This document is also available in non-normative PDF version.
Copyright © 2004 DERI®, All Rights
Reserved. DERI liability,
trademark, document use, and software licensing rules apply.
This document exemplifies the usage of the Web Service Modeling Ontology WSMO for modeling possible Web Service driven applications. The intent of this document is to exemplify use cases with usage scenarios of Semantic Web Services on the one hand, and on the other to showcase modeling with WSMO as an evaluation with real-world testing as support for recursive development of WSMO. For use case modeling, we stick to the latest final working draft of Web Service Modeling Ontology WSMO, Version 0.2.
WSMO Standard: D2 v0.2 Web Service Modeling Ontology (WSMO) , current version at: http://www.wsmo.org/2004/d2/v0.3/
WSMO Primer: D3.1 v01. WSMO Primer
WSMO Reasoning: D5.1 v01. Inferencing Support for Semantic Web Services: Proof Obligations.
This document exemplifies the usage of the Web Service Modeling Ontology WSMO for describing relevant aspects for Semantic Web Services. Therefore, we describe possible use cases of Semantic Web Services and showcase how these can be modeled with WSMO, especially for support of the Semantic Web Service usage scenarios in particular use cases. We briefly replicate the objectives and the approach of WSMO and outline use cases within possible usage scenarios of Semantic Web Services. Then, we showcase how specific use cases can be modeled in WSMO along with explanations on the modeling decisions. Besides, we provide the WSMO models in a computational format.
This Deliverable is intended to evolve in accordance to the ongoing development of the WSMO project, serving as a testing environment and providing input for a recursive, real world testing development of WSMO. In the longer run, additional use cases will be added in order to widen possible solutions for Semantic Web Service technologies around WSMO.
This document is organized as follows: the remainder of Section 1 replicates the objectives and approach of WSMO; Section 2 discusses possible application areas of Semantic Web Services. Section 3 provides the modeling of the use cases in WSMO, pointing out the WSMO approach for Semantic Web Service technologies. Section 4 concludes the document. The complete WSMO models as computational resources are provided in the Appendices.
The use case modeling in this document relies on the latest final working draft of the Web Service Modeling Ontology WSMO, Version 0.2.
A Web Service is a piece of software accessible via the Web. Current Web Service technologies allow exchange of messages between Web Services [SOAP], describing the technical interface [WSDL], and advertising a Web Services in a registry [UDDI]. These technologies do not provide any information about the meaning of information used, neither do they explicitly describe the functionality of a services as needed for automated usage of Web Services. There are also an alternative and complement specifications such as for example [ebXML], which are still perceived to have much potential for widespread adoption as Web Services. ebXML has become the global electronic business specification, that defines a framework for global electronic business. Enhanced Web Service technologies aim at more sophisticated techniques to describe Web Services, emphasizing the concept of Semantic Web Services. In our understanding, a Semantic Web Service is defined as a “self-contained, self-describing, semantically marked-up software resources that can be published, discovered, composed and executed across the Web in a task driven automatic way” [Arroyo et al., 2004]. By machine-processable descriptions of the relevant information and by means of automated mechanisms that utilize this information, the following functionalities for Web Services shall be achieved.
The aim of WSMO and its surrounding efforts is to define a coherent technology for Semantic Web Services (short: SWS). WSMO defines the modeling elements for describing several aspects of Semantic Web Services. The conceptual basis of WSMO is the Web Service Modeling Framework [WSMF], wherein four main components are defined that are needed for a full coverage framework for Semantic Web Services (see Figure 1). The first component is Ontologies which provide the formal semantics of the information used by all other components. The second component is Goals that specify objectives that a client may have when he consults a web service. The third component is Web Services. For supporting automated discovery, composition, and execution of Web Services, descriptions are required on the functionality provided by a Web Service (called “Capability” in WSMO). For supporting automated choreography and execution compensation of Web Services, particular information on the external visible behavior of a Web Service are needed (called “Interface” in WSMO), including information on the technical accessibility and the actual message exchange of Services. The fourth component of WSMO is Mediators, which are used as connectors between particular components and include possibly required mediation facilities needed to make connected components interoperable. WSMO distinguishes different types of Mediators. The components of WSMO along with exhaustive explanations are presented in the WSMO Primer.
 |
| Figure 1. WSMO Components |
Semantic Web Services can be used in manifold application fields. In accordance with the use cases defined in Web Services Architecture Usage Scenarios by the W3C Web Services Architecture Working Group, we discuss two most common use case scenarios to exemplify the usage of SWS technologies:
For describing the use cases, we slightly modify the methodology of the W3C Use Case descriptions and extend by the requirements arising for Semantic Web Services technologies. The following lists the aspects we use for the use case definitions below.
In Web Services Architecture Usage Scenarios, the travel agency use case is separated into two use cases - one with static discovery and one with automated discovery. With Semantic Web Services we clearly want to support automated discovery, thus we restrict the first WSMO use case to a Virtual Travel Agency scenario that supports automated discovery of Web Services.
Imagine a “Virtual Traveling Agency”, called VTA for short, that is an end user platform providing eTourism services to customers. These services can cover all kind of information services concerned with tourism information - from information about events and sights in an area to services that support booking of flights, hotels, rental cars, etc. online. Such VTAs are already existent, but as this point of time they mostly are an information portal along with some web-based customer services (e.g. eTourism.at). By applying Semantic Web Services, a VTA will invoke Web Services provided by several eTourism suppliers and aggregate them into new customer services. Such VTAs will provide automated eTourism services to end users, thus tremendously enhancing the functionality of currently existing VTAs.
The overview of the use case for VTAs that aggregate Web Services of different tourism service providers looks like this: a customer uses the VTA as the entry point for his request. This request must fit to an end-user service that the VTA provides. These end-user services are aggregated by the VTA by invoking and combining Web Services offered by several tourism service providers. Therefore, there must be some kind of contract between the service providers and the VTA for regulating usage and allowance of the Web Services. Figure 2 shows this overview (modified and extended from W3C Travel Agent Use Case overview).
 |
| Figure 2. Use Case Overview: Virtual Travel Agency based on Semantic Web Services |
The overview described above can be seen as a general structure for VTAs that can be extended to more complex scenarios wherein the customer can be a Web Service itself, thus creating a network of composed services that offer complex tourism services. For example, one VTA can provide flight booking services for an airline union, another VTA aggregates booking service for a worldwide hotel chain, and a third VTA provides booking services for rental cars by combining the services of several worldwide operating car rental agencies. Then, another VTA uses these VTA-services for providing an end-user service for booking complete holiday trips worldwide.
In order to showcase and test the applicability of WSMO and not to get lost in real-world modeling of eTourism use cases, we restrict ourselves to a simple VTA use case from booking international online train tickets. This use case is described in more detail in section 3.1.VTA for International Online Train Ticket
In general use case there are 3 actors. The following defines what they are, why they participate in this use case (goal), and with whom they need to interact in what way (role).
We identify the following usage scenarios
In this use case, the VTA is the central point of interaction between the Customer and Web Services. Regarding the technological requirements, it gets obvious from the Usage Scenario descriptions that (1) the Web Services offered by the Service Providers have to carry sufficient descriptive information to support automated Web Service usage, and (2) that the VTA has to hold all mechanisms to handle Semantic Web Services. The basic architecture of such a VTA as a central entity for Semantic Web Services handling is shown in Figure 3. The essential functionalities of Semantic Web Service enabled VTAs – with special regard to the requirements for Semantic Web Service technologies – are:
 |
| Figure 3. General Architecture of a SWS-enabled VTA |
Summarizing, the VTA is a SWS-enabled B2C application that provides an end-user service following a C/S Model. In order to support coherent functionality of the VTA and ensure that the descriptions of Web Services are compatible to this, an overall framework for SWS technologies is needed. This is provided by WSMO. Section 3.1 exemplifies the modeling of the WSMO components for a real world VTA use case in detail.
The second use case is concerned with teh integration of possible heterogeneous resources in B2B settings which is considered as one of the most important application fields of the Web Service technology.
