[bug #61998] doc/groff.texi: update Roman numeral info

[Dave]

URL:
  <https://savannah.gnu.org/bugs/?61998>

                 Summary: doc/groff.texi: update Roman numeral info
                 Project: GNU troff
            Submitted by: barx
            Submitted on: Sat 05 Feb 2022 04:10:31 PM CST
                Category: Core
                Severity: 2 - Minor
              Item Group: Documentation
                  Status: None
                 Privacy: Public
             Assigned to: None
             Open/Closed: Open
         Discussion Lock: Any
         Planned Release: None

    _______________________________________________________

Details:

In its discussion of register formats
<http://git.savannah.gnu.org/cgit/groff.git/tree/doc/groff.texi?id=74f1cccc#n6955>,
the Texinfo manual states: "currently, the correct glyphs for Roman numerals
five thousand ('U+2181') and ten thousand ('U+2182') are not used."

But calling these the _correct_ glyphs may be overstating, while at the same
time mentioning only those two glyphs may be understating.

http://en.wikipedia.org/wiki/Roman_numerals#Large_numbers points out: "there
were various extensions to the system [of Roman numerals] designed to indicate
larger [than 3,999] numbers, none of which were ever standardised."  Unicode's
U+2181 and U+2182 are only two of the many variations listed in this section
of the Wikipedia article.  (For those who distrust Wikipedia, this section
cites nine references that can be cross-checked.)

Further, if U+2181 is taken to be the definitively correct Roman 5000,
shouldn't U+2160 be considered the correct Roman 1, U+2164 the correct 5,
etc.?  That is, the significant point seems to be that groff doesn't use the
Roman numeral portion of Unicode's Number Forms block at all, not that it
doesn't use two particular (nonstandard) characters therein.

(Even this is a bit of an oversimplification, as the Unicode standard itself
says: "For most purposes, it is preferable to compose the Roman numerals from
sequences of the appropriate Latin letters."  But while U+2160 to U+217F do
have Latin-letter equivalents, U+2180 through U+2188 do not, so their nonuse
may be more significant. 
(http://en.wikipedia.org/wiki/Numerals_in_Unicode#Roman_numerals has a compact
chart of all Unicode Roman numerals.)  On the other hand, per above, nothing
in the U+2180 - U+2188 range is considered a standard form.  But my feeling is
that none of this is relevant to groff users just needing to know what the "i"
and "I" register formats do.)

I think the points worth mentioning in the .af section of the manual are:

 * there is no standard representation of Roman numerals above 3,999;
 * nonetheless, troff's use of W for 5,000 and Z for 10,000 is an aberration
with no historical precedent, and no modern use outside of troff's
descendants; and
 * following Unicode's own recommendation, groff outputs Roman numerals using
only plain-ASCII characters, not any Roman numeral characters from Unicode's
Number Forms block.

With the above, there is no particular need to single out U+2181 and U+2182.

If others agree with my reasoning, I'll write up some specific text to cover
these points.

No parallel change need be made in man/groff.7.man
<http://git.savannah.gnu.org/cgit/groff.git/tree/man/groff.7.man>, as this man
page gives no detail about how groff renders Roman numerals (nor did CSTR #54
before it).




    _______________________________________________________

Reply to this item at:

  <https://savannah.gnu.org/bugs/?61998>

_______________________________________________
  Message sent via Savannah
  https://savannah.gnu.org/