The question Davy is referring to is a problem for both print and braille
and causes internationalization issues. Here is some background to remember.
According to Wikipedia the term "decimal mark" is the descriptor for the
character used to separate the whole and fractional portions of a decimal
number. This descriptor is not a Unicode character name. The Unicode
character used for a decimal mark is different in different locales. In the
United States the decimal mark is represented by the Full Stop Unicode
Character. In the UK the decimal mark is represented by the Comma Unicode
Character. UEB allows for both these possibilities with both being referred
to as decimal signs. In UEB a decimal sign must be within the scope of a UEB
Numeric Indicator.
Strings of digits often have embedded full stops as, for example, in the
Liblouis version identifier v2.6.5. These full stops are not intended as
decimal marks. However, UEB rules allow a full stop to be within the scope
of a Numeric Indicator with the exception per Rule 6.4 that if a leading
full stop is obviously a period and not a decimal point it not be prefaced
by a Numeric Indicator.
Distinguishing the semantics in this situation is not always easy. That is
why the preferred print style avoids a leading decimal point by using a zero
before the decimal point.
So in both the US and the UK v2.6.5 would be translated to UEB braille as
v#b4f4e ASCII braille. Thus in the US both the print and braille readers can
only distinguish a period from a decimal point by context. The supposed
advantage to the UEB formulation is that the translation can be done
automatically as long as the print source is appropriate for the locale for
which the braille translation is being produced.
Current American English braille uses dots-256 to translate a full stop
intended as a period punctuation mark and a dots-46 within the scope of a
number sign to translate a full stop intended as a decimal point. This has
the disadvantage that the transcriber or transcription software must
determine the local semantics of the print full stop character but has the
advantage that the braille output has sufficient information to be
backtranslated automatically to either US or UK print.
I recommend that liblouis at least use the heuristics that if an input print
item consists of a sequence of digits with more than one embedded full stop
that these full stops be treated as having the semantics of period
punctuation marks in all locales. There should also be a method for
recognising user markup intended to distinguish the semantics for cases
where the built-in heuristics are inadequate.
Note that currently something similar needs to be done to distinguish the
semantics of the Unicode characters which can represent either an apostrophe
or a closing single quotation mark as the correct UEB translation depends on
the semantics.
HTH,
Susan Jolly
For a description of the software, to download it and links to
project pages go to http://liblouis.org
