Poster Presentation: Xcerpt and XChange – Logic

Programming Languages for Querying and

François Bry, Paula-Lavinia Pătrânjan and Sebastian Schaffert

Institute for Informatics, University of Munich, http://www.pms.ifi.lmu.de/

Motivation
The Semantic Web is an endeavor aiming at enriching the existing Web with
meta-data and data (and meta-data) processing so as to allow computer systems
to actually reason with the data instead of merely rendering it. To this aim, it is
necessary to be able to query and update data and meta-data. Existing Semantic
Web query languages (like DQL2 ) are special purpose, i.e. they are designed for
querying and reasoning with special representations like OWL or RDF, but are not
capable of processing generic Web data. On the other hand, the language Xcerpt [1]
[2] presented in the poster is a general purpose language that can query any kind
of XML data and at the same time, being based on logic programming, provides
advanced reasoning capabilities. It could thus serve to implement a wide range of
different reasoning formalisms.
Likewise, the maintenance and evolution of data on the (Semantic) Web is nec-
essary: the Web is a ”living organism” whose dynamic character requires languages
for specifying its evolution. This requirement regards not only updating data from
Web resources, but also the propagation of changes on the Web. These issues have
not received much attention so far, existing update languages (like XML-RL Up-
date Language [3]) and reactive languages [4] developed for XML data offer the
possibility to execute just simple update operations and, moreover, important fea-
tures needed for propagation of updates on the Web are still missing. The language
XChange also presented in the poster builds upon the query language Xcerpt and
provides advanced, Web-specific capabilities, such as propagation of changes on the
Web (change) and event-based communications between Web sites (exchange).
Xcerpt: Querying Data on the Web
Xcerpt is a declarative, rule-based query language for Web data (i.e. XML doc-
uments or semistructured databases) based on logic programming. An Xcerpt pro-
gram contains at least one goal and some (maybe zero) rules. Rules and goals consist
of query and construction patterns, called terms in analogy to other logic program-
ming languages. Terms represent tree-like (or graph-like) structures. The children
of a node may be either ordered (as in standard XML) or unordered (as is common
in databases).
Data terms are used to represent XML documents and the data items of a
semistructured database. They are similar to ground functional programming ex-
pressions and logical atoms. A database is a (multi-)set of data terms (e.g. the
Web).
Query terms are patterns matched against Web resources represented by data
terms. They are similar to the latter, but augmented by variables (for selecting
data items), possibly with variable restrictions (the possible bindings are restricted
to certain subterms), by partial term specifications (omitting subterms irrelevant
1
This research has been funded by the European Commission and by the Swiss Fed-
eral Office for Education and Science within the 6th Framework Programme project
REWERSE number 506779 (cf. http://rewerse.net).
2
DAML Query Language, http://www.daml.org/dql.
to the query), and by additional query constructs like subterm negation, optional
subterm specification, and descendant.
Construct terms serve to reassemble variables (the bindings of which are specified
in query terms) so as to construct new data terms. Again, they are similar to the
latter, but augmented by variables (acting as place holders for data selected in a
query) and the grouping construct all (which serves to collect all instances that
result from different variable bindings).
Construct-Query Rules (short: rules) relate a construct term to a query con-
sisting of AND and/or OR connected query terms. Rules can be seen as “views”
specifying how documents shaped in the form of the construct term can be ob-
tained by evaluating the query against Web resources (e.g. an XML document or a
database). Xcerpt rules might be chained like active or deductive database rules to
form complex query programs.
XChange: Evolution of Data on the Web
Exchanging Events on the Web. The language XChange aims at establish-
ing reactivity, expressed by reaction rules, as communication paradigm on the Web.
With XChange, communication between Web sites is peer-to-peer, i.e. all parties
have the same capabilities and can initiate communication, and synchronization
can be expressed, so as to face the fact that communication on the Web might be
unreliable and cannot be controlled by a central instance.
The processing of events is specified in XChange by means of event-raising rules,
event-driven update rules, and event-driven transaction rules. Event-raising rules
specify events that are to be constructed and raised as reaction to incoming (internal
or external) events.
Propagating Changes on the Web. XChange provides the capability to spec-
ify relations between complex updates and execute the updates conformly (e.g. in
booking a trip on the Web, one might wish to book an early flight and of course the
corresponding hotel reservation). To deal with network communication problems,
an explicit specification of synchronization operations on updates is needed, a (kind
of) control which logic programming languages lack.
Update rules are rules specifying (possibly complex) updates. The head of an
update rule contains patterns for the data to be modified augmented with update
operations (i.e. insertion, deletion, replacement), called update terms, and the de-
sired synchronization operations.
As sometimes complex updates need to be executed in an all-or-nothing manner
(e.g., in booking a trip on the Web, a hotel reservation without a flight reservation
is useless), the concept of transactions (one or more updates treated as one unit)
is supported by XChange. Transactions may be executed on user requests or as
reactions to incoming events (the latter transactions are specified using event-driven
transaction rules).

References
1. Bry, F., Schaffert, S.: Towards a Declarative Query and Transformation Language for
XML and Semistructured Data: Simulation Unification. In: Proc. Int. Conf. on Logic
Programming (ICLP). LNCS 2401, Springer-Verlag (2002)
2. Bry, F., Schaffert, S.: An Entailment for Reasoning on the Web. In: Proc. of Rules and
Rule Markup Languages for the Web (RuleML03), Springer-Verlag (2003)
3. Liu, M., Lu, L., Wang, G.: A Declarative XML-RL Update Language. In: Proc. Int.
Conf. on Conceptual Modeling (ER 2003). LNCS 2813, Springer-Verlag (2003)
4. Papamarkos, G., Poulovassilis, A., Wood, P.T.: Event-condition-action rule languages
for the semantic web. In: Workshop on Semantic Web and Databases, Berlin (2003)