As seen in comment #0 (original version), "-N --continue" tries to resume the download, while in comment #13 (new version) "--continue" is ignored if "-N" is present. When using a script, it's hard for the script to tell whether the last download completed or not, so you would use -N to prevent creating a new file for every download if the file is already there, also avoiding to download a version that is already there, and you would use --continue to continue an incomplete download. Most people won't care about such differences, but it took two days (with interruptions) to download that big file. Or in short: Your fix seems to change the semantics of wget.