That's a great question. I'm really not sure if the file "exists" while it's being copied to a file server. Our packages work with a lot of really big (>100 megs) files, and FTP'ing them from remote sites can take several minutes.

It's require a bit more complicated code, but it should be possible to write the code in such a way that the file watcher monitors the file for a minute or so to see if the file's size is changing (I'm assuming a large file "grows" as it's downloaded).

The System.IO.FileInfo has a Length property that returns the size (in bytes) of a file. Once the file is 'there', get the size, wait a minute or two, and compare the size with the initial value.

The System.IO.File class that Silver uses in the article has a number of methods that could also be useful: GetCreationTime, GetLastAccessTime, and GetLastWriteTime. I'm not sure which of these change as a file is being written, but I'd bet at least one of them ought to be useful.

Once it (meaning: size, time, etc.) stops changing, it should be safe to assume the file is stable and ready for processing.