The syntax of OpenType

eolson's picture

As most of you are likely aware, the tools and
technology surrounding type design have changed at
an impressive clip in the last few years. The days of
Fontographer and Quark 4 are long gone. As a result,
the types we license have changed as well. OpenType
fonts with rich features are becoming common, though
they're still far from mainstream. With this change we've
become friendly with new programs (FontLab & InDesign)
but only slightly familiar with how to speak of these new fonts.

Is a single file with Small Caps, CE accents, Cyrillic, Greek and
6 styles of numerals really a single font? 4 years ago that might
have been 12 fonts. How about "Pro" or "Basic"? Do we need
language we can all agree on before we slide into the confusions
of SC, LF, OSF, Expert etc?

Just my thoughts on this Friday afternoon

John Hudson's picture

I think the word you want is vocabulary, not syntax, but that aside...

I think the model is that individual fooundries or designers are likely to specify minimum glyph sets that users can expect, and may market these under particular names. There will typically be two considerations: character or language support, and feature or typographic support. The interaction of these considerations may produce modular nomeclature. For example, a character set might be defined that supports particular languages, and some fonts supporting this set might also support a set of typographic features, while other fonts with the same language support may not have any typographic features, or a different set of features. This suggests to me that a clear, if potentially complicated naming system would include one part identifying the language support and another the typographic feature set, and these two parts could be mixed and matched to identify the combined language and feature support of a given font.

eolson's picture

And... after reading through my post I realize why
I don't post often. What an idiot!

You're probably right John. It seems the "potentially complicated"
solution you're referring to may become somewhat of a norm.
I think it's still foggy though.

I suppose what I was trying to say with the original post is
that (unfortunately) many people view OT with a mix of skepticism
and conspiracy theory. Fortunately we're at such an early stage
that it's possible to reverse this. Addressing the often confusing
language surrounding it could clear a few things up and help
type buyers make informed choices.

Thomas Phinney's picture

Hmmm. I agree that many folks are skeptical. As long as they're still on QuarkXPress and Quark doesn't support the extra goodies or Unicode, they're not seeing a whole ton of benefits. Cross-platform fonts are all very well, but only compelling for a few customers.

For those folks, I agree that they have no reason to replace any fonts they already have. However, that is different from whether or not they should buy new fonts in OpenType.

With regards to conspiracy theorists do you mean people who think it's just a scheme to get them to spend more money to replace their existing fonts? There are a few people who have more elaborate conspiracy theories, but they seem like a pretty lunatic fringe.

Sadly, if terms like SC, LF and Expert were confusing to certain people before, then most likely they'll still be confused. But at least they'll have everything in one font, so they just get what's available, and there are fewer decisions for them to make. No more having a customer buy an expert font and then complain that it doesn't have lower case, at least.



Thomas Phinney's picture

With PostScript flavored OpenType fonts, for most non-Adobe apps that are not specifically OpenType aware, the Win/Mac OS (or ATM for older OSes) presents to them the boiled-down class kerning. Basically, it's the class kerning expanded into pairs, and then reduced to just those pairs that for which both characters are in MacRoman/WinANSI.

However, such apps get no kerning at all for these fonts with pairs where either member is outside the basic WinANSI/MacRoman range.

With regards to TrueType flavored OpenType, they can contain both kern table pairs and GPOS kern feature class kerning. However, AFAIK today non-Adobe apps only see the pairs and not the class kerning.

Finally, AFAIK Apple does not yet support any kerning of any sort in Cocoa for either PostScript Type 1 or PostScript flavored OpenType fonts. So the above only applies to Carbon applications. :-(


eolson's picture

Thomas, yes the conspiracy theories in question are related
to those that feal Adobe is peddling OT as another money
making scheme. Hilarious really, because I'm guessing the
type department isn't a major player when it comes to profits
for Adobe (compared to say Photoshop).

Thomas Phinney's picture

You're right.

Sure, Adobe would love to make more money off of fonts, which are a tiny fraction of Adobe's business. But the strategy for doing that is to offer better technology so that people will want to use the new stuff. Users will only make the move if they see value in OpenType, and we're trying to give them that value. Mostly, we expect average users to buy their new fonts in OpenType format, not to replace their existing fonts any time soon.

komitlak's picture

95% of the fonts (cyrillic) i use are created by myself or by my elder dtp fiends. We use custom codepages and encodings wich are tottaly out of any standard (usualy we place the cyrillic letters where the latin ones should to be), this works, but leads us to use sophisticated software in wery primitive way. OT fonts and features (

anonymous's picture


Do most non-Adobe applications like Quark, Word, Freehand etc support class based kerning?

Nigel Hamilton

Syndicate content Syndicate content