2 Unicode versions of the X11 "misc-fixed-*" fonts
3 ------------------------------------------------
5 Markus Kuhn <http://www.cl.cam.ac.uk/~mgk25/> -- 2003-01-17
8 This package contains the X Window System bitmap fonts
10 -Misc-Fixed-*-*-*--*-*-*-*-C-*-ISO10646-1
12 These are Unicode (ISO 10646-1) extensions of the classic ISO 8859-1
13 X11 terminal fonts that are widely used with many X11 applications
14 such as xterm, emacs, etc.
19 None of these fonts covers Unicode completely. Complete coverage
20 simply would not make much sense here. Unicode 3.0 contains over 49000
21 characters, and the large majority of them are Chinese/Japanese/Korean
22 Han ideographs (~28000) and Korean Hangul Syllables (~11000) that
23 cannot adequately be displayed in the small pixel sizes of the fixed
24 fonts. Similarly, Arabic characters are difficult to fit nicely
25 together with European characters into the fixed character cells and
26 X11 lacks the ligature substitution mechanisms required for using
29 Therefore these fonts primarily attempt to cover Unicode subsets that
30 fit together with European scripts. This includes the Latin, Greek,
31 Cyrillic, Armenian, Georgian, and Hebrew scripts, plus a lot of
32 linguistic, technical and mathematical symbols. Some of the fixed
33 fonts now also cover Arabic, Thai, Ethiopian, halfwidth Katakana, and
34 some other non-European scripts.
36 We have defined 3 different target character repertoires (ISO 10646-1
37 subsets) that the various fonts were checked against for minimal
40 TARGET1 616 characters
41 Covers all characters of ISO 8859 part 1-5,7-10,13-16,
42 CEN MES-1, ISO 6937, Microsoft CP1251/CP1252, DEC VT100
43 graphics symbols, and the replacement and default
44 character. It is intended for small bold, italic, and
45 proportional fonts, for which adding block graphics
46 characters would make little sense. This repertoire
47 covers the following ISO 10646-1:2000 collections
48 completely: 1-3, 8, 12.
50 TARGET2 885 characters
51 Adds to TARGET1 the characters of the Adobe/Microsoft
52 Windows Glyph List 4 (WGL4), plus a selected set of
53 mathematical characters (covering most of ISO 31-11
54 high-school level math symbols) and some combining
55 characters. It is intended to be covered by all normal
56 "fixed" fonts and covers all European IBM, Microsoft, and
57 Macintosh character sets. This repertoire covers the
58 following ISO 10646-1:2000 (including Amd 1:2002)
59 collections completely: 1-3, 8, 12, 33, 45.
61 TARGET3 3228 characters
63 Adds to TARGET2 all characters of all European scripts
64 (Latin, Greek, Cyrillic, Armenian, Georgian), all
65 phonetic alphabet symbols, many mathematical symbols
66 (including all those available in LaTeX), all typographic
67 punctuation, all box-drawing characters, control code
68 pictures, graphical shapes and some more that you would
69 expect in a very comprehensive Unicode 3.2 font for
70 European users. It is intended for some of the more
71 useful and more widely used normal "fixed" fonts. This
72 repertoire is a superset of all graphical characters in
73 CEN MES-3A and covers the following ISO 10646-1:2000
74 (including Amd 1:2002) collections completely: 1-12, 27,
75 30-31, 32 (only graphical characters), 33-42, 44-47, 63,
76 65, 70 (only graphical characters).
80 6x13.bdf 8x13.bdf 9x15.bdf 9x18.bdf 10x20.bdf:
82 Complete (TARGET3 reached and checked)
84 5x7.bdf 5x8.bdf 6x9.bdf 6x10.bdf 6x12.bdf 7x13.bdf 7x14.bdf clR6x12.bdf:
86 Complete (TARGET2 reached and checked)
88 6x13B.bdf 7x13B.bdf 7x14B.bdf 8x13B.bdf 9x15B.bdf 9x18B.bdf:
90 Complete (TARGET1 reached and checked)
92 6x13O.bdf 7x13O.bdf 8x13O.bdf
94 Complete (TARGET1 minus Hebrew and block graphics)
96 The supplement package
98 http://www.cl.cam.ac.uk/~mgk25/download/ucs-fonts-asian.tar.gz
100 contains the following additional square fonts with Han characters for
105 Covers TARGET2, JIS X 0208, Hangul, and a few more. This font is
106 primarily intended to provide Japanese full-width Hiragana,
107 Katakana, and Kanji for applications that take the remaining
108 ("halfwidth") characters from 6x13.bdf. The Greek lowercase
109 characters in it are still a bit ugly and will need some work.
