[Babase] Re: Upmpi ready-ish

Karl O. Pinc kop at meme.com
Mon Dec 14 21:16:33 EST 2009


On 12/14/2009 07:51:03 PM, Susan Alberts wrote:
> 
> On Dec 11, 2009, at 5:35 PM, Karl O. Pinc wrote:
> 
> > On 12/11/2009 02:26:21 PM, Lacey Maryott Roerish wrote:
> >> ERROR: Mpiid 2747: Mpidid 8027: VS: Seq 1: MPI_PARTS cannot have
> two
> >> rows
> >> for the same multi-party interaction with a Seq of 1 unless all 
> the
> >> MPIAct
> >> values are VS
> >
> > I thought that was Susan's rule (per something about the
> > dyadic interaction sheet in section 10.1 of the monitoring
> > guide?).
> >
> > I don't see it documented though, so one way or the other
> > I'll need to make a change.  I don't suppose the rule is:
> > "VS must have a seq of 1 and there _must_ be more than one
> > such row?"
> 
> I think the rule is that two rows can't both have seq = 1 unless they 
> 
> are both "versus" interactions, ie that the team saw two "versus"  
> interactions simultaneously begin the mpi.  I think we can probably  
> safely change this rule to just say that there can only be _one_ row  
> with seq = 1. We wouldn't want to always say the VS must have a seq = 
> 
> 1. But I think we can make the only one seq = 1 rule. Does that help?

I think we want the original: "two rows can't both have seq = 1 
unless they are both "versus" interactions" unless you have
some reason to want a change.

Lacey and I have worked it out.




Karl <kop at meme.com>
Free Software:  "You don't pay back, you pay forward."
                 -- Robert A. Heinlein




More information about the Babase mailing list