Scott Ribe wrote: > Personally, I'd put them on the file system, because then backup software > can perform incremental backups. In the database, that becomes more of a > difficulty. One suggestion, don't use a file name from a hash to store the > image, just use the serial id, and break them up by hundreds or thousands, > iow image 1123 might be in images/000/000001/000001123. > > -- > Scott Ribe > scott_ribe@xxxxxxxxxxxxxxx > http://www.killerbytes.com/ > (303) 722-0567 voice I think I know the answer, but if you don't have an "application server" - ie a webserver, etc, and many of the workstations/clients that need access to the images but may not have access to a network share, isn't the database the only choice ? - or is there a postgresql function/utility that will "server" the file from the file system based on the reference/link embeded in the database ?? Geoff.