In the B2B use case, two enterprises called E1 and E2 want electronically exchange
business documents across the network. It is assumed that partners may not know
each other before carrying business transaction and that is why contract negotiation
and contract agreement are essential elements of this use case. It is assumed
that each of the partners expose a set of web services with the given capabilities,
which can handle conversation using any possible known B2B protocols (e.g. ebXML
[ebXML], RosettaNet [RosettaNet] etc.). The contract
agreement defines roles of enterprises in the conversation e.g. one of the enterprise
E1 becomes the seller and the second enterprise E2 becomes the buyer. Agreement
also predefines the order of the message exchange pattern e.g. buyer first sends
purchase order (PO) and after that it receives purchase order acknowledgement
(POA). Differently than in the previous B2C use case, where the client/server
model of interactions has been adopted, the peer-to-peer model is used in this
use case - partners are equal and they carry the conversation. Each of the companies
has own orchestration and the set of web services, which enables to exchange
business documents electronically. Infrastructure provided by SWS takes care
for any necessary mediation between web services (links web services), ontologies
(resolves possible representation mismatches between ontologies used by these
two enterprises), goals (links goals) and web services and goals. SWS infrastructure
supports the execution of the contract to fulfill approved agreement.
 |
| Figure 4. B2B Integration with Semantic Web Services |
In this use case an ultimate goal of an enterprise E1 is to integrate its own back-end system with the back-end system of an enterprise E2. Once integrated, SWS software enables back-end systems of both companies to interact and to preserve the message, process and protocol semantic. The information systems used by enterprises E1 and E2 are autonomous, heterogonous and distributed. Semantic Web Services address each of these three properties and the software based on SWS enables companies to cooperate.
The back-end systems in E1 and E2 are autonomous since each of them changes its state without informing other system about it. SMS software enables to track state changes of back-end applications to facilitate coordination between systems of E1 and E2.
The back-end systems in E1 and E2 are heterogonous, because each of them has different conceptual model for expressing business semantics. SWS software takes care of appropriate mediation of the representation and meaning of the back-end system to the equivalent representation and meaning of the other system. The SWS software ensures to maintain the same semantics between back-end systems of E1 and E2.
The back-end systems in E1 and E2 are distributed because each of them maintains its own state independently from the other system. Back-end applications in companies E1 and E2 do not share data or state at all. SWS software implemented in both companies takes care of transporting data between the systems.
The use case assumes peer-to-peer relationships between two business partners carrying conversation about purchasing/selling of goods. The B2B use case focuses on the technical infrastructure based on the SWS technology, which enable any business company to automatically discover web services which are capable to fulfill its goals, compose simple web services into complex web services to achieve a given goal and to automatically execute given services in a particular order. This use case assumes that there may be no prior business relationships between two enterprises before the discovery. Enterprise E1 must find enterprise E2 and they must agree and enforce the contract in their companies. Agreement should define roles of each of them in the agreed business process – e.g. one of them would become a buyer and one of them would become a seller. The agreement can lead to only one time execution of the agreed business process (e.g. request purchase order) or to long time relationships based on the multiply execution of the agreed contract. Payments are sent through financial institutions and at this stage they are out of the scope of this use case. The same situation concerns the shipment of the goods. This use case consider sending documents as for example purchase orders or invoices, but the physical shipment of goods is out of the scope of this use case.
There are two actors in the B2B use case – actors, which represent two business entities. The size and the importance of companies are not predefined in this use case. They might differ in size but from the perspective of this use case it should not matter which one of them is a more dominant partner. Both of the enterprises undertake a predefined role in the use case. These are:
In this use case the following usage scenarios have been identified:
Web Services Modeling Execution (WSMX) is the infrastructure hosted by each of the enterprises to support services following a peer-to-peer model. WSMX is software implementation of a web service execution environment supporting the development, management and execution of Semantic Web enabled Web Services. WSMX platform does not differentiate between calls coming from the back-end application systems (intra-company information systems) and from the information systems of other enterprises. WSMX can also communicate directly with other WSMX platforms hosted by other enterprises as shown on figure 5.
 |
| Figure 5. B2B Use Case System Architecture |
In the following we model the two use cases described above in WSMO in order to explicate the practical usage and the design of WSMO. We apply the following methodology for this:
For step number 3, the models of the respective WSMO components are described in Listings in this document. Therein, the Listings specify the different WSMO components as conceptual models in accordance to the specifications in WSMO Standard, V0.2. These models have to be transformed for usage within a specific technology, with respect to syntax and technological constructs. For step 4, we provide the models as computational resources for download, testing and development in the Appendices. Currently, we support FLORA-2 as a F-Logic reasoner (see FLORA-2 homepage). The models provided for download in the Appendices are runnable as separate FLORA-2-programms; the structure and the connections of these programs as WSMO-components have to be defined by the application developer (e.g. transforming the "usedMediators"-constructs as well as namespaces in the respective technology support by the tool). Besides, the identifier of every WSMO element is an URI. In the use case, the URL "http://www.wsmo.org/2004/d3/d3.2/v0.1/20040419/resources/FILE" is the general ID for all models of the use case, wherein FILE is replaced by the actual file name of the component as linked to in the Appendices. The elements of specific components have then the identifier "..../resources/FILE#element", which make the identifier-concept of WSMO conformant to the concept of URIs in web technologies. With regard to readbility of the listings, we only specify the file name as the component identifier in the Listings.
For modeling of the WSMO components in the Listings in this document we use WSML-U as a human readable syntax, based on the syntax for F-Logic as defined in [Kifer et al., 1995].
According to the general VTA use case described in Section 2.1 B2C - Virtual Travel Agency we define the following use case here:
The course of the use case is the following:
- the customer poses a request for an international train connection from Innsbruck
to Frankfurt on 23rd May 2004, at 16.00 local time
- the VTA returns a set of possible connections
- the user selects one of these connections and poses a request for booking
the ticket online
- the VTA combines the online train ticket booking services from ÖBB and
DB, executes the booking and payment process, and sends the online ticket per
email to the Customer.
For the aggregated service, the VTA has to determine the itineraries of the international connections, and to split them at the border stations into national itineraries. The VTA has to mediate between the following web services:
The rationale for choosing this use case is that it showcases a possible VTA use case as described above within all the components identified in WSMO. The components are simple, thus this use case allows showing the modeling of WSMO elements without getting lost in complicated definitions of specific elements.
 |
| Figure 6. ÖBB Train Connection Itinerary Service |
The following lists the requirements analysis for modeling the use case. For each of the components of WSMO, a list of requirements is defined that are needed in order to enable the requested functionality of the VTA for online search and booking of international train tickets, regarding the use case described above.
| O1 | We need ontological information on international train itineraries and on the purchasing process. This information should be kept in separated ontologies, following the design principle of modular ontology design. |
| O2 | An itinerary is described by a Start- and End-Location, date and time of departure and arrival, the station where the border is crossed, and the fare. |
| O3 | There has to be customer that buys a train ticket |
| O4 | An itinerary describes a valid international train connection. |
| O5 | There exists a concept that defines whether a location is located at the border between 2 countries |
| O6 | A ticket is valid for exactly 1 itinerary |
| O7 | A ticket is valid for exactly 1 customer |
| O8 | A location description consists of a Location identifier, a Country identifier, and an indicator that the location has a train station |
| O9 | The purchase ontology has to identify the buyer and seller roles, a product with a price, and valid payment methods |
| O10 | The only valid payment method for online tickets is credit card payment |
| O11 | Information on Date and Time should be defined generally in a separate ontology |
| G1 | Booking a Online Train Ticket |
| G1.1 | From Innsbruck to Frankfurt |
| G1.2 | Start time: 17th May 2004, at 16.00 local time |
| W1 | Each National Train Operator provides a Web Service that offes an international train connection timetable and national online ticket booking |
| W2.1 | The international train connection timetable takes a start location and an end location and a departure date and returns a set of itineraries |
| W2.1.1 | Exceptions are: - Service not available - Start or End Location does not exists |
| W2.2 | The national online ticketing service takes an itinerary with start location and end location in the national country, a credit card number and returns a ticket for this itinerary. |
| W2.2.1 | Exceptions are: - Service not available - Credit card not accepted |
| M1 | There exists a WG Mediator to link Goal G1 to a train connection timetable Web Service. |
| M2 | An OO-Mediator has to integrate the two domain ontologies |
| M2.2 | OO-Mediator: All WSMO components apply the integrated ontology |
| M3 | a WW-Mediator mediates between the national online ticketing services. |
The following provides the modeling of the use case in WSMO with respect to the requirements determined above. The models are presented in the same structure as in the requirements analysis. As explained above, the Listings below specify conceptual models of the distinct WSMO components which have to be transformed into the technology provided by the tool used for reasoning on the ontologies. The full WSMO models as downloadable computational resources for this use case are provided in Appendix A. The models apply the syntax defined in section 3.
