Re: Numerals across code pages

From: Robert Abel (freakrob@googlemail.com)
Date: Fri Feb 19 2010 - 13:44:38 CST

  • Next message: Ed Trager: "Re: Numerals across code pages"

    Maybe the following is at least of some interest:
    http://blogs.msdn.com/michkap/archive/2005/01/27/361413.aspx

    Basically, this talks about an implementation in Windows Notepad that
    converts numbers from many Unicode ranges to their respective values (0
    to 9 for each figure) which can then be used to in the Goto-dialog. It
    seems some farther restrictions on input were made since the blog entry
    was originally posted (i.e. ā¯¶ doesn't work anymore), but I think that's
    the basic idea behind the whole thing the OP was asking for.

    As for output script: if output is surrounded by other text, then maybe
    you can get script conventions out of that.
    However, it might be of use to consider if and when "1 ļ¼‹ äø€" (Latin
    one, full-width plus, ideograph one) should be equal to "2" (Latin two).
    Also, there may be even more signs for addition and multiplication
    operations than you are aware of. So if your implementation computes "1
    + äø€" (Latin one, regular plus, ideograph one) but not the variant with
    full-width plus (see above), then it's even more confusing for regular
    users. In the end it boils down to a lot of code that will cover a
    little percentage of extreme cases and might be perceived as "wrong" by
    regular users, depending on scenario and program of course.

    Robert



    This archive was generated by hypermail 2.1.5 : Fri Feb 19 2010 - 13:48:33 CST