I've got some word documents coming from a weird source. They originate in Framemaker. Framemaker does a much better job exporting to html than rtf, so I pickup the html and doctor it a bit to remove junk that Framemaker produces. Then I open the html in Word, save it as Word files and then do the confluence import. Problem is, the images get lost. They're there in word, from the html, but confluence doesn't pick them up.
Does anybody know what might be happening here? It's really a bummer to have to figure out which images from the original framemaker belong where in confluence.