The use case modeling in this document relies on the Web Service Modeling Ontology WSMO, Version 0.3. This version of WSMO does not provide elaborated description elements for all WSMO components, especially for Web Services a sophisticated specification of the WSMO modeling primitives only exists for Web Service Capabilities. As we can only apply such elaborated description structures for use case modeling, we restrict this version of the use case modeling to the WSMO description elements defined.
We define 3 domain ontologies that provide the terminology definitions for the use case. The first ontology "International Train Ticket" describes the domain of train tickets, the second ontology "Date and Time" defines a general model for specifying time and dates and relationships of them, and the third ontology "Purchase" describes generic elements of purchasing a product between a buyer and a seller.
We apply the following conventions in the Listings for ontology specifications (those which are not ontology-specific hold for all other WSMO component specifications as well):
The "International Train Ticket" Ontology defines an itinerary and the surrounding concepts as defined in Listing 1. Additionally, a axiom is defined that checks the validity of the traveling dates (the start date / time has to be in the future and the arrival date / time as to be later than the starting date / time) as well as some instances needed in the further use case modeling. Listing 1 shows the F-Logic specification of the ontology.
Ontology
International Train Connections Domain Ontology
non-functional Properties
title
International Train Connections Domain Ontology
creator
DERI International
subject
International, Train Itineraries, Ticket Booking
description
International Train Itineraries for Online Ticket Booking
publisher
DERI International
contributor
Michael Stollberg, Rubén Lara, Holger Lausen, Axel Polleres
date
20040419
type
domain ontology
format
text
identifier
http://www.wsmo.org/2004/d3/d3.2/v0.1/20040517/resources/tc.flr
source
http://www.wsmo.org/2004/d3/d3.2/v0.1/20040517/resources/tc.flr
language
English
relation
http://www.wsmo.org/2004/d3/d3.2/v0.1/20040419/resources/dt.flr
coverage
Europe
rights
DERI
version
1.3
usedMediators
// OO Mediator for using the Date and Time Ontology
VTA-OOM-trainConnection.wsml
CONCEPTS
// a 'location' is the general notion of all locations in a country
location
name oftype string
locatedIn oftype set country
/* Axioms for location */
// 'locatedIn' is a transitive property. this allows to specify only the city for a desired train connection
X ofclass location[locatedIn ofvalues Y] :-
X ofclass location[locatedIn ofvalues Z] and
Z ofclass location[locatedIn ofvalues Y].
// Integrity Constraint
invalid(X) :- X ofclass location[locatedIn ofvalues X].
//subclasses of location
station subclassof location
country subclassof location
city subclassof location
village subclassof location
// 'borderStation' defines a train station at the border
// 'borderToCounty' denotes all adjacent countries
borderStation subclassof station
borderToCounty oftype set (country)
itinerary
startLocation oftype station
endLocation oftype station
via oftype set station
departure oftype dateAndTime
arrival oftype dateAndTime
// Integrity Constraint: departure has to be before arrival
invalid(I) :-
I ofclass itinerary[
departure ofvalue X2 and
arrival ofvalue X3]
and after(X2,X3).
traveller
name oftype string
// 'ticket' defines a ticket that relates an itinerary to a customer
ticket subclassof product
itinerary oftype itinerary
traveller oftype traveller
// below, some instances are defined that are needed throughout the use case modeling
germany subclassof country
name ofvalue 'Germany'
austria subclassof country
name ofvalue 'Austria'
innsbruckHbf subclassof station
name ofvalue'Innsbruck Hbf'
locatedIn ofvalues innsbruck
innsbruck subclassof city
name ofvalue 'Innsbruck'
locatedIn ofvalues austria
boesby subclassof village
name ofvalue 'Boesby'
locatedIn ofvalues germany
frankfurtHbf subclassof station
name ofvalue 'Frankfurt Hbf'
locatedIn ofvalues germany
kufsteinHbf subclassof borderStation
name ofvalue 'Kufstein Hbf'
locatedIn ofvalues austria
borderToCounty ofvalues germany
salzburgHbf subclassof borderStation
name ofvalue 'Salzburg Hbf'
locatedIn ofvalues austria
borderToCounty ofvalues germany
freilassingBf subclassof borderStation
name ofvalue 'Freilassing Bf'
locatedIn ofvalues germany
borderToCounty ofvalues austria
kiefersfeldenBf subclassof borderStation
name ofvalue 'Kiefersfelden Bf'
locatedIn ofvalues germany
borderToCounty ofvalues austria
|
The "Date and Time" Ontology in Listing 2 defines models for dates (i.e. certain days) and time (i.e. definition of certain points in time). Further, it defines axioms that represent conventional aspects of date and time, like ´before´ and ´after´, etc. In the use case, this is needed to determine validity of train connections, e.g for ensuring that a ticket is not for an itinerary that is in the past. It also can be used generally for expressing dates and time and relationships between them. Listing 2 only displays the ontology schema and the algebra for date and time, while the downloadable file contains instances and queries for testing.
The main ontology taken into consideration for developing this representation in F-Logic is an entry sub-ontology of time, available at http://www.isi.edu/~pan/damltime/time-entry.owl. This ontology uses abstract temporal concepts like instant, interval and event and uses the Gregorian calendar as representation (partly using own encoding and partly using XSD encoding). Axioms are defined in first order logic in the accompanying paper [Pan and Hobbs]; there also is a LISP version of these axioms available at http://www.cs.rochester.edu/~ferguson/daml/daml-time-20030728.lisp.Other ontologies like COBRA calenderclock ontology (http://daml.umbc.edu/ontologies/cobra/0.4/calendarclock) are only a straight forward representation of the Gregorian calendar, without any abstraction of concepts and description of axioms. Widely used concrete representations for date and time are defined in ISO 8601 (Numeric representation of Dates and Time) and in the XML Schema Definition (http://www.w3.org/TR/xmlschema-2/), which is based on ISO 8601. The ontology defined in Listing 2 uses the Gregorian calendar with the representation based on the definition in http://www.w3.org/TR/xmlschema-2/.
Ontology
Date and Time Ontology
non-functional properties
title
Date and Time Ontology
creator
DERI International
subject
Date, Time, Date and Time Algebra
description
generic representation and alegbra for date and time
publisher
DERI International
contributor
Holger Lausen, Axel Polleres, Rubén Lara,
date
20040517
type
domain ontology
format
text
identifier
http://www.wsmo.org/2004/d3/d3.2/v0.1/20040517/resources/dt.flr
source
http://www.wsmo.org/2004/d3/d3.2/v0.1/20040517/resources/dt.flr
language
English
relation
http://www.isi.edu/~pan/damltime/time-entry.owl,
http://daml.umbc.edu/ontologies/cobra/0.4/calendarclock,
´ http://www.w3.org/TR/xmlschema-2/
coverage
general
rights
DERI
version
1.15
Concepts and Integrity Constraints
// An instant represents a particular point in time
instant
// An interval represents a duration between 2 points in time
interval
start oftype instant
end oftype instant
/* Axioms for Interval */
//computes if a interval X contains a second interval Y
contains(X and Y) :-
X ofclass interval and Y ofclass interval and
(before(X.start and Y.start) or equal(X.start and Y.start)) and
(after(X.end and Y.end) or equal(X.end and Y.end)).
//computes if a interval X contains a instant Y
contains(X and Y) :-
X ofclass interval and Y ofclass instant and
(before(X.start and Y) or equal(X.start and Y)) and
(after(X.end and Y) or equal(X.end and Y)).
// class date and its representation according to the Gregorian calendar"
date[
dayOfMonth oftype dayOfMonth
monthOfYear oftype monthOfYear
year oftype year
// Integrity Constraints for date
invalid(X) :-
X ofclass date and
(invalid(X.dayOfMonth) or invlaid(X.monthOfYear) or invalid(X.year)).
// day of a month is represented by an integer
dayOfMonth subclassof integer.
//integrity constraint for valid dayOfMonths:
invalid(X) :-
X ofclass dayOfMonth and
(X < 0 or X > 31).
