New to Typophile? Accounts are free, and easy to set up.
Kerning pairs
Should the font have kerning pairs for all the possibilities, or is it acceptable to skip pairs that will never exist in words?
In designing the kerning pairs for the Armenian font I came across some pairs (like co-yi or tso-hi) that will never occur in a real word (not even in foreign words), but the pair is drawn in such a way that they will partly overlap if typed by mistake.
Should I create a kerning for such pairs?
User login
Support our Sponsors
16 Sep 2005 — 1:31pm
The paradox is that it can take longer to think about and analyse it than to actually do it! :-) In Armenian especially the non-alphabetic characters aren't many; although the alphabetic ones (not to mention the floating marks) require much heavier kerning than in the Latin alphabetics.
More than kerning though the real problem in Armenian is the detminental effect the requirement of tight spacing can have on the actual forms of the glyphs, especially descenders.
hhp
16 Sep 2005 — 4:38pm
That is one of the main issues I am facing. I decided I will do all the features for this first OTF project and save a reference for future fonts.
As I understand you Hrant, if it is faster to skip it is OK? I am still analysing the kerning and positioning of the marks, which will be a lot of work. Even there, the marks go over a few of the glyphs only, so I will do the features for only those pairs.
Tight spacing is a big issue in Armenian. Most of the fonts available today are not designed in a very good way. They either lack the needed spacing/kerning, or the glyph is designed in within its own bounds and requires minimal or no kerning, which of course is bad font design.
16 Sep 2005 — 6:05pm
I was actually saying that sometimes simply deciding what to skip can take longer than not skipping. I think generally a good compromise strategy is to make some intelligent macro decisions about kerning (for example "I will kern the alphabetics with anything/everything" or "I will not kern between consonants and floating marks") and just do all the rest blindly. It's only when the deadline* is too tight for the kerning job at hand that you have get more clever (or lower your standards :-) for example by using linguistic frequencies, grammatical realities, etc.
* If there is one; for labors of love (like Armenian type design usually is) often there is no deadline... The problem is you then have to define some other "barrier" in order not to still be working on the damn thing on your deathbed! :-)
> Most of the fonts available today are not designed in a very good way.
That's the understatement of the "tar"... :-/
hhp