International Conference on Semantic Computing
Remote Experiments as Semantic Web Services C´assio V. S. Prazeres University of S˜ao Paulo
[email protected]
Maria da Grac¸a C. Pimentel University of S˜ao Paulo
[email protected]
Abstract
cation [5, 6, 10]. Online experimentation allows students from anywhere to operate remote instruments at, usually, any time. In order to make experiments available on the web, WebLabs need to pay attention to three phases of online experimentation: pre-experimentation, execution and post-experimentation [14]. The pre-experimentation phase specifies preset information including input and output resources as well as temporal information related to the scheduling of the resources involved in the execution of the experiment. In the execution phase, the experiments need to be monitored. The monitoring facility of online experimentation may be used for teachers as a base for students evaluations or for the control of the experiment execution which may demand some special online attention of a technician. The post-experimentation phase defines arrangements including how to make available information relative to the results of the experiment execution and as well as any logging information relative to the experiment. Such information may be useful or, most likely, necessary for students to improve their reports, for teachers who need evaluate the students, and for technicians who need evaluate some technical aspects of the experiment execution. In order to offer their services on the Web, WebLabs must produce information associated with the three phases of the remote experimentation. However, most of the efforts related to remote experimentation are dedicated to solve the technical problems of automation involved in offering the remote access to the experiments. In this context, the three phases relative to management support, as described here, and a formalization of the experiments, when handled by WebLabs, are usually in an ad hoc way. As a result, is an opportunity for specifying a standard way to describe experiments and make possible the integration of distinct experiments in Web-based environments like an e-learning environment. In order to allow the integration of remote experiments in Web-based environments, we propose supporting remote experiments as Semantic Web Services. The use of ontologies to make experiments available as Semantic Web Ser-
Online experimentation allows users to operate remote instruments from anywhere and, usually, at any time. The amount of experiments that can be accessed trough the Web is growing quickly. These remote experiments have distinct characteristics, need a variety of resources, as well as present different availability and access procedures. Remote experiments can be supported as Semantic Web Services, demanding special attention in terms of discovering, selection, composition, scheduling, monitoring and making available the results. In order to allow the integration of remote experiments in Web-based environments, we propose supporting remote experiments as Semantic Web Services. Our approach, which we call RALOWS - Remote Access Laboratory Ontology and Web Service, reuse and extend known ontologies to provide formal description for experiments. More specifically, we combine temporal concepts to service and resource descriptions so as to make possible to specify scheduling and to impose restrictions in the use of the resources involved in the experimentation.
1
Introduction
Broadband networks allow quality user-interaction and, as a result, many laboratories have been extended so that their experiments can be accessed by remote users. These laboratories are usually named WebLabs when the experiments and associated instruments are accessed and controlled via the Web. There are several reasons to execute an experiment remotely. The instruments may not be available locally for the experiment, it is not possible or convenient for someone to be present at the site where the experiment must be carried out, etc. Examples of successful WebLabs are found in several domains such as robotics [8], telemedicine [13] and engineering [1], giving remote access to resources that are rare or expensive, to name a few. In the educational domain, in particular, remote laboratories have been proposed to complement in loco experiences so as to improve edu-
0-7695-2997-6/07 $25.00 © 2007 IEEE DOI 10.1109/ICSC.2007.54
Cesar A. C. Teixeira Federal University of S˜ao Carlos
[email protected]
791
vice promotes the standardization of tasks such as experiment discovery, selection, composition and monitoring, at the same time that allows automate these tasks. Our approach, which we call RALOWS - Remote Access Laboratory Ontology and Web Service, reuse and extend known ontologies to provide formal description for experiments. More specifically, we combine temporal concepts to service and resource descriptions so as to make possible to specify scheduling and to impose restrictions in the use of the resources involved in the experimentation. This paper is organized as follows. Section 2 presents some related ontologies that we reuse in RALOWS. A description of RALOWS is presented in Section 3. A discussion of the using RALOWS in a real scenario of experimentation is made in Section 4. Section 5 presents some related works. Section 6 summarizes our contributions and highlights future works.
2
represent information transformation; b) preconditions and effects represent state changes produced by the execution of the service, which forms the IOPE (inputs, outputs, preconditions and effects). For a detailed perspective on how a service operates, a service is better viewed it as a process. Specifically, OWL-S defines a subclass of ServiceModel, the ProcessModel. A process in OWL-S has two functions. One is to produce a data transformation from a set of inputs to a set of outputs. Second, to produce a transition in the world from one state to another. This transition is described by means of the preconditions and the effects of the process. The grounding of a service specifies the details of how to access the service – details having mainly to do with protocol and message formats, serialization, transport, and addressing. A grounding can be thought of as a mapping from an abstract to a concrete specification of those service description elements that are required for interacting with the service – in particular, the inputs and outputs of atomic processes. In OWL-S, both the ServiceProfile and the ServiceModel are thought of as abstract representations; only the ServiceGrounding deals with the concrete level of specification.
