[groff] 06/08: groff_mdoc(7): Revise "Predefined Strings" again.

[G. Branden Robinson]

gbranden pushed a commit to branch master
in repository groff.

commit 9a3ccc6772998093d5b0cbfd69703a4e179f7b14
Author: G. Branden Robinson <g.branden.robinson@gmail.com>
AuthorDate: Sun Nov 1 13:33:10 2020 +1100

    groff_mdoc(7): Revise "Predefined Strings" again.
    
    Per historical briefing from Ingo Schwarze, the strings documented in
    this section are deprecated and have been for at least a decade; say so.
    
    Update table to recommend alternative input for each.  (You know what's
    coming next--I fix the page to take its own advice.)
    
    Sync glyph descriptions with those in groff_char(7).
    
    Shorten "Unicode" column heading to "UCS" because we ran out of room.
    
    Give some historical background to motivate the preference for special
    character escapes (or literal input in a few cases).
---
 tmac/groff_mdoc.7.man | 80 ++++++++++++++++++++++++++++++---------------------
 1 file changed, 48 insertions(+), 32 deletions(-)

diff --git a/tmac/groff_mdoc.7.man b/tmac/groff_mdoc.7.man
index bac1238..b872c32 100644
--- a/tmac/groff_mdoc.7.man
+++ b/tmac/groff_mdoc.7.man
@@ -4072,40 +4072,39 @@ It is neither callable nor parsed and takes no 
arguments.
 .
 .Sh "Predefined strings"
 .
-The following strings are predefined,
-and should be used in preference to any corresponding special character
-escapes in
+The following strings are predefined for compatibility with legacy
 .Nm \-mdoc
 documents.
 .
-They degrade gracefully when rendered on devices with more limited glyph
-repertoires.
+Contemporary ones should use the alternatives shown in the
+.Ql Prefer
+column below.
 .
 .
 .Pp
 .TS
-lb lb lb lb lb
-l l l l l.
-String 7-bit   8-bit   Unicode Meaning
-<=     <=      <=      \*[<=]  less than or equal to
->=     >=      >=      \*[>=]  greater than or equal to
-Rq     "       "       \*[Rq]  right double quote
-Lq     "       "       \*[Lq]  left double quote
-ua     ^       ^       \*[ua]  up arrow
-aa     \[aq]   \'      \*[aa]  acute accent
-ga     \`      \`      \*[ga]  grave accent
-q      \&"     \&"     \*[q]   neutral double quote
-Pi     pi      pi      \*[Pi]  pi
-Ne     !=      !=      \*[Ne]  not equal to
-Le     <=      <=      \*[Le]  less than or equal to
-Ge     >=      >=      \*[Ge]  greater than or equal to
-Lt     <       <       \*[Lt]  less than
-Gt     >       >       \*[Gt]  greater than
-Pm     +\-     \[+-]   \*[Pm]  plus or minus
-If     infinity        infinity        \*[If]  infinity
-Am     \*[Am]  \*[Am]  \*[Am]  ampersand
-Na     \*[Na]  \*[Na]  \*[Na]  not a number
-Ba     \*[Ba]  \*[Ba]  \*[Ba]  vertical bar
+lb lb lb lb lb lb
+l l l l l l.
+String 7-bit   8-bit   UCS     Prefer  Meaning
+<=     <=      <=      \*[<=]  \[rs](<=        less than or equal to
+>=     >=      >=      \*[>=]  \[rs](>=        greater than or equal to
+Rq     "       "       \*[Rq]  \[rs](rq        right double quote
+Lq     "       "       \*[Lq]  \[rs](lq        left double quote
+ua     ^       ^       \*[ua]  \[rs](ua        vertical arrow up
+aa     \[aq]   \'      \*[aa]  \[rs](aa        acute accent
+ga     \`      \`      \*[ga]  \[rs](ga        grave accent
+q      \&"     \&"     \*[q]   \[rs](dq        neutral double quote
+Pi     pi      pi      \*[Pi]  \[rs](*p        lowecase pi
+Ne     !=      !=      \*[Ne]  \[rs](!=        not equals
+Le     <=      <=      \*[Le]  \[rs](<=        less than or equal to
+Ge     >=      >=      \*[Ge]  \[rs](>=        greater than or equal to
+Lt     <       <       \*[Lt]  <       less than
+Gt     >       >       \*[Gt]  >       greater than
+Pm     +\-     \[+-]   \*[Pm]  \[rs](+\-       plus or minus
+If     infinity        infinity        \*[If]  \[rs](if        infinity
+Am     \*[Am]  \*[Am]  \*[Am]  &       ampersand
+Na     \*[Na]  \*[Na]  \*[Na]  NaN     not a number
+Ba     \*[Ba]  \*[Ba]  \*[Ba]  |       bar
 .TE
 .
 .
@@ -4114,18 +4113,35 @@ The names of the columns are shorthand for standard 
character encodings;
 \*[Lq]7-bit\*[Rq] for ISO 646:1991 IRV (US-ASCII),
 \*[Lq]8-bit\*[Rq] for ISO 8859-1 (Latin-1) and IBM code page 1047,
 and
-\*[Lq]Unicode\*[Rq] for ISO 10646.
+\*[Lq]UCS\*[Rq] for ISO 10646 (Unicode character set).
 .
-The strings are defined such that the best available representation will
-be used.
+The strings were originally defined
+(circa 1989)
+such that the best available representation would be used.
 .
 For example,
-a Unicode-enabled terminal device will have optimal glyph
-representations for all strings,
+a Unicode-enabled terminal device using a well-chosen font will have
+optimal glyph representations for all of the above,
 whereas a Latin-1 terminal's only advantage over an ASCII one
 (for these glyphs)
 is the plus-minus sign.
 .
+Old typesetters lacked directional double quotes, \" like the C/A/T
+and would produce repeated directional single quotes,
+\[oq]\[oq]like this\[cq]\[cq];
+early versions of
+.Nm \-mdoc
+in fact defined the
+.Ql Lq
+and
+.Ql Rq
+strings this way. \" thanks to Ingo Schwarze for the research
+.
+In the years since,
+output drivers have taken on the responsibility of glyph substitution;
+they are the part of the formatting process with relevant knowledge of
+the available repertoire.
+.
 .
 .Pp
 String names which consist of two characters can be written as