113 Covers all JIS X 0208, JIS X 0212, GB 2312-80, KS X 1001:1992,
114 ISO 8859-1,2,3,4,5,7,9,10,15, CP437, CP850 and CP1252 characters,
115 plus a few more, where priority was given to Japanese han style
116 variants. This font should have everything needed to cover the
117 full ISO-2022-JP-2 (RFC 1554) repertoire. This font is primarily
118 intended to provide Japanese full-width Hiragana, Katakana, and
119 Kanji for applications that take the remaining ("halfwidth")
120 characters from 9x18.bdf.
124 Covers the same repertoire as 18x18ja plus full coverage of all
125 Hangul syllables and priority was given to Hanja glyphs in the
126 unified CJK area as they are used for writing Korean.
128 The 9x18 and 6x12 fonts are recommended for use with overstriking
129 combining characters.
131 Bug reports, suggestions for improvement, and especially contributed
132 extensions are very welcome!
137 You install the fonts under Unix roughly like this (details depending
138 on your system of course):
140 System-wide installation (root access required):
145 mv -b *.pcf.gz /usr/lib/X11/fonts/misc/
146 cd /usr/lib/X11/fonts/misc/
150 Alternative: Installation in your private user directory:
154 mkdir -p ~/local/lib/X11/fonts/
155 mv *.pcf.gz ~/local/lib/X11/fonts/
156 cd ~/local/lib/X11/fonts/
158 xset +fp ~/local/lib/X11/fonts (put this last line also in ~/.xinitrc)
160 Now you can have a look at say the 6x13 font with the command
162 xfd -fn '-misc-fixed-medium-r-semicondensed--13-120-75-75-c-60-iso10646-1'
164 If you want to have short names for the Unicode fonts, you can also
165 append the fonts.alias file to that in the directory where you install
166 the fonts, call "mkfontdir" and "xset fp rehash" again, and then you
171 Note: If you use an old version of xfontsel, you might notice that it
172 treats every font that contains characters >0x00ff as a Japanese JIS
173 font and therefore selects inappropriate sample characters for display
174 of ISO 10646-1 fonts. An updated xfontsel version with this bug fixed
175 comes with XFree86 4.0 or newer.
177 If you use the Exceed X server on Microsoft Windows, then you will
178 have to convert the BDF files into Microsoft FON files using the
179 "Compile Fonts" function of Exceed xconfig. See the file exceed.txt
180 for more information.
182 There is one significant efficiency problem that X11R6 has with the
183 sparsely populated ISO10646-1 fonts. X11 transmits and allocates 12
184 bytes with the XFontStruct data structure for the difference between
185 the lowest and the highest code value found in a font, no matter
186 whether the code positions in between are used for characters or not.
187 Even a tiny font that contains only two glyphs at positions 0x0000 and
188 0xfffd causes 12 bytes * 65534 codes = 786 kbytes to be requested and
189 stored by the client. Since all the ISO10646-1 BDF files provided in
190 this package contain characters in the U+00xx (ASCII) and U+ffxx
191 (ligatures, etc.) range, all of them would result in 786 kbyte large
192 XCharStruct arrays in the per_char array of the corresponding
193 XFontStruct (even for CharCell fonts!) when loaded by an X client.
194 Until this problem is fixed by extending the X11 font protocol and
195 implementation, non-CJK ISO10646-1 fonts that lack the (anyway not
196 very interesting) characters above U+31FF seem to be the best
197 compromise. The bdftruncate.pl program in this package can be used to
198 deactivate any glyphs above a threshold code value in BDF files. This
199 way, we get relatively memory-economic ISO10646-1 fonts that cause
200 "only" 150 kbyte large XCharStruct arrays to be allocated. The
201 deactivated glyphs are still present in the BDF files, but with an
202 encoding value of -1 that causes them to be ignored.
