Further work on ancillary files.

* webfont.txt: Use Unicode, rewrite character table and Greek
transliteration sections.
* pronunc.txt: Update.
* tagset.txt: Update.

1 files changed, 529 insertions, 517 deletions

diff --git a/webfont.txt b/webfont.txt
index 591e980..d432fe5 100644
--- a/webfont.txt
+++ b/webfont.txt
@@ -1,88 +1,70 @@
                  WEBSTER FONTS
                  =============
 
-          Fonts for the Webster 1913 Dictionary.
-          For version 0.50
-          Last edit May 5, 2001
-          ______________________________________
- (This file contains some extended ASCII characters, and should be
-transmitted in binary mode)
----------------------------------------------------------------------- 
-
-    This file describes a modified font for use in visualizing the
-text of the 1913 "Webster's Revised Unabridged Dictionary" (W1913),
-usable for the DOS operating system of IBM-compatible personal computers.
-The electronic version of that dictionary and this font were prepared by
-MICRA, Inc., Plainfield NJ, and are copyrighted (C) 1996 by MICRA, Inc.
-For details of permissions and restrictions on using these files, see
-the accompanying file "readme.web".
-    The special characters used in the electronic version of the Webster
+* Overview
+
+This file describes special symbols and markup entities used in the
+GNU Collaborative International Dictionary of English.
+
+* Introduction
+
+The special characters used in the electronic version of the Webster
 1913 are required for visualizing unusual characters used in the
 etymology and pronunciation fields of the dictionary, in a form
-comparable to the way they appear in the original.  Since there are
-more than 256 characters used in that dictionary, not all can be
-represented by single-byte codes, and are instead represented by
-SGML-style "short-form" symbols.  (rather than the "entity" format
-"&xx;"  The ampersand is used frequently, and we prefer to leave
-the "<" as the only "escape" character) of the type <x/ where x
-is a specific code for the symbol in the dictionary.
-See the "Short Form" section below for details about such characters.
-Note that the symbols used here are in some cases abbreviations
-(for compactness) of the ISO 8879 recommended symbols.  If necessary,
-the table below allows simple replacement by alternate encodings.
-    This symbol font can be loaded in IBM-compatible (x86) computers
-running the DOS operating system by using the "font.bat" command file
-in the "utils" directory.  The fonts files for 8x14 and 8x16 fonts are
-"web14.fnt" and "web16.fnt" respectively.
-    For those loading the Webster onto some machine other than an
-IBM-compatible running DOS, it will be necessary to provide a
-translation table, to convert these characters into a code that
-can be handled by that computer.  For this reason, I attach an
-"explanation" for each character, for those who cannot view
-the original DOS font.
-    The DOS-loadable font does not contain all of the characters needed
-to depict the etymologies or the pronunciations.  In addition to an
-absence of several characters used in the pronunciations, no Greek letters are
-included.  The Greek words appearing in the etymologies,
-when they are included, will be typed in a
-roman-letter transcription (See section on Greek transcription, below).
-Only a very few Greek words have been thus transcribed as of the
-present version (version 0.41).
-    Wherever the typists did not know the character to use, they
-usually inserted a reverse-video question mark (decimal 176).
-This appears in full-ASCII versions as <?/.  This mark was used both for
-characters in non-ASCII fonts, and for unreadable characters (i.e.,
-characters smeared in the original or distorted in the copies available
-to the typists. The type in the original was in many places smeared and
+comparable to the way they appear in the original.
+
+The GCIDE markup provides two ways for representing such characters:
+using special "escape sequences" and using special markup entities.
+Historically, "escape sequences" were used to indicate the
+character's ordinal position in a special font, prepared by MICRA,
+Inc. to represent it on screen.  Although nowadays this method is
+obsolete, the dictionary corpus still uses these sequences.  This file
+describes their mapping to Unicode characters.
+
+An escape sequence has the form \'xx, where "x" represent lowercase
+hexadecimal digits.  For example, \'94 stands for "o" with diaeresis.
+There are only 256 such sequences.
+
+Special markup entities are able to represent a wider range of
+characters.  A markup entity is similar to SGML one, but has a
+different format.  The traditional &xx; format was judged inconvenient
+because the ampersand is used frequently in the corpus.  Instead,
+GCIDE entities have the format <WORD/, where "<" and "/" represent the
+beginning and end of the entity and WORD represents the character
+itself.  Valid WORDs are in some cases abbreviations (for compactness)
+of the ISO 8879 recommended symbols.  Characters representable by
+escape sequences can also be represented by entities, but the reverse
+is not true, due to a limited range of the former.
+
+The Greek words appearing in the etymologies, when they are included,
+are typed in a roman-letter transcription, which is described below in
+chapter "Greek transliteration".
+
+* Unrecognized characters
+
+Wherever the typists did not know the character to use, they usually
+inserted a reverse-video question mark (decimal 176).  This appears in
+full-ASCII versions as <?/.  This mark was used both for characters in
+non-ASCII fonts, and for unreadable characters (i.e., characters
+smeared in the original or distorted in the copies available to the
+typists. The type in the original was in many places smeared and
 illegible at the left and right page margins; occasionally, small
 parts of words were blotted out by plain white space).
-    A character table for the high-order characters appears below.
-Under that is a list and description of most of the special characters
-used in the Webster files.
-     Note that there are yet some characters used in the etymologies,
-and some other symbols, which are not in this list.  For example, the
-vowels with a double dot *underneath*, e.g. a (as in all) have no representation
-in this character set, and, where explicitly entered in the dictionary,
-are represented by <xdd/ where "x" is the letter, as in "<add/".
-
-ITALICS
--------
-   In most places, italic font is represented by the tags <it>...</it>
+
+* Italics
+
+In most places, italic font is represented by the tags <it>...</it>
 surrounding the italic text, or by some other tag which also implies
-italic font.  In the pronunciations, however, where italicized vowels 
+italic font.  In the pronunciations, however, where italicized vowels
 are used among non-italic and other special characters to indicate
-pronunciation, the special codes <ait/, <eit/, <iit/, <oit/, <uit/, 
+pronunciation, the special codes <ait/, <eit/, <iit/, <oit/, <uit/,
 are also used to indicate the italicized vowel.
 
