Nick Webb wrote:
> From the way I read this, there is an alternative to having a real-time
> clock in each fax machine. The rule is that the required time stamp is the
> date and time the transmission is *sent* (not received) so this implies the
> sender could pass a time stamp with the job rather than the printer
> generating one. If the sender is usually a PC then this would be simple as
> all PCs have real time clocks. If the required water mark information could
> be covered by a usual banner page then why have a real time clock?
>> One possibility for time of day acquisition in a printer is "network time"
> (RFC 867/868 - may be amended by now), however I have no idea how
> ubiquitous this is.
It's pretty ubiqitous, supported by every Unix implementation I'm aware
of, Cisco routers, and available for WinNT. I'm sure other router
vendors implement NTP in their routers as well.
An end-node, such as a printer, need only implement SNTP. SNTP is
basically the 'client' end of NTP and is a simplier implementation. On
another project I'm working on I found a reference to sources for SNTP
in Stevens' book "Unix Network Programming", and there's code on the net
for this, too.
DAYTIME could be used, too (RFC868) and is even lighter weight than