I’ve encountered a bizarre spell-checking anomaly when converting text documents from one word processor format to another. I have a collection of short stories that I’ve written over the years, and over time I’ve used different word processors, specifically, Microsoft Word for Mac, OpenOffice, NeoOffice (Mac OS X native version of OpenOffice), and Apple’s “Pages” word processing app. The conversion process has gone something roughly like this:
• Write the original story in a text editor, usually a Mac shareware program called Tex-Edit Plus, which does plain text but has many options for formatting the text; save as .txt or .rtf
• Import or copy/paste the .txt or .rtf document into MS Word and save as a .doc file
• Decide to de-Microsoft myself and install OpenOffice (and later, NeoOffice), and convert .doc files to .odt files
• Get fed up with Open/NeoOffice sluggishness and go back to .doc
• Install Apple’s Pages and begin process of converting everything once again
At some point I started noticing a strange spell-checking problem: all of these documents were created using US English spelling, but when I opened the MS Word or Pages versions, all sorts of words were marked with red squiggly underlines, indicating misspellings. But the words were spelled correctly. Bringing up the spell-checking menu revealed the problem: both MS Word and Pages were using the UK English dictionary for spell-checking! This was apparent by the way they kept suggesting changes like “colour” instead of “color”. On top of that, both apps claimed to be using US English dictionaries, and neither would let me change anything in that regard.
I finally narrowed the problem down, after much aggravation: the problem was only occurring in documents I had converted from Open/NeoOffice .odt format into .doc or .pages format (documents originally created in MS Word or Pages and never converted to .odt did not have this problem). In some cases I still had the .odt version, and opening it in NeoOffice showed no spell-checking issues. But the .doc or .pages versions that had been converted from .odt all suddenly seemed to think they should be written with UK English spelling.
(Regarding the Pages versions, I should note that neither MS Word nor Pages will open .odt documents, and neither NeoOffice nor MS Word will open .pages documents, so converting from NeoOffice to Pages involved opening the files in NeoOffice and then resaving them as .doc files, and then opening the .doc in Pages and resaving as .pages files.)
So it appears to me that Open/NeoOffice does something with its formatting code that is being completely misinterpreted by both MS Word and Pages. And, whatever this code problem is, it’s carried over during a copy/paste operation, i.e. if I merely selected and copied the text in the .odt document and pasted it into a new .doc or .pages document, the spelling bug would be retained. What I ultimately ended up having to do was export the buggy .doc or .pages file back to a plain-text, completely unformatted document, save the .txt file, and then reopen the .txt file in MS Word or Pages and redo all the formatting. This would result in an honest-to-goodness US English document as far as the spell-checkers are concerned.
I’m going to guess that OpenOffice/NeoOffice simply has/had a bug in its “Save as MS Word Document” filter that was miscoding the language in the resulting exported .doc file, causing both MS Word and Pages to think they were reading a UK English document. But still, that wouldn’t seem to explain why the error would be carried over during a copy/paste.
Any idea what was going on here?