I want to check an assumption - when you make an Ohorn ( U+01A1, U+01A0 ) it is necessary that the ’horn’ be joined with the O shape so that the O shape in the Ohorn cannot be a component.
(Although the Ohorn shape can be a component of other Unicode glyphs.)
Is this correct?
Put another way though - I might ask. Will a composed overlapping shape create an unwanted white space the way that a temporary ( unconnected but overlapping) leg on an upper case R does?
Thanks!
What happens can be rasterizer-dependent and in some cases size-dependent. But also consider what will happen if somebody ever does an outline or stroke effect on the text. You really don’t want overlapping paths in that case.
Regards,
T
Great Point. Thanks Thomas!
...if somebody ever does an outline or stroke effect on the text. You really don’t want overlapping paths in that case.
That’s why I weld (remove overlap) cedillas and ogoneks onto accented characters. Shipping a font with diacritics in component form only saves a few kilobytes in font file size and adds nothing to the quality of the product.
Most fonts I’ve produced so far are display types, or semi-display-demi-text, and designers will inevitably render them in outline form. As long as your horn glyph, cedilla glyph and ogonek glyph exist stand-alone in their own cells with legit unidoce index numbers assigned, they can be components of diacritic characters combined on-the-fly by applications—-but you knew that already.
Will a composed overlapping shape create an unwanted white space the way that a temporary ( unconnected but overlapping) leg on an upper case R does?
The whitespace is generated in OT fonts with Type 1 outlines, but for some reason not in TT fonts with TT outlines.
j a m e s
Captured from web.archive.org on 4 Feb 2008. view this capture