//a year is represented by an integer
year subclassof integer.
//a monthOfYear is represented by an integer and has additional properties
monthOfYear subclassof integer
name oftype string
daysAfterBeginOfYear oftype integer
//integrity constraint for valid monthOfYear:
invalid(X):-
X ofclass monthOfYear and
(X < 0 or X > 12).
// concrete month are defined as instances
1 ofclass monthOfYear
daysAfterBeginOfYear ofvalue 31
name ofvalue 'January'
2 ofclass monthOfYear
daysAfterBeginOfYear ofvalue 59
name ofvalue 'February'
3 ofclass monthOfYear
daysAfterBeginOfYear ofvalue 90
name ofvalue 'March'
4 ofclass monthOfYear
daysAfterBeginOfYear ofvalue 120
name ofvalue 'April'
5 ofclass monthOfYear
daysAfterBeginOfYear ofvalue 151
name ofvalue 'May'
6 ofclass monthOfYear
daysAfterBeginOfYear ofvalue 181
name ofvalue 'June'
7 ofclass monthOfYear
daysAfterBeginOfYear ofvalue 212
name ofvalue 'July'
8 ofclass monthOfYear
daysAfterBeginOfYear ofvalue 243
name ofvalue 'August'
9 ofclass monthOfYear
daysAfterBeginOfYear ofvalue 273
name ofvalue 'September'
10 ofclass monthOfYear
daysAfterBeginOfYear ofvalue 304
name ofvalue 'October'
11 ofclass monthOfYear
daysAfterBeginOfYear ofvalue 334
name ofvalue 'November'
12 ofclass monthOfYear
daysAfterBeginOfYear ofvalue 365
name ofvalue 'December'
// class time
time
hourOfDay oftype hourOfDay
minuteOfHour oftype minuteOfHour
secondOfMinute oftype secondOfMinute
// integrity constraint for valid time:
invalid(X) :-
X ofclass time and
(invalid(X.hourOfDay) or invalid(X.minuteOfHour) or invalid(X.secondOfMinute)).
// a secondOfMinute is represented by an integer
secondOfMinute subclassof integer.
// integrity constraint for valid secondOfMinute:
invalid(X) :-
X ofclass secondOfMinute and
(X < 0 or X > 59).
// a minuteOfHour is represented by an integer
minuteOfHour subclassof integer.
// integrity constraint for valid minuteOfHour:
invalid(X) :-
X ofclass minuteOfHour and
(X < 0 or X > 59).
// a hourOfDay is represented by an integer
hourOfDay subclassof integer.
// integrity constraint for valid hourOfDay:
invalid(X) :-
X ofclass hourOfDay and
(X < 0 or X > 23).
// class date and time and representing together a specific point of time (instant)
dateAndTime subclassof instant
date oftype date
time oftype time
//integrity constraint for valid dateAndTimes:
invalid(X) :-
X ofclass dateAndTime and
(invalid(X.date) or invalid(X.time)).
/***************************
algebra for date and time
***************************/
// computes equality of a date
equal(X and Y) :-
Y ofclass date and X ofclass date and
X.dayOfMonth = Y.dayOfMonth and
X.monthOfYear = Y.monthOfYear and
X.year = Y.year.
// computes if a given date X is before another date Y
before(X and Y) :-
Y ofclass date and X ofclass date and
((X.dayOfMonth Y.dayOfMonth and X.monthOfYear = Y.monthOfYear and X.year = Y.year) or
(X.monthOfYear < Y.monthOfYear and X.year = Y.year) or
(X.year < Y.year)).
// computes if a given date X is after another date Y
after(X and Y) :-
Y ofclass date and X ofclass date and
((X.dayOfMonth > Y.dayOfMonth and X.monthOfYear = Y.monthOfYear and X.year = Y.year) or
(X.monthOfYear > Y.monthOfYear and X.year = Y.year) or
(X.year > Y.year)).
/* this is simplified and ignores leap years a proper algorithm as e.g. found at
http://quasar.as.utexas.edu/BillInfo/JulianDatesG.html calulates correctly for Gregorian Calendar dates
(after 1582 - at keast for most countries, see http://members.brabant.chello.nl/~h.reints/cal/whenjul2greg.htm
for an exact dates per country (e.g. Yugoslavia changed 1919)) can be represented in f-logic as follows: */
julianDayNumber(X and JDN) :-
X ofclass date and
X.monthOfYear < 3 and
Y is X.year - 1 and
M is X.monthOfYear + 12 and
D is X.dayOfMonth and
A is truncate(Y and 100) and
B is truncate(A and 4) and
C is 2 - A + B and
E is floor(365.25 * (Y + 4716)) and
F is floor(30.6001 * (M + 1)) and
JDN is C + D + E + F - 1524.
julianDayNumber(X and JDN) :-
X ofclass date and
X.monthOfYear > 2 and
Y is X.year and
M is X.monthOfYear and
D is X.dayOfMonth and
A is truncate(Y and 100) and
B is truncate(A and 4) and
C is 2 - A + B and
E is floor(365.25 * (Y + 4716)) and
F is floor(30.6001 * (M + 1)) and
JDN is C + D + E + F - 1524.
/* however due to a bug in XSB (float number arithemtics do not work proper) this could not be implemented
and a simplified version that aproximates the days after Christ is used. */
daysAfterChrist(D and X) :-
D ofclass date and Z is D.monthOfYear and
X is (D.dayOfMonth + Z.daysAfterBeginOfYear + (D.year*365)).
// the difference in days between 2 dates
daysBetween(D1 and D2 and X) :-
D1 ofclass date and D2 ofclass date and
daysAfterChrist(D1 and DAC_D1) and
daysAfterChrist(D2 and DAC_D2) and
X is DAC_D1 - DAC_D2.
// computes if two given times are the same
equal(X and Y) :-
X ofclass time and Y ofclass time and
X.secondOfMinute = Y.secondOfMinute and
X.minuteOfHour = Y.minuteOfHour and
X.hourOfDay = Y.hourOfDay.
// computes if a given time X is before another time Y
before(X and Y) :-
X ofclass time and Y ofclass time and
((X.secondOfMinute < Y.secondOfMinute and X.minuteOfHour = Y.minuteOfHour and X.hourOfDay = Y.hourOfDay) or
(X.minuteOfHour < Y.minuteOfHour and X.hourOfDay = Y.hourOfDay) or
(X.hourOfDay < Y.hourOfDay)).
// computes if a given time X is after another time Y
after(X and Y) :-
X ofclass time and Y ofclass time and
((X.secondOfMinute > Y.secondOfMinute and X.minuteOfhour = Y.minuteOfhour and X.hourOfDay = Y.hourOfDay) or
(X.minuteOfhour > Y.minuteOfhour and X.hourOfDay = Y.hourOfDay) or
(X.hourOfDay > Y.hourOfDay)).
// computes the amount of seconds from midnight
secondsFromMidnight(T and X) :-
T ofclass time and
X is T.secondOfMinute + (T.minuteOfHour*60) + (T.hourOfDay*60*60).
// the difference in seconds between 2 times
secondsBetween(T1 and T2 and X) :-
T1 ofclass time and T2 ofclass time and
secondsFromMidnight(T1 and SFM_T1) and
secondsFromMidnight(T2 and SFM_T2) and
X is SFM_T1 - SFM_T2.
// computes if Date and Time are equal
equal(X and Y) :-
X ofclass dateAndTime and Y ofclass dateAndTime and
equal(X.date and Y.date) and
equal(X.time and Y.time).
// computes if a given date and time X is before another date and time Y
before(X and Y) :-
X ofclass dateAndTime and Y ofclass dateAndTime and
((equal(X.date and Y.date) and before(X.time and Y.time)) or
before(X.date and Y.date)).
// computes if a given date and time X is after another date and time Y
after(X and Y) :-
X ofclass dateAndTime and Y ofclass dateAndTime and
((equal(X.date and Y.date) and after(X.time and Y.time)) or
after(X.date and Y.date)).
// computes the difference in seconds between two different DateAndTime
secondsBetween(D1 and D2 and X) :-
D1 ofclass dateAndTime and D2 ofclass dateAndTime and
daysAfterChrist(D1.date and DAC_D1) and
daysAfterChrist(D2.date and DAC_D2) and
secondsFromMidnight(D1.time and SFM_T1) and
secondsFromMidnight(D2.time and SFM_T2) and
X is SFM_T1 + DAC_D1 * 24 * 60 * 60 -
(SFM_T2 + DAC_D2 * 24 * 60 * 60).
