Aldermann Type

Pro edition

Character Set.

672 drawings across ten blocks, the six OpenType features that ship, and the six things that go wrong between a font file and a rendered page — none of which are the type designer’s fault and all of which are the type designer’s problem.

By block

Basic Latin is 95 of 672 and it is a fifth of the work
BlockGlyphsWhat is in it
Basic Latin95ASCII. Every font has it and it is 20% of the work.
Latin-1 Supplement96Western European accents. The point at which a font becomes usable outside English.
Latin Extended-A128Central and Eastern European. Polish ł, Czech háček, Hungarian double acute, Turkish dotless ı.
Latin Extended-B (part)42Romanian comma-below among others. Partial because the block is enormous and most of it is unused.
Vietnamese134Stacked diacritics. Bellrope Sans only, and the single largest decision in the family.
Punctuation and marks61Including both dashes, guillemets, and the case-sensitive alternates.
Figures, four sets40Lining, oldstyle, tabular, proportional. Ten drawings each and they are not scalable copies of one another.
Currency18Including ₺, ₴ and ₹, which date a font faster than anything else in it.
Maths and misc34Real minus, real multiply, and fractions drawn rather than assembled.
Ligatures and alternates24fi, fl, ffi, ffl and the case-sensitive forms.

Features, and what they silently change

Two are on by default and one of those is wrong in a code editor
FeatureStateWhat it does
ligaBy defaultStandard ligatures. Off in a code editor, which is why Quire Mono ships with them disabled in the font rather than expecting the app to do it.
loclBy defaultLanguage-aware forms. Turkish i keeps its dot in caps, Romanian gets the comma below. Silently wrong in every application that does not set a language.
tnumOn requestTabular figures. Turn it on in any table of numbers and the columns line up; leave it off and they do not, which is the single commonest typographic bug in web applications.
onumOn requestOldstyle figures. Right in running text, wrong in a table, and the choice is the writer’s rather than the designer’s.
caseOn requestCase-sensitive forms. Raises hyphens, parentheses and guillemets beside capitals. Nobody turns it on and everybody should.
ss01On requestA single-storey a in Bellrope Sans. Exists because an interface client asked; kept because it is genuinely better at 11 px.

Six things that go wrong after delivery

Every one of these gets reported as a bug in the typeface, and none of them is one.

  1. The subset drops the accentsA build tool subsets to “latin” and Polish disappears. The font is fine; the pipeline threw the glyphs away. Check a built page, not the source file.
  2. No lang attributeWithout lang="ro" the locl feature never fires and Romanian silently gets a cedilla. One attribute, and it is missing on most multilingual sites.
  3. The fallback stack is wrongA missing glyph falls back to whatever is next in the stack, which is a different design at a different size. It looks like a bug in the font.
  4. Variable axes clampedA weight of 450 requested from a static 400/700 pair renders as 400. Nothing errors and the page is subtly wrong everywhere.
  5. Faux bold and faux italicThe browser synthesises them when the weight or style is not loaded. It is always visible and it is always blamed on the typeface.
  6. Hinting on WindowsThe one platform difference that still matters at small sizes. Test on it, or find out from a client on it.