Data elements as unique identifiers

In a database every entry needs a unique identifier, something distinguishing one record from all others and which is used to identify a record. There are certain possibilities of using a unique identifier: At first glance, one could think of using the orthographic representation of a concept as unique identifier. It is -- relatively -- easy to implement a search function for this unique identifier because every record has it; and of course every concept in a termbank should have an orthographic representation.

In a fully defined semantic net this is not always the case. Unfortunately, there are cases where there are nodes  in a semantic net or in an ontological hierarchy which are not named, there are lexical gaps.  For example it is possible that there is a common hyperonym of two terms, and the terms are more closely related to each other than another term having the same hyperonym. If someone presents this relation in an ontological hierarchy, an unnamed node for an extra hyperonym of the more closely related terms may be the result. In these cases terminologists sometimes suggest a new term, but that does not mean that it is possible to establish this term. In these cases the de facto value of the node name remains empty.

 
• The problem of homographs
• Forming unique identifiers: adding information
• Forming unique identifiers: Enumeration of the nodes of a semantic net
• Forming unique identifiers: using arbitrary numbers
• Forming unique identifiers: summary

Thorsten Trippel
Fri May 21 13:04:11 MET DST 1999