[Groff] Box-building special characters

[Bernd Warken]

###

Does anyone have information on the special characters \(br and \(bv ?
Do the abbreviations mean `bar' and `bar vertical'?

devps maps \(bv to Postscript braceex (brace extension).  This is in
accordance with the old Bourne troff documentation, which mentions it
among the large bracket building characters and draws it to fit these.
Kernighan draws it a quite unclear way as a vertical bar that does not
fit any other characters.

\(br seems to be interpreted as an alias for the normal bar `|' in
devps.  In the Bourne documentation, it is placed among the mathematical
symbols, drawing at the same vertical position as \(or, but a bit longer
and thinner.  The Kernighan documentation places it besides "or", but
draws it in an off-center shape.

As to the Unicode device, groff correctly maps \(bv to U+2502 BOX DRAWINGS
LIGHT VERTICAL.

But \(br should be mapped differently.  Possible characters:

U+007C VERTICAL LINE = VERTICAL BAR
U+01C0 LATIN LETTER DENTAL CLICK
U+2223 DIVIDE
U+2578 LIGHT VERTICAL BAR

and further Chinese/Japanese and Dingbats characters.

###

The Unicode mapping of the large square bracket building characters si
dangerous:

lc U+2308
rc U+2309
lf U+230A
rf U+230B

Tho this mapping establishes a similar naming scheme both in groff and
Unicode, these codes cannot be taken because there is no fitting
for the vertical extension character \(bv.

So the BOX DRAWINGS LIGHT codes should be the suitable mapping:

lc      24      0       0x250C
rc      24      0       0x2510
lf      24      0       0x2514
rf      24      0       0x2518

###

There seems to be no box building character for the horizontal
extension.

\(rn is a look-alike, but it is mapped to U+203E OVERLINE.  As this can
be used for composing mathematical characters this mapping seems
alright.

Has anyone an idea, why it was called "rn"?

###

The 4 characters needed to build large paranthesis are mapped correctly.

But in groff, there are 2 more special characters large curly brackets
(used in mathematics and music), \(lk and \(rk.

The bad news is that groff wrongly maps them to the simple braces
{ and }.

The very bad news is that these characters seem to have no Unicode
equivalents.  That's strange, the Unicode Consortium should have
included at least all characters from Kernighan's standard troff
documentation.  So could someone please check this.

Do we need to start another Unicode character submission war?  Grrr...


Bernd Warken <address@hidden>