>So your suggesting to recommand to this SMALL insurance company to purchase a server and then the application will work?

Mike, I suggest you hire someone to look at the problem who's more in touch with the reality of the client's enviroinment. With an insurance company, my gut reaction is there's a security issue now, the likelihood of adequate packup and control of the shared resources is poor, and, to be frank, administering a peer-to-peer LAN is at least as difficult as administering a small domain; with centralized control of access and permissions you greatly reduce risks; you can successfully deploy fault-tolerant storage at low cost (ie, if you're worried about having 15GB of shared hard disk going south and losing the inadequately-backed-up-data as a result, the cost of mirrored drives on the Win2K server is a whopping $120 for a second 15GB disk drive, since fault tolerance is part of the Server's native ability), you can provide access control at the level of file granularity on a per-user basis, create groups and assign rights to groups, and then assign permissions to groups to minimize the admin tasks, there's a web-accessible interface for managing the server, and remote admin capabilities. IOW, what you're doing is, from my POV, creating a disaster that will be pinned on you because you flat out gave the client incredibly poor advice about how their automation should be done. You've set yourself up to take the blame when some poor dumb secretary sits at the wrong machine and Ctrl-Alt-Dels the entire LAN into the void.

If you're MS-phobic, you can use NetWare to provide similar authentication, file service and ACL, but that's not what I would do. Christ, if your clients are going to cringe over the cost of a server, you could use Linux, but you're going to run up the cost of administration and service on the server in the long run.

I am a small independant (admittedly, many of my clients are large corporations, but nearly everyone hires me because I have a clue about their actual IT requirements, not to play codemonkey). I go in and make sure that my client has a clue about the consequences of doing things wrong. For me, that means I require little things like UPS systems on all stations, real servers with decent AC, and (after making suitable backups ahead of time) make sure they know what the vulnerabilities are of their systems, by practical demonstration if necessary. Make'm watch Eyes Wide Shut, and (after making a known good backup), go over to some secretary's desk, access the shared database directory of the so-called server, and delete it. Or log out at its console, and just forget to log into Windows (hit the Cancel button at the Windows login.) Or pull the plug. After the horrorific screams, drop the backup tape in and restore it. That may give them a clue about where they should spend a few shekels for a little peace of mind.

>I have a feeling it's a surefire way to loose the contract.

What you're doing is an almost sure-fire way to be negligent. I carry professional liability insurance. Do you?

>Microsoft's Peer-to-Peer is meant for a small office environment were a server requires (sometimes) the assistance of an addministrator (hence to cost of $30,000.00 (canadian) mentionned earlier.)

YMMV, I think they're spending as much with no ACL and decentralized control with infinite vulnerability.