I suppose if the code was light enough i wouldnt have too much of a performance issue.

I could store current access information in some form of object which would persist in server memory.

I still have concerns over the potential size of the object should high use be an issue.

One thing i was thinking was regarding what you were saying about moving the image from a private location be it physical or database to a public location and having clean up code.

With unix file systems and symlinks this would be cool as i would only have to check when the server creates a 404 error but with bloody NT.

Thanks for the idea i will try out a couple of methods this weekend.

--
nick