Thanks Tom. I wish I had read this email prior to sending my last one. :-(
On Fri, 2007-11-16 at 13:36 +0000, Thomas Beale wrote: > Sorry to take so long to respond to this. I have looked at the specs > now. The specification is correct - width is required in all > INTERVAL_EVENTs, in fact it is the defiining attribute of an > INTERVAL_EVENT. There is nothing to stop someone creating an > INTERVAL_EVENT that has a width of 0 at this stage - the specifications > don't prevent it; however, instantaneous events are normally represented > as POINT_EVENTs. > > The ambiguity seems to have come from the comment (obviously a hangover > from some previous version of the specification) in section 6.2.4 (class > definition of NITERVAL_EVENT) where it says: "Void if an instantaneous > event." This comment should be removed, and possibly some other comment > added to clarify if width of 0 should be allowed. > > I will add this to the list of Release 1.0.2 CRs. > > - thomas beale > > Tim Cook wrote: > > On Mon, 2007-11-12 at 22:46 +1100, Andrew Patterson wrote: > > > > > >> Hi Tim, I just was going through my spam folder on gmail and found > >> a few openehr emails from you - I don't know why gmail is flagging > >> them as spam.. > >> > > > > Maybe I should take that as a hint? ;-> > > > > > >>> The width attribute is defined as a DV_DURATION type and it is required. > >>> However the text says that it should be Void if the Interval_Event is > >>> instantaneous. There is also an invariant preventing width from being > >>> Void. > >>> > >> I always read this to mean "where the event is instantaneous > >> (i.e a POINT_EVENT) the duration is void". So I would contend > >> that an INTERVAL_EVENT cannot be instantaneous, > >> rather than allowing the INTERVAL_EVENT duration to be 0. > >> > >> I agree that the wording could be better to clarify which interpretation > >> is right.. > >> > > > > > > Well, I certainly agree that an instantaneous event cannot be an > > interval event. However, there is some confusion in (my mind at least) > > using the various DV_DURATIONS in History. But the biggest issue I have > > is in dealing with the fact that an Interval Event has one and only one > > DV_DURATION. DV_DURATION can never be Void. It MUST return a valid ISO > > Date_Time value. > > > > > > So maybe there is a typo somewhere in this chain but the fact is that we > > have a serious conflict here. *If* an interval exists, then it's "width" > > can never be Void because DV_DURATION does not allow a Void. > > > > Cheers, > > Tim > > > > > > > > > > -- Timothy Cook, MSc Health Informatics Research & Development Services http://timothywayne.cook.googlepages.com/home LinkedIn Profile:http://www.linkedin.com/in/timothywaynecook