204 The ISO10646-1 fonts can not only be used directly by Unicode aware
205 software, they can also be used to create any 8-bit font. The
206 ucs2any.pl Perl script converts a ISO10646-1 BDF font into a BDF font
207 file with some different encoding. For instance the command
209 perl ucs2any.pl 6x13.bdf MAPPINGS/8859-7.TXT ISO8859-7
211 will generate the file 6x13-ISO8859-7.bdf according to the 8859-7.TXT
212 Latin/Greek mapping table, which available from
213 <ftp://ftp.unicode.org/Public/MAPPINGS/>. [The shell script
214 ./map_fonts automatically generates a subdirectory derived-fonts/ with
215 many *.bdf and *.pcf.gz 8-bit versions of all the
216 -misc-fixed-*-iso10646-1 fonts.]
218 When you do a "make" in the submission/ subdirectory as suggested in
219 the installation instructions above, this will generate exactly the
220 set of fonts that have been submitted to the XFree86 project for
221 inclusion into XFree86 4.0. These consists of all the ISO10646-1 fonts
222 processed with "bdftruncate.pl U+3200" plus a selected set of derived
223 8-bit fonts generated with ucs2any.pl.
225 I recommend to play around with the UTF-8 editor Yudit. To use for
226 example the 6x13 font with Yudit 1.5, you just have to select the
237 in the Font menu or in the ~/.yuditrc config file. Yudit is a nice
238 text file editor with UTF-8 support, available from
240 http://www.yudit.org/
241 ftp://sunsite.unc.edu/pub/Linux/apps/editors/X/yudit-1.5.tar.gz
243 You can also use these fonts with Emacs 20.6 or higher. For more
246 http://www.cs.ust.hk/faculty/otfried/Mule/
248 Every font comes with a *.repertoire-utf8 file that lists all the
249 characters in this font.
255 If you want to help me in extending or improving the fonts, or if you
256 want to start your own ISO 10646-1 font project, you will have to edit
257 BDF font files. This is most comfortably done with the xmbdfed font
258 editor (version 4.3 or higher), which is available from
260 ftp://crl.nmsu.edu/CLR/multiling/General/
262 Once you are familiar with xmbdfed, you will notice that it is no
263 problem to design up to 100 nice characters per hour (even more if
264 only placing accents is involved).
266 Information about other X11 font tools and Unicode fonts for X11 in
267 general can be found on
269 http://www.cl.cam.ac.uk/~mgk25/ucs-fonts.html
271 The latest version of this package is available from
273 http://www.cl.cam.ac.uk/~mgk25/download/ucs-fonts.tar.gz
275 If you want to contribute, then get the very latest version of this
276 package, check which glyphs are still missing or inappropriate for
277 your needs, and send me whatever you had the time to add and fix. Just
278 email me the extended BDF-files back, or even better, send me a patch
279 file of what you changed. The best way of preparing a patch file is
281 ./touch_id newfile.bdf
282 diff -d -u -F STARTCHAR oldfile.bdf newfile.bdf >file.diff
284 which ensures that the patch file preserves information about which
285 exact version you worked on and what character each "hunk" changes.
287 I will try to update this packet on a daily basis. By sending me
288 extensions to these fonts, you agree that the resulting improved font
289 files will remain in the public domain for everyone's free use. Always
290 make sure to load the very latest version of the package immediately
291 before your start, and send me your results as soon as you are done,
292 in order to avoid revision overlaps with other contributors.
294 Please try to be careful with the glyphs you generate:
296 - Always look first at existing similar characters in order to
297 preserve a consistent look and feel for the entire font and
298 within the font family. For block graphics characters and geometric
299 symbols, take care of correct alignment.
301 - Read issues.txt, which contains some design hints for certain
304 - All characters of CharCell (C) fonts must strictly fit into
305 the pixel matrix and absolutely no out-of-box ink is allowed.
307 - The character cells will be displayed directly next to each other,
308 without any additional pixels in between. Therefore, always make
309 sure that at least the rightmost pixel column remains white, as
310 otherwise letters will stick together, except of course for
311 characters -- like Arabic or block graphics -- that are supposed to
314 - Place accents as low as possible on the Latin characters.
316 - Try to keep the shape of accents consistent among each other and
317 with the combining characters in the U+03xx range.
319 - Use xmbdfed only to edit the BDF file directly and do not import
320 the font that you want to edit from the X server. Use xmbdfed 4.3
323 - The glyph names should be the Adobe names for Unicode characters
324 <http://partners.adobe.com/asn/developer/typeforum/unicodegn.html>,
325 as xmbdfed can set them automatically if it is configured
326 with the location of the Adobe "glyphlist.txt" file in
327 "adobe_name_file" in "~/.xmbdfed". For xmbdfed 4.5 and older, use
328 <http://partners.adobe.com/asn/developer/type/glyphlist-old.txt>.