-DIACRITICS
--------------
-     The European grave and acute accents are represented by the
-standard (IBM PC) high-order codes.  Other characters with diacritics
-are represented by special "entity" codes, and in some cases also
-are found in this special WEB1913 font, described below.
-     Vowels with a circle above (as in Swedish) are coded <xring/
-(x with a ring, or "degrees" mark over it); vowels with tilde over them
-are represented by <xtil/, where "x" is the vowel, as in <etil/ (<atil/
+* Diacritics
+
+Vowels with a circle above (as in Swedish) are coded <xring/ (x with a
+ring, or "degrees" mark over it); vowels with tilde over them are
+represented by <xtil/, where "x" is the vowel, as in <etil/ (<atil/ 
 also has code 238); letters with a dot above are represented by <xdot/
 -- letter with a dot below are represented by <xsdot/ ("subdot");
 vowels with the semi-long mark (a macron with a short perpendicular
@@ -93,70 +75,57 @@ the "oo" with an unbroken macron above the two letters, <aemac/ = the
 ligature ae with a macron [also 214 = \'d6], and <oemac/ the ligature
 oe with a macron [also 215 = \'d7]); vowels with umlauts or a crescent
 (breve) above have codes in this list, but may also be represented by
-<xum/ and <xcr/ respectively.  There is an occasional hacek or caron mark
-(an inverted circumflex) in the original; such letters are coded <xcar/.
-The o with a caron has code 213, but no others are in this font list.
+<xum/ and <xcr/ respectively.  There is an occasional hacek or caron
+mark (an inverted circumflex) in the original; such letters are coded
+<xcar/.  The o with a caron has code 213, but no other letter with a
+caron is representable by an escape sequence.
+
 The diaeresis is treated typographically as identical to the umlaut.
-   A special modification, used only for poetry (see entry "saturnian verse"
-under "saturnian") is a vowel with a macron, in which the macron is lighter
-than the usual macron, signifying a stressed syllable which has a short
-vowel sound.  This is represented by <xsmac/ ("short mac").
-   Another special character used in pronunciations is an "n" with an underline (like
-a macron, but below the letter), used to represent the "ng" sound.  This is coded
-<nsm/ ("n sub-macron").  The ligated th used in pronunciations to depict the
-"th" sound of "the" is coded as <th/.
-    NOTE: the letter combinations "fi" and "fl" are invariably printed as the
+A special modification, used only for poetry (see entry "saturnian
+verse" under "saturnian") is a vowel with a macron, in which the
+macron is lighter than the usual macron, signifying a stressed
+syllable which has a short vowel sound.  This is represented by
+<xsmac/ ("short mac").
+
+Another special character used in pronunciations is an "n" with an
+underline (like a macron, but below the letter), used to represent the
+"ng" sound.  This is coded <nsm/ ("n sub-macron").  The ligated th
+used in pronunciations to depict the "th" sound of "the" is coded as
+<th/.
+
+NOTE: the letter combinations "fi" and "fl" are invariably printed as the
 ligatures &filig; and &fllig;, but these ligatures are not marked as such
 in this transcription, and the two letters are left as individuals.
 