// the difference in (decimal) days between two different DateAndTime
daysBetween(D1 and D2 and X) :-
D1 ofclass dateAndTime and D2 ofclass dateAndTime and
secondsBetween(D1 andD2 and Z) and
X is Z/60/60/24.
//The current Date is defined here since we do not have build in function yet
currentDate ofclass dateAndTime
date ofvalue CurrentDate ofclass date
dayOfMonth ofvalue 28 and
monthOfYear ofvalue 2 and
year ofvalue 2001 and
time ofvalue CurrentTime ofclass time
hourOfDay ofvalue 23 and
minuteOfHour ofvalue 40 and
secondOfMinute ofvalue 12
|
The "Purchase" ontology defines general concepts for purchasing a product (there is a buyer, a seller, a product with a price, a payment method, and delivery).
Ontology
Purchase Domain Ontology
non-functional properties
title
Purchase Domain Ontology
creator
DERI International
subject
buyer, seller, product, price, payment method, delivery
description
general purchase ontology
publisher
DERI International
contributor
Michael Stollberg, Axel Polleres, Rubén Lara, Holger Lausen
date
20040517
type
domain ontology
format
text
identifier
http://www.wsmo.org/2004/d3/d3.2/v0.1/20040517/resources/po.flr
source
http://www.wsmo.org/2004/d3/d3.2/v0.1/20040517/resources/po.flr
language
English
relation
coverage
general
rights
DERI
version
1.7
CONCEPTS
address
name oftype string
street oftype string
number oftype integer
zipcode oftype string
city oftype city
state oftype state
country oftype country
country
name oftype string
location
name oftype string,
country oftype country
state
name oftype string
partOfCountry oftype country
buyer
shipTo oftype address
billTo oftype address
purchaseIntention oftype set (tradeItem)
hasPayment oftype set (paymentMethod)
seller
address oftype address
saleIntention oftype set (tradeItem)
tradeItem
product oftype product
price oftype price // Note that the item price given by the buyer denotes a price *limit*
product
name oftype string
price
amount oftype real
currency oftype string
paymentMethod
name oftype string
// A trade is an actual agreement on trading items between two partners.
trade
items ofclass set (tradeItem)
buyer oftype buyer
seller oftype seller
payment oftype paymentMethod
// Delivery of a good as an effect of a purchase
delivery
products oftype set (product)
receiver oftype buyer
sender oftype seller
AXIOMS
/* purchasing only possible if
1) Only products can be traded/delivered if the Buyer requires them (purchaseIntention),
and if the seller provides them (saleIntention), and if the selling price is not higher than the price intended by the buyer: */
invalid(T) :- T ofclass trade[items ofvalues I] and tnot T.seller[saleIntention ofvalues I].
priceok(I) :- T ofclass trade[items ofvalues I] and T.buyer[purchaseIntention ofvalues I1] and
I.product = I1.product and
I.price.currency = I1.price.currency and
I.price.amount =< I1.price.amount.
/* 2. Only products can be delivered if
the receiver requires them (purchaseIntention) and if the sender provides them (saleIntention) */
invalid(D) :- D ofclass delivery[products ofvalues P] and
tnot D.receiver.purchaseIntention[product ofvalue P].
invalid(D) :- D ofclass delivery[products ofvalues P] and
D.receiver.saleIntention[product ofvalue P].
INSTANCES (needed throughout the use case modeling)
creditcard subclassof paymentMethod
creditcard
name oftype string
number oftype string
expmonth oftype month
expyear oftype year
type oftype string
cash ofclass paymentMethod
currency oftype currency
currency
// some sample currencies
euro ofclass currency
usd ofclass currency
gbp ofclass currency
|
Goals denote what a user wants to receive when using a Web Service. More precisely, a Goal describes constraints of an object that the user desires. A Goal is modeled as a fact, i.e. it specifies the ontological structure of the object of desire without variables. Goals in WSMO are understood as "instantiated Goals", i.e. a concrete desire to be resolved. For WSMO-based applications, separation of Goal Schemas and Goal Instances seems to be an appropriate solution for handling of Goals. The former shall describe the general structure of a Goal, while a Goal Instances instantiate a Goal Schema by specifying certain attributes for the which is instantiated by a user for expressing a certain desire. Goal Schemas can also be thought of as pre-defined Goals, while Goal Instances are concrete requests for Web Services during runtime.
In the use case, we have one Goal: a user wants to buy a ticket online for a train connection from Innsbruck to Frankfurt on a certain date. The Goal Template states that the desire is to get a train ticket for an itinerary and for a customer, according to the knowledge defined in the ontologies. The Goal Instance specifies concrete values for the template structure. Listing 4 shows this Goal with the following elements:
Goal
buying online train ticket
non-functional Properties
title
buying online train ticket
creator
DERI International
subject
Train Tickets, Online Ticket Booking
description
a desire for booking an international train ticket online
publisher
DERI International
contributor
Michael Stollberg, Rubén Lara, Holger Lausen, Axel Polleres
date
20040517
type
WSMO Goal
format
text
identifier
http://www.wsmo.org/2004/d3/d3.2/v0.1/20040517/resources/goal.flr
source
http://www.wsmo.org/2004/d3/d3.2/v0.1/20040517/resources/goal.flr
language
English
relation
http://www.wsmo.org/2004/d3/d3.2/v0.1/20040517/resources/tc.flr
http://www.wsmo.org/2004/d3/d3.2/v0.1/20040419/resources/dt.flr
http://www.wsmo.org/2004/d3/d3.2/v0.1/20040419/resources/po.flr
coverage
Europe
rights
DERI
version
1.14
usedMediators
// OO Mediator for using the ontologies
VTA-OOM-Goal1.wsml
// define an instance of Goal
mygoal ofclass goal.
// Goal Postcondition:
// a ticket for an itinerary from Innsbruck to Frankfurt
mygoal[postcondition ofvalue
myTicket ofclass ticket[
itinerary ofvalue exists itinerary ofclass itinerary[
startLocation ofvalue innsbruckHbf
endLocation ofvalue frankfurtHbf
departure ofvalue exists dateandtime ofclass dateAndTime[
date ofvalue exists date ofclass date[
dayOfMonth ofvalue 17
monthOfYear ofvalue 5
year ofvalue 2004
],
time.hourOfDay ofvalue 18 ofclass hourOfDay
]
],
traveller ofvalue exists traveller ofclass traveller[
name ofvalue 'Dieter Fensel'
]
]
].
// Goal Effect: a trade for the ticket, payed by creditcard
mygoal[effect ofvalue
mytrade ofclass trade[
items ofvalues myTicket,
buyer ofvalue exists buyer ofclass buyer[
shipTo ofvalue myAddress
billTo ofvalue myAddress
],
payment ofvalue myCreditCard ofclass creditCard[
name ofvalue 'Dieter Fensel'
number ofvalue 1234567890
expMonth ofvalue 9
expYear ofvalue 2006
type ofvalue 'MasterCard'
]
],
myAddress ofclass address[
name ofvalue 'Dieter Fensel'
street ofvalue 'Technikerstrasse'
number ofvalue 13
zipcode ofvalue 6020
city ofvalue innsbruck
state ofvalue tirol
country ofvalue austria
]
].
|
For our use case we define one Web Service: an (imaginary) online train ticket booking services for international train itineraries, offered by the Austrian national train operator ÖBB. Of course, this Web Service can be split up into several Web Services wherefore technologies for composition and Orchestration would be needed. But as a our intention within the current version of this use case modeling is to test and showcase the basic modeling of Web Services, we restrict ourselves to only one Web Services at this point in time.
As the referenced version of WSMO does only provide specifications for the description elements for Web Service Capability modeling, we restrict the Web Services models to the Capabilities at this point in time. A Web Service Capability in WSMO is described by pre- and postconditions, assumptions and effects. The primary information for suitability of a Web Service for satisfying a given Goal is the postcondition (the Web Service postcondition has to logically satisfy the Goal postcondition, which is the core of the discovery mechanisms). The other description elements are secondary information for determining suitability, i.e. filtering the set of Web Services that potentially match the Goal. More detailed discussion of the Discovery mechanism of WSMO Goals and Capabilities is provided in section 3.1.3.
