[XeTeX] on fontspec 1.10

Jonathan Kew jonathan_kew at sil.org
Mon Jun 26 13:54:32 CEST 2006


On 26 Jun 2006, at 12:34 pm, Will Robertson wrote:

> Hi,
>
> I suppose I should have forseen this complaint.
> Before I address it, I would like to ask if anyone minds that
> fontspec features are accessed with mixed case keyval syntax. It's
> come up previously that all other LaTeX packages would use
> [letters=smallcaps], so fontspec's a little in the wrong, but I've
> been a bit loathe to change it and make things even more confusing
> (in the short term).

My 2¢: PleaseStayWithMixedCaseForTheOptions;  
itssomucheasiertoreadwhenthingsgetlong. :)

but maybe the latex style gurus have other views....

> Anyway...
>
> Ralf Stubner wrote:
>> I have run into a similar problem with a font I am currently
>> developing,
>> which (like GFS Didot) has small caps only for the regular shape. I
>> tried to avoid this by changing
>>
>> \newfontinstance\spacedsc[LetterSpace=5.0,%
>>                           Letters=SmallCaps,%
>>                           Letters=UppercaseSmallCaps]{FPL Neu}
>
> No, this one is better than calling \addfontfeature every time, since
> the option processing occurs only once (the actual font definitions
> will only occur the first time, though). It's a negligible
> difference, but...
>
> Now, back to the problem. Since [Letters=SmallCaps] is only defined
> for the upright shape, to avoid warnings for the other shapes you
> should activate this feature for the upright shape only:
....

I wonder whether it would make sense for fontspec to provide a  
version of \newfontinstance that only defines the specific font you  
ask for, rather than attempting to find and set up the four  
'standard' style variants.

In fact, \newfontinstance sounds to me like a good name for such a  
thing, and the current \newfontinstance could be named  
\newfontfamily. But maybe it's too late to reconsider the name of an  
existing command, and instead you could have \newfontface to define a  
single face rather than a family?

JK




More information about the XeTeX mailing list