-SPECIAL SYMBOLS
-   The dagger <dag/, double dagger <ddag/, and paragraph mark <para/ are rarely used.
-    The double prime, or "seconds" of a degree is sometimes represented by
-a double "light accent" (code 183 = \'b7).  In other places, and in later
-versions, it is represented by <sec/ = hex a9, in the webfont.
-   The symbols "greater than" <gt/ and "less than" are encountered only
-once, but are distinguished from the right- and left-angle brackets
-(> and <) because of possible typographical differences in some fonts.
-   The schwa is symbolized by <schwa/.  It is not used in the 
-pronunciations, but is mentioned as a symbol.
-   The right-pointing arrow is <rarr/, consistent with ISO 8879.
-
-----------------------------------
-Table 1     
-----------------------------------
-Numbers
-�������                   Hex codes
-1   
    
-11                    (12 is a hard page break, 13 CR, 14 sect break)
-21    
-31   !"#  $%&'(         
-121 yz{|}  ~���         79-7d 7e-82
-131 �����  �����         83-87 88-8c
-141 �����  �����         8d-91 92-96
-151 �����  �����         97-9b 9c-a0
-161 �����  �����         a1-a5 a6-aa
-171 �����  �����         ab-af b0-b4
-181 �����  �����         b5-b9 ba-be
-191 �����  �����         bf-c3 c4-c8
-201 �����  �����         c9-cd ce-d2
-211 �����  �����         d3-d7 d8-dc
-221 �����  �����         dd-e1 e2-e6
-231 �����  �����         e7-eb ec-f0
-241 �����  �����         f1-f5 f6-fa
-251 �����                fb-ff
-
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
-Below is a complete list of the symbols used in the Webster ("webfont") 
-which are encoded in the special font listed above, together with
+* Special symbols
+
+The dagger <dag/, double dagger <ddag/, and paragraph mark <para/ are
+rarely used.
+
+The double prime, or "seconds" of a degree is sometimes represented by
+a double "light accent" (code 183 = \'b7).  In other places, and in
+later versions, it is represented by <sec/ = \'a9.
+
+The symbols "greater than" <gt/ and "less than" are encountered only
+once, but are distinguished from the right- and left-angle brackets (>
+and <) because of possible typographical differences in some fonts.
+
+The schwa is symbolized by <schwa/.  It is not used in the
+pronunciations, but is mentioned as a symbol.  The right-pointing
+arrow is <rarr/, consistent with ISO 8879.
+
+* Symbol summary
+
+Below is a complete list of the symbols used in the Webster, together
+with their "webfont" number (escape sequence), corresponding markup
+entity, and corresponding symbols in ISO 8879 and Tex coding.  Much of
+this table was prepared by Rik Faith, to whom we express our
+appreciation.
+
+The "Uc" column gives the Unicode representation of the character.
+The "nearest ASCII" equivalents are given for those who want to
+display the data as best one can in 7-bit simple ASCII symbols without