Listing 5 shows the Capability specification for the Web Service, including the following WSMO description elements for Web Service Capabilities. Each notion is modeled as a rule that says when there is an instance that fulfills its body than the specific notion is satisfied:
Web Service Capability
selling online train tickets for Austria and Germany
non-functional Properties
title
selling online train tickets for Austria and Germany
creator
DERI International
subject
Train Tickets, Online Ticket Booking
description
a Web Service Capability for selling international train tickets online
publisher
DERI International
contributor
Michael Stollberg, Rubén Lara, Holger Lausen, Axel Polleres
date
20040517
type
WSMO Web Service Capability
format
text
identifier
http://www.wsmo.org/2004/d3/d3.2/v0.1/20040517/resources/capability.flr
source
http://www.wsmo.org/2004/d3/d3.2/v0.1/20040517/resources/capability.flr
language
English
relation
http://www.wsmo.org/2004/d3/d3.2/v0.1/20040517/resources/tc.flr
http://www.wsmo.org/2004/d3/d3.2/v0.1/20040419/resources/dt.flr
http://www.wsmo.org/2004/d3/d3.2/v0.1/20040419/resources/po.flr
coverage
Europe
rights
DERI
version
1.3
usedMediators
// OO Mediator for using the Date and Time Ontology
VTA-OOM-WS1Cap.wsml
// define an instance of Web Service Capability
oebbCap ofclass capability.
Precondition
// the input has to be an itinerary wherefore
// the start- and endlocation have to be in Austria or in Germany and
// the departure date has to be later than the current Date
oebbCap[precondition] :-
X ofclass itinerary[
startLocation ofvalue StartLoc
endLocation ofvalue EndLoc
departure ofvalue Departure
] and
(StartLoc.locatedIn = austria or StartLoc.locatedIn = germany) and
(EndLoc.locatedIn = austria or EndLoc.locatedIn = germany) and
after(Departure,currentDate).
Assumption
// there needs to be a buyer that accepts payment by cerditcard
// and this creditcard has to be valid (not expired).
oebbCap[assumption] :-
X ofclass buyer[
acceptsPayment =>> Payment
] and
Payment ofclass creditCard and
(currentDate.date.year < Payment.expYear) or
((currentDate.date.year = Payment.expYear) and
((currentDate.date.monthOfyear < Payment.expMonth) or (currentDate.date.monthOfyear = Payment.expMonth))).
Postcondition
// the output of the service is a ticket (train ticket)for an itinerary wherefore
// the start- and endlocation have to be in Austria or in Germany and
// the departure date has to be later than the current Date
oebbCap[postcondition] :-
X ofclass ticket[
itinerary ofvalue_Itinerary ofclass itinerary[
startLocation ofvalue StartLoc
endLocation ofvalue EndLoc
departure ofvalue Departure
]
] and
(StartLoc.locatedIn = austria or StartLoc.locatedIn = germany) and
(EndLoc.locatedIn = austria or EndLoc.locatedIn = germany) and
after(Departure,currentDate).
Effect
// there shall be a trade for the train ticket of the postcondition
oebbCap[effect] :-
Y ofclass trade[
items ofvalues Ticket ofclass ticket[
itinerary ofvalue_Itinerary ofclass itinerary[
startLocation ofvalue StartLoc
endLocation ofvalue EndLoc
]
] and
payment ofvalue Payment
] and
Payment ofclass creditCard and
(StartLoc.locatedIn = austria or StartLoc.locatedIn = germany) and
(EndLoc.locatedIn = austria or EndLoc.locatedIn = germany).
|
Regarding the requirement analysis for our use case, we have identified the need the following types of Mediators:
In the following we model the concrete mediators needed for the use case. These Mediators are applied by the different components described before using the “usedMediators” modeling element.
OO-Mediators
OO Mediators "connect" the ontology/ies with the component that is using it. According to the WSMF/O-principle of strong de-coupling, all issues related to integrate (meaning to resolve all namespace and access issues) and mediate (meaning to resolve heterogeneities) the ontologies to be used are located in the OO Mediator. The user who uses a OO Mediator with the "usedMeditaors"-tag should therefore not care about which ontology he refers to - he just gets his "Information Space" by the OO Mediator and uses it.
In order to import the ontologies into all our components, we need three OO Mediators:
Listings 6-8 specify these three OO Mediators. The general structure of an OO Mediator is that the sourceComponent is the imported ontologies (can also be other OO Mediators), and the targetComponent is the user of the imported ontologies. The internal structure and functionality of a OO Mediator is not stable specified yet, thus we only specify the connection facility in the Listings.
OO Mediator
Train Connection Ontology uses Date and Time Ontology
non-functional Properties
title
Train Connection Ontology uses Date and Time Ontology
creator
DERI International
subject
description
importing the Date and Time Ontology into the Train Connection Ontology
publisher
DERI International
contributor
Michael Stollberg
date
20040430
type
WSMO OO Mediator
format
text
identifier
http://www.wsmo.org/2004/d3/d3.2/v0.1/200400517/resources/VTA-OOM-trainConnection.wsml
source
http://www.wsmo.org/2004/d3/d3.2/v0.1/200400517/resources/VTA-OOM-trainConnection.wsml
language
English
relation
http://www.wsmo.org/2004/d3/d3.2/v0.1/20040517/resources/tc.flr
http://www.wsmo.org/2004/d3/d3.2/v0.1/20040419/resources/dt.flr
coverage
rights
DERI
version
1.1
sourceComponent
// identifier of "Date an Time Ontology"
http://www.wsmo.org/2004/d3/d3.2/v0.1/200400517/resources/dt.flr
http://www.wsmo.org/2004/d3/d3.2/v0.1/200400517/resources/po.flr
targetComponent
// identifier of the "Train Connection Ontology"
http://www.wsmo.org/2004/d3/d3.2/v0.1/200400517/resources/tc.flr
mediationService
// not needed here
|
OO Mediator
OO Mediator for Goal
non-functional Properties
title
OO Mediator for Goal
creator
DERI International
subject
description
importing all ontologies in to the Goal
publisher
DERI International
contributor
Michael Stollberg
date
20040430
type
WSMO OO Mediator
format
text
identifier
http://www.wsmo.org/2004/d3/d3.2/v0.1/200400517/resources/VTA-OOM-Goal1.wsml
source
http://www.wsmo.org/2004/d3/d3.2/v0.1/200400517/resources/VTA-OOM-Goal1.wsml
language
English
relation
http://www.wsmo.org/2004/d3/d3.2/v0.1/20040517/resources/tc.flr
http://www.wsmo.org/2004/d3/d3.2/v0.1/20040419/resources/po.flr
coverage
rights
DERI
version
1.1
sourceComponent
// identifiers of "Train Connection Ontology" and "Purchase Ontology"
http://www.wsmo.org/2004/d3/d3.2/v0.1/200400517/resources/tc.flr
http://www.wsmo.org/2004/d3/d3.2/v0.1/200400517/resources/po.flr
targetComponent
// identifier of the Goal
http://www.wsmo.org/2004/d3/d3.2/v0.1/200400517/resources/goal.flr
mediationService
// not needed here
|
OO Mediator
OO Mediator for Web Service Capability
non-functional Properties
title
OO Mediator for Web Service Capability
creator
DERI International
subject
description
importing all ontologies in to the Web Service Capability
publisher
DERI International
contributor
Michael Stollberg
date
20040430
type
WSMO OO Mediator
format
text
identifier
http://www.wsmo.org/2004/d3/d3.2/v0.1/200400517/resources/VTA-OOM-WS1Cap.wsml
source
http://www.wsmo.org/2004/d3/d3.2/v0.1/200400517/resources/VTA-OOM-WS1Cap.wsml
language
English
relation
http://www.wsmo.org/2004/d3/d3.2/v0.1/20040517/resources/tc.flr
http://www.wsmo.org/2004/d3/d3.2/v0.1/20040419/resources/po.flr
coverage
rights
DERI
version
1.1
sourceComponent
comment: identifiers of "Train Connection Ontology" and "Purchase Ontology"
http://www.wsmo.org/2004/d3/d3.2/v0.1/200400517/resources/tc.flr
http://www.wsmo.org/2004/d3/d3.2/v0.1/200400517/resources/po.flr
targetComponent
comment: identifier of the Web Service Capability
http://www.wsmo.org/2004/d3/d3.2/v0.1/200400517/resources/capability.flr
mediationService
comment: not needed here
|
WG-Mediators
A WG Mediator denotes the differences between a Goal and a Web Service Capability
in order to make them matching by restricting the range of valid information
to be exchanged between the Goal Owner and the Web Service. The difference is
stated in a reduction. In fact, the reduction defines the intersection between
the information space of the Goal and the information space of Capability. Thereby,
only such information will be interchanged between the Goal and the Web Service
that are valid, meaning that with all outputs, postconditions, and effects of
the Web Service the Goal is satisfiable.
