Bry, François; Furche, Tim; Linse, Benedikt
AMaχoS—Abstract Machine for Xcerpt. Architecture and Principles.
PPSWR 2006, 4th International Workshop Principles and Practice of Semantic Web Reasoning, 10. - 11. Juni 2006, Budva, Montenegro.
Web query languages promise convenient and efficient access
to Web data such as XML, RDF, or Topic Maps. Xcerpt is one such Web
query language with strong emphasis on novel high-level constructs for
effective and convenient query authoring, particularly tailored to versatile
access to data in different Web formats such as XML or RDF.
However, so far it lacks an efficient implementation to supplement the
convenient language features. AMaχoS is an abstract machine implementation
for Xcerpt that aims at efficiency and ease of deployment. It
strictly separates compilation and execution of queries: Queries are compiled
once to abstract machine code that consists in (1) a code segment
with instructions for evaluating each rule and (2) a hint segment that
provides the abstract machine with optimization hints derived by the
query compilation. This article summarizes the motivation and principles
behind AMaχoS and discusses how its current architecture realizes