Finally, there's the issue of ''ordinal'' identification ie situations where there may be multiple values '''OR''' associations between values, such as degree(s)<->year(s). Right now, relationships don't handle either very well (typically, they need an additional category identifier), but attributes are just the ticket. So even if something was logically a relationship, if you're going to have a complex data set, you may decide to use attributes instead. [[User:Centiare|Centiare]] 09:11, 13 January 2007 (PST) | Finally, there's the issue of ''ordinal'' identification ie situations where there may be multiple values '''OR''' associations between values, such as degree(s)<->year(s). Right now, relationships don't handle either very well (typically, they need an additional category identifier), but attributes are just the ticket. So even if something was logically a relationship, if you're going to have a complex data set, you may decide to use attributes instead. [[User:Centiare|Centiare]] 09:11, 13 January 2007 (PST) |