L2/07-140

Source: Deborah Goldsmith
Date: 04/18/2007 02:16 PM
Subject: Prefix_Numeric and Postfix_Numeric in line break vs. word break

I would like to have a topic on Prefix_Numeric and Postfix_Numeric in
line break vs. word break. Here is a short summary of the issue:

Currently, UAX 29 places word boundaries between currency symbols and
numbers, and also between numbers and trailing symbols such as %. In
line break, these characters have the categories Prefix_Numeric and
Postfix_Numeric, respectively. In these examples:

$1,234.56
-47.5
+33
87.5%
44‹
20Ž

a word break occurs between the prefix and the number, and between
the number and the suffix. However, a line break is not allowed in
the same locations.

Interestingly, the ICU documentation on this topic is incorrect. If
you look at:

http://icu-project.org/userguide/boundaryAnalysis.html

you will see that the example image shows no word boundary between $
and the following number. Also, the summary below the image says:

"2. Numbers are kept together, including any commas, points or
currency symbols."

which is clearly incorrect according to the current rules.

I would like to have a discussion about whether we should change this
behavior (or not), and treat characters with the line break
properties Prefix_Numeric and Postfix_Numeric (or some subset of
those) in the same way for word break. Apple has received user
complaints about the current behavior. I am hard pressed to come up
with an example as to why word breaks between numbers and their
prefixes and suffixes would be useful.

There was a brief discussion about this at a prior UTC, but no
conclusion was reached.

Thanks,
Deborah