and that the overall design is strongly influenced by ‘broken’ typography (such as Textura, Element, etc.).

Note: Screenshot shows an old version of the capital R.