> I have diverted from tetex to texlive,even now,I feel the Latex's
> font is so complicated for me to deploy them from scratch! A shame
> to me!
:-)
> 1. If I use your subfonts.pe to generate the subfonts, can I use
> the non-unicode fonts with Unicode.sfd?
Yes. As an argument you specify an SFD file which gives the mapping
between the encoding of the font and the encoding of the subfonts.
> I like to work under Unicode environment，but the Unicode fonts are
> few and most of them are non-free!
Well, today virtually *all* CJK fonts are encoded in Unicode! Be
careful to distinguish between the input encoding and the covered
character set. For example, gbsn00lp.ttf has a TrueType 3,1 cmap
(which means Unicode), but the character set is only simplified
Chinese -- normally the characters which are contained in the GB2312
character set.
To repeat: There are various terms which you have to understand:
. A character set. This is just a collection of characters. For
example, the list of about 2000 Japanese character all children
have to learn in school (Tooyoo kanji, IIRC).
. An ordered character set. Normally, most character standards
define an order of the contained characters, for example GB 2312.
Traditionally, a character position is given in the range
0x2121-0x7E7E.
. A character encoding. This specifies how a character position in
an ordered character set is really accessed. For example, in
EUC-China encoding, the character 0x2121 is accessed as 0xA1A1.
In Unicode, the same character is accessed as U+6863 (I hope I got
the right mapping value :-)
SFD files which start with letter `U' map fonts encoded in Unicode to
a script specific encoding (Big 5, for example). If you want to
access the fonts with UTF-8 (within your LaTeX document), you should
use Unicode.sfd (which just splits the font without reordering).
A final `complication' in TeXLive is that, for the sake of saving disk
space, only Unicode subfonts do exist as Type 1 fonts; the script
specific versions are realized with virtual font (using the uni2sfd
perl script).
> 2. I have read one article about your nomenclature for cjk
> fonts,such as truetype font bkai00mp.ttf, when we use unicode,
> you say:
>
> (font0) = bkai00mp
> (font1) = bkaiu
> (font2) = bkai
>
> I'm confused about the names of one fonts used by tex itself and
> users can you give me a example for the of unicode, say the
> simkai.ttf?
There is no fixed standard. In former times, font names for TeX
should be restricted to 8 characters (the `Berry' or `fontname'
scheme), but today this has become rather obsolete.
Here an example: bkai00mp.ttf is the real font name. This can't be
directly used in TeX, because (1) TeX doesn't allow digits in font
names, and (2) this is a font with more than 256 characters. You need
subfonts as created by the subfonts.pe script. For convenience, I've
chosen `bkai' as the name stem of the subfonts which expect Big5 as
the input encoding: bkai01.tfm, bkai02.tfm, ... Subfonts which expect
Unicode as the input encoding have an additional `u' in the name:
`bkaiu00.tfm', `bkaiu01.tfm', etc.
> Furthermore, how can I know a font is Unicode or not?
For TrueType fonts, you might try ttf2tfm with an invalid PID/EID
pair:
ttf2tfm <font> -P 9 -E 9
it then displays the available cmaps. If you see the value `(3,1)',
it contains a Unicode cmap.
> 3. When I use the subfonts.pe or other tools, after I generate the
> enc, afm, tfm and pfb files, I found it is so difficult to
> generate the corresponding map files, such as ttfonts.map,
> cid-x.map and etc., and I feel also embarrassed about the
> appropriate place of them. For example, when I use the GBK
> code, I can do all of them properly, but when I plan to use
> Unicode, I find I can't let them work at all!
Regarding font maps, there is the script `updmap' which creates proper
map files for all programs which use Type 1 fonts. The central
configuration file is texmf/web2c/updmap.cfg; simply add a new map
file to the list and execute the script.
ttfonts.map is by default empty in TeXLive -- if you want to use
outline fonts only ttf2pk doesn't help you, obviously (because it
creates PK files).
Since your interest is to make the CJK fonts work with all dvi
drivers, you *must* use subfonts in Type 1 format. Doing that, you
can completely ignore cid-x.map and friends.
> What should I do if I want to work under unicode, so that all of
> the commands such as latex, dvips, ps2pdf, dvipdf, dvipdfm(x),
> pdflaex, can work properly, and I don't want the pk fonts, how can I
> use only type1 or truetype in dvi, ps or pdf files? Can you give me
> a example under the texlive for use the full-unicode environment to
> do the above things?
Have a look at the file doc/latex/CJK/examples/CJKutf8.tex; there you
can see how to use Unicode fonts in a document. Note that you
shouldn't use dvipdf; dvipdfmx is a much better replacement.
Werner
_______________________________________________
Cjk maillist - [email protected]
http://lists.ffii.org/mailman/listinfo/cjk
