Table that won't sort correctly: wash-up

[Simon BUCH]

Hello,

I suspect the effect that you may be observing is Microsoft's handling 
of some multibyte characters, where they get mutated to the Unicode 
replacement character '�' U+FFFD - which appears to happen very often 
when content is imported or pasted into FrameMaker from Microsoft word.

The character is not shown in FrameMaker, but if the file is exported to 
XML or HTML, it is rendered as a question mark in a rhombus diamond 
symbol: � - as shown on the web page:
http://www.fileformat.info/info/unicode/char/fffd/index.htm
Some browsers and applications display it as a hollow box character.

If you know the exact position of one of the errant character in the 
FrameMaker document, you can select the single character by using the 
cursor keys, and press the delete key. For some document I found it was 
possible to select Select Edit > Find/Change... ; specify text to find 
as: \x0d ; and then click Change All.

Regards
// Simon BUCH




On 09/12/2012 12:52, Steve Rickaby wrote:
> At 09:24 +0000 8/12/12, Steve Rickaby wrote:
>
>> Plain old MIF-ing fixes it.
> This is interesting, as it provided a specific example of an invisible (to the eye) problem that MIF-washing fixes.
>
> To recap, I had a large table whose first column contained entries of the form '[xxxx]', where 'xxxx' was a reference marker. The text had originated in Word, and the table didn't sort properly when the first column was used as the sort key.
>
> In the MIF, quite a lot of the first column entries were of the form:
>
>        <ParaLine
>          <String `['>
>          <String `xxxx]'>
>        > # end of ParaLine
>
> In other words, the '[xxxx]' entry consisted of two separate but adjacent strings. This messed up the sort. The only visible evidence of this on the screen was that it took one more right-cursor movement to advance from the start of the line to past the '[' as expected. Also, the wildcard pattern '^[' didn't get a match.
>
> Clearly one of the things that FrameMaker does when parsing MIF is to spot adjacent strings like this and concatenate them. Useful to know. Why the text came out of Word like this is another question altogether. But with Word, who knows?
>