RE: When is a character a currency sign?

[Asmus Freytag]

Unicode assigns the general category value, "Sk", or "Symbol, [k]urrency" 
to all characters whose *primary* function is to act as a currency symbol.

That excludes all characters that have other, unrelated uses, as long as 
those are not more specialized than the use as currency sign. That's an 
uwritten, but workable definition and it has so far not lead to too many 
issues in deciding whether or not to classify a character as Sk.

Even undisputed currency symbols, such as $ have by now picked up other 
uses, for example as part of identifiers for certain programming languages. 
Any currency that uses (or used to use) a single non-specialized letter as 
its common symbol would have the opposite problem: in that case the 
character is *used* as currency symbol, but *isn't* a currency symbol. In 
other words, the label now attaches to the function not the essence of the 
symbol.

In that light it makes sense to define as 'currency sign' any *single* 
character that is used to denote a currency. As long as one doesn't insist 
that 'currency sign' implies 'not a letter or ideograph' such a definition 
might be useful, for example in a parser that attempts to isolate possible 
currency expressions from running text.

Such a parser would need a list of such currency signs, as well as a list 
of 'currency words', for currency designators that are more than one 
character. Perhaps the broader, operational definition of 'currency sign' 
is then not so helpful after all, certainly not as it regards simple 
letters. However, with ideographs, where there might be simila but not 
identical characters, noting explicitly which one is used as the 'currency 
sign' might be useful -- as long as it isn't understood as an exclusive 
property.

A./