[XeTeX] OpenType Math kerning (cut-ins)

[Diederick C. Niehorster]

Dear John,

I have personally no interest in this feature right noww, but can
definately see its use for many people, as you have correctly
sketched. This way, the XeTeX community could slowly build its own
patches for fonts that are broken, making XeTeX more and more usable.
I see you wrote this has come up before. Has anybody made concrete
suggestions as to a syntax for the .krn file or even for the
implementation? That might really help in getting this going.

Just what i thought of when i read the above..
Dee

On Fri, Oct 30, 2009 at 6:44 PM, John Was <john.was at ntlworld.com> wrote:
> Dear Adam
>
> That's interesting, and I think reinforces the case for introducing a
> 'custom kern' option into XeTex at the point where one loads a font.
> Sometimes the unacceptable spacing will be a fault of the font provider
> (e.g. collisions between some letters and closing parentheses in Monotype
> Baskerville, or overtight apostrophe kerning in Minion Pro), sometimes, as
> apparently here, the result of a failure of the typesetting program
> correctly to read the information in the font.  But either way, in real life
> one just wants to achieve a nice-looking final PDF, and sometimes rather
> quickly.  The XeTeX gurus are usually very quick to fix reported bugs, while
> one could wait for ever for Monotype or Adobe to bring out an improved
> version of the font, so in practice what one wants is a way of fixing the
> flaw immediately, pending a more permanent solution.  So, just as one can
> say e.g.
>
> :letterspace=1.5 or :mapping=tex-text
>
> when invoking a font, I would like to have the option of, say,
>
> :kern=d:/kerns/mykern.krn
>
> where the file mykern.krn is a user-generated UTF8 file listing kerns (in a
> specified simple syntax) that should be added to those that the program has
> detected (whether or not with complete success) in the font file itself.
> (The entries  would of course overwrite kerning pairs already loaded with
> the font but leave in force all the loaded kerning pairs that the user
> doesn't want to interfere with.)
>
> My recollection from previous discussions is that this was considered quite
> practicable but I guess not high priority - though when the subject comes up
> there are usually several people indicating that they would appreciate the
> facility.  If an improved version of the font or of the XeTeX engine
> eliminated the flaws, one would then simply remove the ':kern=' option from
> the font call.
>
>
> John
>
>
>
>
>
>
>
>
> :letterspace=1.5
>
>
> ----- Original Message ----- From: "Adam Twardoch (List)"
> <list.adam at twardoch.com>
> To: "XeTeX List" <xetex at tug.org>
> Sent: Friday, October 30, 2009 9:35 AM
> Subject: Re: [XeTeX] OpenType Math kerning (cut-ins)
>
>
>> Dear list members,
>>
>> I'm forwarding the reply from Ross Mills, one of the principal designers
>> of Cambria Math. The attachment indicates that the kerning of the said
>> pair is correct in Word 2007.
>>
>> Best,
>> Adam
>>
>> -------- Original Message --------
>> Subject: Re: [Fwd: [XeTeX] OpenType Math kerning (cut-ins)]
>> Date: Fri, 30 Oct 2009 02:01:22 -0700
>> From: Ross Mills <ross at tiro.ca>
>>
>> Adam Twardoch wrote:
>>>
>>> -------- Original Message --------
>>> Subject: [XeTeX] OpenType Math kerning (cut-ins)
>>> Date: Wed, 28 Oct 2009 11:43:39 -0400
>>> From: Joel C. Salomon <joelcsalomon at gmail.com>
>>> Reply-To: Unicode-based TeX for Mac OS X and other platforms
>>> <xetex at tug.org>
>>> To: X�TeX <xetex at tug.org>
>>>
>>> Consider the attached files (requires a recent version of
>>> unicode-math).  There does not seem to be any kerning of f_k, even in
>>> Cambria Math where it's most obviously needed.
>>>
>>> I recall from Rhatigan's TUG 2008 presentation "Three typefaces for
>>> mathematics" (based on his thesis; see
>>> <http://ultrasparky.org/school/pdf/DanielRhatigan_Dissertation.pdf>,
>>> pp. 36�37) that OpenType Math fonts have somewhat complicated kerning
>>> (cut-ins) for super-/sub-scripts.  Is support for these implemented in
>>> X�TeX?
>>>
>>> �Joel Salomon
>>>
>> Apparently not. The subscript positioning may also be off, judging by
>> this example. This would tend to indicate lack of support or incorrect
>> interpretation of Math table values and constants (cut-ins/kerning and
>> subscript positioning, respectively) within XeTex. I don't know how
>> complicated the kerning can be categorized as, but the values are stored
>> in the Math table, and its just a matter of implementing those
>> correctly, but yes, it is more sophisticated then 'normal' kerning. This
>> may also account for the bad spacing in the quad integral example you
>> sent before, but the kerning for integrals is handled in a different
>> manner, and is likely to cause confusion for implementations.
>> Positioning of scripts off of the integral are indicated by italic
>> correction values, rather then by cut-ins. Don't ask me why. Sergey
>> Malkin may have a better memory then I on that issue.
>>
>> -Ross
>>
>>
>> --
>>
>> Adam Twardoch
>> | Language Typography Unicode Fonts OpenType
>> | twardoch.com | silesian.com | fontlab.net
>>
>> Reporter: "So what will your trip to Ireland look like?"
>> Lech Wałęsa: "I get into a car, then onto a plane, and then the other
>> way around."
>>
>>
>
>