Related Ontologies
This section presents the OWL-S ontology and some related ontologies that we use in RALOWS to turn a remote experiment into a Semantic Web Service. In all figures representing an ontology, in this paper we use the following notation: an ellipse indicating an ontology class; a directed labeled line indicating a property (the label is the name of the property); a directed hatched line indicating class hierarchy; and, all extensions, which we made in the ontologies, and the document segments we present are in accordance with the OWL/XML syntax.
2.1
2.2
Resource Ontology
The Resource1 ontology proposed for OWL-S intends to be an ontology of resources stated at a level that is abstract enough to cover physical, temporal, computational, and other sorts of resources. Martin et al. [9] say that “specific kinds of resources will, of course, have specific properties; in this development we sketch out the principal classes of properties a resource might have”. In RALOWS, we use this ontology with some extensions (see Section 3.1) for the domain of remote experiments.
OWL-S
Figure 1 shows the OWL-S upper ontology for services. The intent of this ontology is to provide three essential types of knowledge about a service: the ServiceProfile, the ServiceModel and the ServiceGrounding (shown in Figure 1) [9].
Figure 1. OWL-S Ontology for Services [9]. The class ServiceProfile provides a superclass of every type of high-level description of the service. The ServiceProfile is used for publishing (providers) and discovering (requesters) services. The OWL-S Profile focuses on two aspects of the service functionality: a) inputs and outputs
Figure 2. The Resource ontology for OWL-S. As shown in Figure 2, the OWL-S resource ontology 1 http://www.daml.org/services/owl-s/1.0/Resource.owl
792
takes assumptions on three aspects: Allocation Types, Capacity Types and Resource Composition. Note that, again in Figure 2, resources can also have a capacityGranularity property, that is, the units in terms of which the capacity is measured. The resource ontology not makes some assumption about the range of the capacityGranularity property (represented by OWL Thing in Figure 2).
2.3
Time Ontology
A complete and complex ontology for time (the OWLTime2 ) was defined by Bobbs et al. [7], who afirm that OWL-Time is an abstract ontology of time intended to be a complete specification of a theory of time as, required for Semantic Web applications. Pan and Hobbs [11] present a sub-ontology3 of OWLTime named Time-Entry. The purpose of this sub-ontology is to provide quick access to the essential vocabulary in OWL for the basic temporal concepts and relations. For some simple applications, the OWL-Time is far more than required, then, in RALOWS, we will use the Time-Entry sub-ontology for time.
Figure 4. Our approach, RALOWS (Remote Access Laboratory Ontology and Web Service), reuse and extend those nown ontologies to provide formal description for experiments. A combination of temporal concepts and service and resource descriptions allow the specification of scheduling information as well as restrictions in the use of the resources involved in the experimentation.
Figure 4 shows the OWL-S upper ontology with the extensions that we propose in RALOWS and will be explained in details in the following sub-sections.
3.1
Resources are collections of one or more resources (recursively) and a resource is anything made available in the context of an experiment: instruments, software, chemicals, live animals, etc. We also refer as a resource to a human being when he or she is a person who supports the execution of an experiment (e.g., a technician) or he or she is an active agent with respect to the experiment (e.g, a tutor who interacts with remote students) [12]. The OWL-S ontology presents a property named provides (see Figure 1) which indicates that a Resource may provide a Service. For this work, in the domain of remote experiences, there are two types of resources as shown in Figure 4. Document 1 shows our extensions on the Resource ontology (see Section 2.2). We create two subclasses for Resource: InputResource (lines 6 to 9 of Document 1) and Instrument (lines 10 to 13 of Document 1).
Figure 3. Subclass hierarchy of temporal concepts. Figure 3 presents the subclass hierarchy of temporal concepts proposed for the Time-Entry sub-ontology. The most basic temporal concepts in the sub-ontology are are Instant, Interval, InstantEvent, and IntervalEvent.
3
Experiment Resources
RALOWS
Section 2 presented the OWL-S ontology for services and related ontologies. In this section, we present our approach to make experiments available as Semantic Web Services using OWL-S and related ontologies with some extensions made to adapt the ontologies for the domain of remote experiments.
0 1 2 3 4 5 6 7 8 9 10 11 12 13
2 http://www.isi.edu/∼pan/damltime/time.owl 3 http://www.isi.edu/∼pan/damltime/time-entry.owl
793
Resource Input Resource Instrument Resource
In this paper we do not present details about how a service is accessed (ServiceGrouding of the OWL-S). The grounding of a service specifies the details of how to access the service - details having mainly to do with protocol and message formats, serialization, transport, and addressing. Wrapping instruments as semantic web services makes possible the automatic discovery of experiments that are composed (see Section 3.3) by these instruments. The ServiceProfile is used to advertise the experiments for discovering purposes. The ServiceProfile requested (by a user that searches for some experiment) is matched with several ServiceProfile’s which are advertised by the providers.
