Thanks PsychoSpunk and tinman for the info. Looks like I'm bagging MSSQL::sqllib and looking at DBD::ODBC. I've written some things on Win32 using the DBD::ADO and it worked great with DBI, so I'm certainly more partial to DBI. Switching to a different DBD is fine with me, just wasn't sure and wanted the wisdom of the monks.
Thanks again.
Comment on Re: Differences with methods to access SQL Server 2000.
I have been using DBD::Sybase (having used Sybperl before), with FreeTDS on Linux, and it's quite slow.
It might be worth a try though, as these are easier to get going than ODBC.