> >(1) It would be trivial to modify xtnpd and ntpd to do the right
 > >thing.  Long term, this is the right thing to do.
 > 
 > Is it trivial enough that you'll implement it and maintain a
 > nonstandard API forever?

>TIODCDTIMESTAMP isn't a "standard" API, either.  I.e. I can't find
>references to it in any POSIX or X/Open standards documents.

That's right. Like I said earlier, it's the _defacto_ standard.  
NTP is the predominant application which would use this, and for xntp,
TIODCDTIMESTAMP is ``the standard'' supported way to get per-second,
on-the-second timestamps from serial ports.

FreeBSD and Linux and (and xntpd) use TIODCDTIMESTAMP.
comp.protocols.time.ntp says Solaris PPS support is about to ship,
ten years after Dave Mill's code.  

My take is that the existing API is just hitting wide deployment and
it's probably too late in the technology/deployment curve to start
setting a new `standard'.  Even if it is better.  *I* don't want to
fight that fight, and I dont see anyone else here who does.

I just want to get code that works with the existing xtpd
(ntpv3)codebase and with the new ntpd (v4) codebase.