[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: CAP draft 07 issues - sections 3 and 4
To anyone who still hasn't quite grasped the complications that
arise when you assume that SQL can be used to query the
iCalendar data model, please reread John's earlier post that
I quote below, as he explains the issues at hand extremely clearly;
dropping CAP-QL and using the 'simpler' SQL-MIN doesn't begin to
address these issues:
At 03:26 PM 25/03/2002 -0500, John Stracke wrote:
If we dropped USING_PROPERTY, we'd be left with three choices for how to
deal with "property relop1 value1 AND property relop2 value2" (where relop
is a relational operator such as = or <=):
(a) Both references to "property" always refer to the same instance. In
this case, "ATTENDEE='foo@xxxxxxxxxxx' AND ATTENDEE='bar@xxxxxxxxxxx'"
will always be false--not very useful.
(b) "property relop value" always means "there exists some instance of
property such that property relop value". In this case,
"VALARM.DTSTART>=20020325T000000Z AND VALARM.DTSTART<=20020325T235959Z"
means "there exists a VALARM A that starts after 20020325T000000Z and
another VALARM B, possibly the same as A, that starts
before 20020325T235959Z". Again, not very useful.
(c) Define, on a property-by-property basis, which of (a) or (b) applies
for each property in 2445. Lots of work, lots of potential for error,
difficult to maintain as 2445 evolves.
I don't think any of these is viable; I think we need to keep
USING_PROPERTY (and USING_PARAMETER, too, by the same reasoning).