In our use case, we only need 1 WG Mediator for connecting the Goal and the Web Service, specified in Listing 9. The WG Mediator, as all other WSMO components, has to be aware of the ontologies used in the components to be connected. As the source component is the Goal, and the target component is the Web Service with the Capability, the the OO Mediators defined for the Goal and for the Web Service inplicitely. The "usedMediators" tag specifies usage of additional ontologies that are needed for specifying the reduction. The Reduction in the WG Mediator restricts the set of values for the Goal and the Web Service to those for which the usage of the Web Service is valid for satisfying the Goal. More precisely, the Reduction is the intersection of valid knowledge items of the Goal and of valid items of the Web Service.
WG Mediator
WG Mediator
non-functional Properties
title
Web Service - Goal connection Capability uses all ontologies
creator
DERI International
subject
description
connecting the Goal and the Web Service Capability
publisher
DERI International
contributor
Michael Stollberg
date
20040430
type
WSMO OO Mediator
format
text
identifier
http://www.wsmo.org/2004/d3/d3.2/v0.1/200400517/resources/VTA-WGM1.wsml
source
http://www.wsmo.org/2004/d3/d3.2/v0.1/200400517/resources/VTA-WGM1.wsml
language
English
relation
http://www.wsmo.org/2004/d3/d3.2/v0.1/20040517/resources/capability.flr
http://www.wsmo.org/2004/d3/d3.2/v0.1/20040419/resources/goal.flr
coverage
rights
DERI
version
1.1
usedMediators
// no additional Mediators needed
sourceComponent
// identifier of the Goal
http://www.wsmo.org/2004/d3/d3.2/v0.1/200400517/resources/goal.flr
targetComponent
// identifier of the Web Service
http://www.wsmo.org/2004/d3/d3.2/v0.1/200400517/resources/ws.flr
reduction
comment: to be specified
|
GG-Mediators
A GG Mediator connects Goals by specifying a reduction between them. For example,
a GG Mediator would connect a Goal "buy a ticket" with another Goal
"buy a train ticket" by stating the ontological correspondance between
the Goals as a reduction. If 'train ticket' is a subclass of 'ticket', than
the reduction in the GG Mediator would specify that valid instances for the
second Goal have to be 'train ticket subclassof ticket'.
There is no GG Mediator needed in the use case.
WW-Mediators
A WW Mediator connects Web Services used by another Web Service in ther Orchestration,
resolving heterogeneities at all levels (data, process, protocol). There is
no WW Mediator in this use case, since the Web Service does not apply other
Web Services in order to realize its functionality (at least this is not modelled
at the moment). .
On basis of the models for the WSMO components specified above, we can define the following automated mechanisms: Web Service Discovery, Web Service Composition, and Web Service Execution. In the following we explain how these mechanisms work and which parts of the WSMO models they use.
Web Service Discovery is concerned with inference-based mechanisms that detect suitable Web Service for a given Goal. This means that the discovery mechanism searches available Web Service descriptions and determines whether these can be used to fulfill a certain Goal. The overall structure of WSMO supports Web Service discovery explicitly by introducing the notions of Goals and Web Services as top level building blocks. The requirements and the approach for Web Service Discovery in WSMO is exhaustively discussed in [Keller et al., 2004]. Here, we shortly summarize the most important aspects and explain how the discovery mechanism works in the use case models as specified above.
The functionality of the discovery mechanism can be separated into three major aspects:
With regard to the WSMO models of the use case defined above, we can now show how the heart of Web Service Discovery in WSMO, i.e. the Goal-Capability-Matching works. Therefore, we have to show that the Proof Obligation for Goal-Capability-Matching holds for the models of the Goal and the Capability defined in the use case.
The Proof Obligation for Goal-Capability-Matching is defined as follows
 |
This Proof Obligation states that under consideration of all Ontologies and Mediators used in the Goal and the Capability description (2nd line), if the user is able to provide concrete values for the input parameters of a service such that the Preconditions and Assumptions defined in the Capability (3rd line) are satisfied, and if the Capability postconditions imply the Goal postconditions and the Capability effects imply the Goal effects, that the Capability matches the Goal. In the use case, we have defined 1 Goal and 1 Capability (see Listing 5). Obviously, the Proof Obligation is fulfilled in this case because the Goal and Capability are homogenenous.
The realization of the Goal-Capability-Matching works as follows - all compuational resources for this are provided in Appendix A:
The idea for general Goal-Capability-Matching, based on the constrcution of Goals and Capabilities, is that there is a query that asks whether the Capability is fulfilled. This means there has to be fact that satisfies the body of the distinct description elements of the Capability. In the use case, this fact is the Goal. Considering the Goal as specfied in Listing 4, the Goal postcondition is fact that satisfies the body of the Capability postcondition in Listing 5: the Goal Postcondition is a ticket for a specific itinerary (from Innsbruck to Frankfurt with a certain departure), and the body of the Capability postcondition requires a ticket for an itinerary with start- and endlocation in Austria or Germany, respectively, and with a departure that has to be later than the current date (which is explicitely modeled in the Date and Time Ontology). The same holds for the Goal Effect in correlation to the Capability Effect. Therein, the domain knowledge specified in the ontologies is applied by the reasoner in order to determine the matching; for instance, the Goal specifies "innsbruckHbf" as the start location - and instance defined in the Train Connection Ontology, and the Capability postcondition states that the start location of the itinerary has to be located in Austria or Germany. The instance "innsbruckHbF" is located in "innsbruck", and "innsbruck" is located in Austria - so "innsbruckHbf" satisfies the condition of the Capability postcondition. Thus, the following query returns "oebbCap" as a Capability that can satisfy the Goal:
?- X:capability[postcondition] and X:capability[effect].
This structure of Goal and Capability description also supports detection of "overspecified" Capabilities for resolving a Goal. Imagine a Capability with a postcondition that contains a planeticket from A to B, and a train ticket within Austria and Germany. Here, also the Goal would satisfy the Capability postcondition; in order to ensure that only train tickets are booked, the reduction in the WG Mediator that connects this "overspecified" Web Service to the Goal restricts the set of valid information to train tickets only (see an example for this in Appendix A as well).
Observant readers will notice that the realization for Goal-Capability-Matching
depicted here does only consider the Capability Postcondition and the Capability
Effect with respect to all used Ontologies and Mediators for Goal-Capability-Matching
- in contrast to the general Proof Obligation introduced above. The reason for
this is both on a conceptual level and on the modeling realization. The underlying
understanding of a Goal is that it describes a desire "I want this"
without any respect to how this desire could be solved. This desire is modeled
as a specific ontology structure - the Goal model it says that the desire is
a "ticket for a specifc itinerary", and that there also should be
a trade for this ticket. The Goal does not state anyting about the input that
will be provided to a suiatble Web Service, neither the Goal does specify any
functional capabilities that might be needed for interacting with a Web Service.
On the other Hand, the underlying understanding of the Web Service Capability
is that this is a functional description of the Web Service, i.e. it describes
what the Web Service does. In general, the information provided in the Web Service
Capability are "If an input is provided for which certain conditions hold,
then the Web Service will return some result in relation to the input wherefore
also some conditions hold". With regard to this, the Capability description
elements are conceptualy separated into two groups: elements that define conditions
that have to hold before the Web Service is executed, and elements that describe
conditions that hold after the Service is executed. The former group consists
of the precondition (describes the requested input along with conditions on
it) and the assumptions (arbitrary state of the world that has to hold before
the Web Service can be executed), and the latter group is the postcondition
(describes the output of the service after execution along with conditions on
it) and the effect (arbitrary state of the world that holds after the execution
of the Web Service, i.e. changes in the world).
