What may an OpenType font know about alternative glyphs?

Discussions about the development of TrueType and OpenType fonts.
Post Reply
Top Typographer
Top Typographer
Posts: 2025
Joined: Tue Sep 14, 2004 6:41 pm
Location: Worcestershire, England

What may an OpenType font know about alternative glyphs?

Post by William »

In the viewtopic.php?t=2214 thread is the following question.
Does anyone know if it would be possible for the program to integrate multiple images per character, and then randomly select one live while one types? I'd like to make a font that looks like handwriting, without repeat images for the same characters in close proximity.
In that thread Bhikkhu Pesala writes as follows.
It would not be possible AFAIK to make random glyphs appear.
I got to wondering exactly what would be needed to make the required effect happen. Maybe it is something for the future for products such as Adobe InDesign or Serif PagePlus.

I know that InDesign, which I have not used, I have only read about it, can be used to access alternative glyphs for a character from an OpenType font, a chosen alternative glyph rather than a random alternative glyph, so presumably (I do not like using the word presumably but my knowledge of this matter is limited and this is a discussion) the font has knowledge that a particular glyph is an alternative version of a letter e or of a letter g or of whatever. I am thinking of this situation contrasting with the TrueType possibility of including in the font an alternative version of a letter e or of a letter g or of whatever within the Private Use Area, where the glyph is available yet the font has no knowledge that, say, an alternative version of a letter e is in fact an alternative version of a letter e rather than being some other character.

So, as InDesign can obtain information from a font about which alternative versions of e letter are available, maybe an application could use such information in order to include one of them at random, as the original enquiry asks. However, it occurs to me that some human discretion is still needed, unless the OpenType rules for labelling of alternative glyphs is more comprehensive than I presently know. For example, please consider the letter e. Suppose that a font has three glyphs for letter e, namely the regular glyph and two alternative glyphs. One of the alternative glyphs might, say, have a slightly different angle than the regular e. That glyph could perhaps be used randomly with no problems. The other of the two alternative glyphs might have a long sweeping tail, useful for only the end of a word or maybe for only for the end of a line of poetry. Use of that glyph randomly might cause a poor result as it might in such a case be used within a word and look wrong.

So, I am wondering, how much knowledge about an alternative glyph can an OpenType font contain? Would extending that knowledge perhaps be useful?

As someone who has a long-standing interest in random numbers (as in random walk simulation of conductive heat transfer in arbitrary-shaped objects) I am thinking as to quite how a random use of alternative glyphs would work. Maybe there would need to be a facility to enter a seed value into the software package using the font so that, if desired, the display effect produced one day could be repeated on another day, rather than the display effect necessarily being different every time the text file were opened.

William Overington

10 April 2008
Post Reply