330 - Be careful to not change the FONTBOUNDINGBOX box accidentally in
333 You should have a copy of the ISO 10646 standard
335 ISO/IEC 10646-1:2000, Information technology -- Universal
336 Multiple-Octet Coded Character Set (UCS) -- Part 1: Architecture
337 and Basic Multilingual Plane, International Organization for
338 Standardization, Geneva, 2000.
339 http://www.iso.ch/cate/d29819.html
341 and/or the Unicode 3.0 book:
343 The Unicode Consortium: The Unicode Standard, Version 3.0,
344 Reading, MA, Addison-Wesley Developers Press, 2000,
346 http://www.amazon.com/exec/obidos/ASIN/0201616335/mgk25
348 All these fonts are from time to time resubmitted to the XFree86
349 project (they have been in there since XFree86 4.0), X.Org, Sun, and
350 to other X server developers for inclusion into their normal X11
353 Starting with XFree86 4.0, xterm has included UTF-8 support. This
354 version is also available from
356 http://dickey.his.com/xterm/xterm.html
358 Please make the developer of your favourite software aware of the
359 UTF-8 definition in RFC 2279 and of the existence of this font
360 collection. For more information on how to use UTF-8, please check out
362 http://www.cl.cam.ac.uk/~mgk25/unicode.html
363 ftp://ftp.ilog.fr/pub/Users/haible/utf8/Unicode-HOWTO.html
365 where you will also find information on joining the
366 linux-utf8@nl.linux.org mailing list.
368 A number of UTF-8 example text files can be found in the examples/
371 http://www.cl.cam.ac.uk/~mgk25/ucs/examples/
375 Robert Brady <rwb197@ecs.soton.ac.uk> and Birger Langkjer
376 <birger.langkjer@image.dk> contributed thousands of glyphs and made
377 very substantial contributions and improvements on almost all fonts.
378 Constantine Stathopoulos <cstath@irismedia.gr> contributed all the
379 Greek characters. Markus Kuhn <Markus.Kuhn@cl.cam.ac.uk> did most 6x13
380 glyphs and the italic fonts and provided many more glyphs,
381 coordination, and quality assurance for the other fonts. Mark Leisher
382 <mleisher@crl.nmsu.edu> contributed to 6x13 Armenian, Georgian, the
383 first version of Latin Extended Block A and some Cyrillic. Serge V.
384 Vakulenko <vak@crox.net.kiae.su> donated the original Cyrillic glyphs
385 from his 6x13 ISO 8859-5 font. Nozomi Ytow <nozomi@biol.tsukuba.ac.jp>
386 contributed 6x13 halfwidth Katakana. Henning Brunzel
387 <hbrunzel@meta-systems.de> contributed glyphs to 10x20.bdf. Theppitak
388 Karoonboonyanan <thep@linux.thai.net> contributed Thai for 7x13,
389 7x13B, 7x13O, 7x14, 7x14B, 8x13, 8x13B, 8x13O, 9x15, 9x15B, and 10x20.
390 Karl Koehler <koehler@or.uni-bonn.de> contributed Arabic to 9x15,
391 9x15B, and 10x20 and Roozbeh Pournader <roozbeh@sharif.ac.ir> and
392 Behdad Esfahbod revised and extended Arabic in 10x20. Raphael Finkel
393 <raphael@cs.uky.edu> revised Hebrew/Yiddish in 10x20. Jungshik Shin
394 <jshin@pantheon.yale.edu> prepared 18x18ko.bdf. Won-kyu Park
395 <wkpark@chem.skku.ac.kr> prepared the Hangul glyphs used in 12x13ja.
396 Janne V. Kujala <jvk@iki.fi> contributed 4x6. Daniel Yacob
397 <perl@geez.org> revised some Ethiopic glyphs. Ted Zlatanov
398 <tzz@lifelogs.com> did some 7x14. Thanks also to everyone who
399 contributed additions to the UTF-8 example texts and to Bruno Haible
400 <haible@ilog.fr> for valuable comments.
402 The creation of these fonts would certainly not have been possible
403 without Mark Leisher's wonderful xmbdfed software.
408 Markus G. Kuhn, Computer Laboratory, University of Cambridge, England