With respect to this underlying understanding of the WSMO description elements
it is obvious that for Goal-Capability-Matching only the 2nd group of Capability
description elements has to be considered , i.e. Capability postcondition and
Capability effect.The reason is that these information state what the Web Service
can provide in the end, and this is what we want to know within Goal-Capability-Matching.
This means by Goal-Capability-Matching we get the information that certain Web
Services (or a set of Web Services) can be used to solve the Goal; in the next
step we have to check whether we really can use the Web Serivce (can we provide
the required input? Do we support other technical features? etc.). Besides,
from a modeling perspective, you always need a counterpart for a description
notion that shall be taken into account for Goal-Capability-Matching. Within
the understanding of Goals outlined above and the description elements existing
for Goals in WSMO, there is not such a counterpart for Capability preconditons
and assumptions. The conceptual reason for this we explained above.
The solution for Goal-Capability-Matching outlined and implemented here seems
to be a proper approach towards a generic Goal-Capability-Matching in WSMO.
Nevertheless, it is not finished by now, and we briefly explain the background
and the arising challenges for further elaboration of this approach.
In the solution outlined above, the Goal is a fact. Such facts are needed as
they have to satisfy the body of the respective Capabilty notions. But the restriction
of modeling Goals as facts obviously limits the expressiveness of Goals - in
the end, we should be able to define an arbitrary object as Goal that might
contain variables and axioms. Also, from an application perspective, the facts
that will be applied to check satisfaction of the Capability postcondition and
effects would be generated by the internal functionality of the Web Service
(or at least by the Web Service owner), and not be stated by the Goal.
Such an approach wherein Goals and Capabilities are modeled as arbitrary logical
expressions requires a different, more compley and advanced mechanism for Goal-Capability-Matching.
More precisely, we would have to test whether a Capability logically entails
a Goal (as stated in the general Proof Obligation above). The result of Goal-Capbility-Matching
would be the same as in the solution outlined here: When there are facts that
satisfy the Capability postcondition and effect, and if the Capability logically
entails the Goal, than those facts will also satisfy the Goal, so the Capability
matches the Goal. This approach seems to be the most reasonable solution for
Goal-Capability-Matching within WSMO
Unfortunaltely, there are some problems for such a solution. Logical entailment
can be realized in different ways - by implication (the Capability postcondition
implies the Goal), or by query containment when the Goal and the Capability
are modeled as queries. But, logical entailment is not decidable, nor satisfiable
for First Order Logic in general. For a proper solution, the logical language
to be used for modeling Goals and Capabilities needs to be restricted to a decidable
subset of FOL - such as Horn Logic or Description Logic, or a combination of
these. Furthermore, there does not exists a proper implementation for determining
logical entailment (at least for the existing reasoners that we have considered
so far).
In conclusion, the approach for Goal-Capability-Matching presented here seems
to be a promising solution because it is heading towards the right direction.
For further development of the Goal-Capability-Matching technology within WSMO
as the heart of Web Service Discovery, very complex and challenging efforts
have to faced: definition of a decidable subset for the specification language
of WSMO as well as implementation of logical entailment for the reasoner to
be supported within WSMO.
We have described a real-world setting of using Semantic Web Services for a Virtual Travel Agency (VTA) that provides an end-user service for booking international train tickets, thereby aggregating Web Services of different e-Tourism Service Providers. The set up of this use case and the system architecture of the VTA here is conform to the general structure of the VTA use case described in Section 2.1.
Within the WSMO models defined in this use case we have shown how to model the different components of WSMO, with regard to the stable WSMO modeling elements avaliable at this point of time:
Furthermore, we have outlined the general workflow of the WSMO Discovery mechanism that works on the WSMO models for Goals and Capabilities.
The outcome of the first use case modeling are manifold. First of all, it shows how the different WSMO componts are modeled concretely. This gives answers to many questions that have been arising within WMSO: a more concrete understanding of Goals in WSMO and what they actually express, what is defined in a Web Service Capability with special regards to the difference between preconditions and assumptions, postconditions and effects respectively, and the concept of Mediators in WSMO. Further major outcomes of the use case and testing efforts so far is a concrete specification of how to model the different types of axiom definitions in WSMO, as well as further insights on Goal-Capability Matching as the heart of Web Service Discovery mechanism that work on the WSMO models for Goals and Web Service Capabilities. The scope of the use case is restricted to the most essential building blocks of WSMO at this point of time, and it will be updated and extended in the future for testing and showcasesing further WSMO constructs.
[not in this version]
similar to 3.1.1
similar to 3.1.2
similar to 3.1.3
similar to 3.1.4
Appart from discussing possible usage scenarios of Semantic Web Services, the major interest in this deliverable is to test and verify WSMO modeling for recursive development of WSMO, and to serve as a testbed for development of WSMO-based technologies. The deliverable is intended to exemplify and showcase the usage of WSMO for modeling different aspects related to Semantic Web Services, and it will continuously be updated according to further development of WSMO.
According to the current status of WSMO, the most interesting aspects are:
The major outcome of the Use Case modeling provided in this deliverable are:
The directions for future work in this deliverable are:
[Arroyo et al., 2004] Arroyo, S.; Lara, R.; Gómez, J.; Berka, D.; Ding, Y.; Fensel, D. (2004): Semantic Aspects of Web Services. In Munindar. P. Singh (Ed.), Practical Handbook of Internet Computing. Baton Rouge: Chapman Hall and CRC Press, Baton Rouge. 2004.
[Arroyo and Stollberg, 2004] Arroyo, S.; Stollberg, M.: WSMO Primer. WSMO Deliverable D3.1, available at: http://www.wsmo.org/2004/d3/d3.1/v0.1/
[ebXML] ebXML Deliverables http://www.ebxml.org/specs/
[Fensel & Bussler, 2002] D. Fensel and C. Bussler: The Web Service Modeling Framework WSMF, Electronic Commerce Research and Applications, 1(2), 2002.
[Keller et al., 2004] Keller, U.; Lara;, R.; Polleres, A.; Lausen, H.; Stollberg, M.: Inferencing Support for Semantic Web Services: Proof Obligations. WSML Deliverable D5.1 v0.1, WSML Working Draft 05 Arpil 2004. available at http://www.wsmo.org/2004/d5/d5.1/v0.1/20040405//d5.1v0.1_20040405.pdf
[Kifer et al., 1995] M. Kifer, G. Lausen, and James Wu: Logical foundations of object oriented and frame-based languages. Journal of the ACM, 42(4):741-843, 1995.
[Pan and Hobbs] F. Pan and R. Hobbs: Time in OWL-S, avaliable at: http://www.isi.edu/~pan/damltime/AAAIsymp2004.pdf.
[Oren et al, 2004] Oren, E. (Ed.): BNF grammar for WSML user language., WSMO Deliverable D16.1, Working Draft 18 April 2004 avaliable at: http://www.wsmo.org/2004/d16/d16.1/v0.2.
[Roman et al., 2004] D. Roman, U. Keller, H. Lausen (eds.): Web Service Modeling Ontology - Standard (WSMO - Standard), version 0.2 available at http://www.wsmo.org/2004/d2/v02/
[RosettaNet] RosettaNet http://www.rosettanet.org/
[SOAP] Mitra, N.: SOAP Version 1.2 Part 0: Primer. W3C Recommendation 24 June 2003. available at: http://www.w3.org/TR/soap12-part0/
[UDDI] Bellwood, T.; Clément, L.; von Riegen, C. (Ed.): UDDI Version 3.0.1. UDDI Spec Technical Committee Specification, Dated 20031014. available at: http://uddi.org/pubs/uddi_v3.htm
[WSDL] Chinnici, R.; Gudgin, M.; Moreaum, J.-J.; Weerawarana, S. (2003): Web Services Description Language (WSDL) Version 1.2. W3C Working Draft 3 March 2003. available at http://www.w3.org/TR/wsdl20/.
The work is funded by the European Commission under the projects DIP, Knowledge Web, Ontoweb, SEKT, and SWWS; by Science Foundation Ireland under the DERI-Lion project; and by the Austrian government under the CoOperate programme.
The editors would like to thank to all the members of the WSMO working group for their advises and inputs to this document. Very special thanks go to Michael Kifer who supported the development of the models for Goals and Capabilities and especially on elboration of generic solutions for the