Page 1 of 1

Composite vs. Simple

Posted: Thu Dec 17, 2015 7:37 am
by PJMiller
Does anyone have a feel for the relative sizes of composite glyphs as opposed to simple glyphs? Obviously it will depend on the number of control points but in general I would imagine that composites are smaller, but I think there are advantages to making a glyph simple.

Is there a case for making all glyphs simple?

If a glyph is composite and has a scale factor applied or is mirrored does this cause any problems?

I realise that mirroring it once will cause the contour to be the wrong direction but mirroring twice ( i.e. a turned glyph ) will cause the contour to be the correct direction.

What about scaling the glyph? Does this cause any problems?

Does anyone have any opinions or advice?

:?:

Re: Composite vs. Simple

Posted: Thu Dec 17, 2015 9:13 am
by Bhikkhu Pesala
  1. Using composites makes your fonts smaller because they are just references to the composite glyph members
  2. Editing fonts that use composites is more efficient and less error-prone as you only need to edit the composite glyph members. However, changes to the metrics may need copying to the composites
  3. There is no benefit in making all glyphs simple as far as I know
  4. Scaling, flipping, and rotating composite glyph members may cause problems, so I steer clear of it.
  5. Scaling glyphs for superscripts, subscripts, etc., is bad design. Scaled glyphs need to be made heavier to match the stroke weights of other glyphs in the font. Compare, for example: ™ and TM, or 1 and ¹ in Times New Roman.
Glyph Scaling.png
Glyph Scaling.png (4.59 KiB) Viewed 9084 times

Re: Composite vs. Simple

Posted: Thu Dec 17, 2015 6:45 pm
by PJMiller
Thankyou Bhikkhu for your concise and informative answer.

You have confirmed most of what I had assumed.

My only comment on point 4. would be that composites which have been reflected twice (i.e. turned) seem to work well in Microsoft Word 2007, Libre Office and in Inkscape, of course they might fail horribly in all other software, I don't know.