Semantically correct mappings

From: <andymaule_at_gmail.com>
Date: 2 Dec 2005 08:14:29 -0800
Message-ID: <1133540069.763534.314270_at_f14g2000cwb.googlegroups.com>



Hi there,

I'm a UK PhD student looking for some clarifications. Hopefully someone can help :)

I've been reading "Preserving mapping consistency under schema changes" by Yannis Velegrakis, RenŽee J. Miller and Lucian Popa. http://portal.acm.org/citation.cfm?id=1023919&coll=Portal&dl=ACM&CFID=61962995&CFTOKEN=90717062

I want to confirm that I'm correct in thinking the following according to the terminology defined in this paper, and the papers that are referenced here.

Sound mapping, complete mapping, or exact mappings are phrases to determine the extent of a mapping i.e. a sound mapping means everything in the source schema queries Q(S) is included in the target schema queries Q(T) but Q(T) may be a superset of Q(S).

A semantically correct mapping is a mapping that correctly preserves the constraints of S in T, so that the semantics of S are not violated in T but they don't have to be explicitly replicated.

Therefore it is possible to have a semantically correct mapping, where Q(T) is only a small subset of Q(S). It is also possible to have a mapping that is both semantically correct AND sound.

Is this correct? Sorry if I've lost some of you along the way, but it's quite a specific question. Let me know if I need to clarify this a bit more clearly.

Thanks

Andy Maule Received on Fri Dec 02 2005 - 17:14:29 CET

Original text of this message