Re: [Freefont-bugs] Add extra characters to FreeSerif

[Steve White]

Hi Graeme,

The plot thickens!  See below.

On Sat, Feb 26, 2011 at 3:42 AM, Graeme Kidd <address@hidden> wrote:
>
>
> ----------------------------------------
>> Date: Sat, 26 Feb 2011 00:56:37 +0100
>> Subject: Re: [Freefont-bugs] Add extra characters to FreeSerif
>> From: address@hidden
>> To: address@hidden
>> CC: address@hidden
>>
>> Hi Graeme,
>>
>> Some ideas below:
>
> Thanks for the tips but I am on a Windows Vista machine (Frontforge compiled 
> via cygwin), but before I go debugging it further can I just confirm that "ℤ" 
> (0x2124) and "ℝ" (0x211d) do exist. Looking through the FreeSerif glyphs I 
> found two group of blackboard bold letters. The first were in uppercase A 
> (6644 0x19f4 Abb) to G (6649 0x19f9 Gbb), I (6650 0x19fa Ibb) to O (6655 
> 0x19ff Obb) and S (6656 0x1a00 Sbb) to Y (6662 0x1a06 Ybb). The second were 
> in lowercase and went from a (6663 0x1a07 abb) to z (6688 0x1a20 zbb).
>
> Should I not expect to see the R between O and S and the Z after Y? I can 
> also confirm that I am using the 20100919 file.
>
Regarding Vista, it is also possible for different installations of
one font to get tangled up.  I have seen it happen.  If you go to a
DOS prompt and navigate to the system fonts directory, sometimes you
will see ghost versions of previously installed fonts.  I recall there
was no obvious way to clean them out, and I had to do something
sneaky.

Just what are you seeing that appears wrong?  (What are you looking at
in what application program, and what do you expect to see?)

Regarding the curious distribution of the blackboard bold letters: you
are seeing an effect of the uncertain development of Unicode.

The "ℤ" (0x2124) and "ℝ" (0x211d)  in question are in the Unicode range
    Letterlike Symbols
    http://www.unicode.org/charts/PDF/U2100.pdf
The other blackboard bold symbols are from the monster range
    Mathematical Alphanumeric Symbols
    http://www.unicode.org/charts/PDF/U1D400.pdf

Historically, the Letterlike Symbols range appeared first, as a rather
haphazard collection of often-appearing symbols, the  R and Z being
very commonly used in math to represent the algebras of the real
numbers and the integers [German: Zahl], respectively.  Later, they
decided, what the heck, let's add whole ranges of these styled
characters.  But they have some policy not to repeat a letter, so
there are holes in the mathematical blackboard bold range that serve
specifically as placeholders for the previously-defined symbols.

Unfortunately the decision to add whole ranges of styled letters
played conceptual havoc with Unicode users.  Originally, Unicode was
meant to encode symbols of differing *semantic* values, and
specifically *not* differing *styles*.  This latter range seems to
consist of ranges of styles of the Latin alphabet, which have no
particular semantic meaning.  Well OK you could say they can hold
arbitrary but different semantic meaning in math text.  Anyway, the
problem is, I have a font which has a roman, italic, and bold style.
Inside the roman style, this range holds bold and italic versions of
the same face.  Now... what goes in the same place of the bold version
of the font?  A *more* bold style of the bold and italic symbols? What
happens in my bold-italic style?  is the bold-italic range bolder and
italicer?  I guess so.  What a mess.

The answer to your question is: No. In accordance with the Unicode
standard, the blackboard bold R and Z are in Letterlike Symbols, and
there are holes in the Mathematical Alphanumeric Symbols where you
would expect to find those styled letters.

Cheers!