The Instrument subclass of Resource corresponds to every resource that may be accessed and controlled through the Web to provide some service (e.g., an oscilloscope). This type of resource is detailed in Section 3.2. The InputResource subclass of Resource is a type of resource that may appear as an input for a service. Some examples are: a chemical solution demanded in a chemical experiment; a technician needed to supervise the execution of an experiment; a rat used in a behavioral experiment, and so on. As shown in Figure 4 and detailed in Document 2, an InputResource is a subclass of Input (lines 3 to 6 of the Document 2) that is the class for all inputs in OWL-S. We create this special type of input because some InputResource should be scheduled as a resource necessary to the execution of the experiment. 0 1
3.3
The task of service composition in OWL-S involves the automatic selection, composition, and interoperation of Web services to perform some task, given a high-level description of an objective. With the OWL-S markup of Web services, the information necessary to select and compose services will be encoded at service Web sites. Software can be written to manipulate these representations, together with a specification of the objectives of the task, to achieve them automatically [9]. An instrument, wrapped as a semantic web service, may represent an experiment; an experiment, otherwise, may require the connection of several instruments forming. In the latter case, we have a composed experiment. The IOPEs of an experiment are composed by the union of all IOPEs of all instruments involved composition of the experiment. However, some IOPEs may be default for all experiments:
2 3 4 5 6 7 8
Input Resource 9 10 11
Figure 4 also shows a property from ServiceProfile to InputResource named hasInputResource. This property, as specified in Document 2 (lines 7 to 11), is an extension as a sub-property of the property hasInput in OWL-S.
3.2
Wrapping Instruments as Semantic Web Services
An instrument is a type of resource that may be accessed and controlled through the Web to provide some service. Then, we propose that the instruments be wrapped as a Semantic Web Service in OWL-S. If all instruments “provide a service”, then each instrument presents a ServiceProfile, is describedBy a ServiceModel (with the ProcessModel) and supports a ServiceGrounding, as shown in Figure 4. Document 3 has the OWL of the property provides. 0 1 2 3 4 5 6
Locating, Selecting and Composing Experiments
• Inputs: UserName In, Password In, and IntervalEvent In. • Outputs: Results Out and ExperimentLog Out. • Preconditions: UserExist, ScheduledIntervalEventExist, and IsValidIntervalEvent. • Effects: ExperimentExecuted, ResultsGenerated, and ScheduledIntervalEventInvalided.
In order to someone to execute an experiment, he or she has to be an autorized user, and has to book the experiment and, as a result, has to be the owner of the slot time (an IntervalEvent in the Time-Entry sub-ontology) at the moment of the experiment execution. This interval event is an input for the execution of the experiment as the experiments need to know the time reserved to the user. The user name and password are inputs needed to autenticate the user. The results of the execution of the experiment and the experiment log are outputs. Finally, the effects are the experiment executed, the results generated, and a booked interval event.
The Inputs, Outputs, Preconditions and Effects of an instrument are created in the ServiceProfile. Each instrument has its proper IOPE’s which are intrinsic to the instrument. The process involved in the use of an instrument, to execute some experiment, is represented as a ProcessModel in OWL-S.
794
may want to monitor a student executing an experiment for evaluation or orientation purposes; other students may want to monitor the execution of the experiment, to mprove their learning; a technician may be required to monitor the execution of the experiment for some special control; etc. A web service also described by OWL-S can be used to monitor others services (experiments). This monitoring service may have, at least, two inputs: ProcessModel and the current status (in the ProcessModel) of the monitored service. All service that “wants” to be monitored must supply an output providing its ProcessModel and another output providing the current status of the execution. The monitoring service get these outputs as its inputs and generates an output with the current status of the service.
When a user requests an experiment that cannot be matched, the OWL-S matcher will try to select some services (experiments or instruments) that, combined, can satisfy the request of the user. As a result, the selected services will compose a new experiment. The selection and composition may be more or less automatic, depending on how detailed are the information in the ServiceProfile and ProcessModel.
3.4
Scheduling Experiments
In RALOWS, we use two classes of the Time-Entry subontology: Interval and IntervalEvent. The property isAvailableBy (lines 3 to 6 of Document 4) from the Resource class has a zero or more cardinality (not shown in Document 4). Then, a resource can have zero or more intervals of availability. If it has zero, the resource would not be scheduled and all experiments, that use (InputResource) or is provided (Instrument) by this resource, would not be scheduled too. 0 1 2 3 4 5 6 7 8 9 10
4. Analysis Using a Case Study In this section we present an experiment refereed to as the Skinner Box Experiment4 . This experiment is used as basis for analysis and evaluation of RALOWS. Section 4.1 presents the experiment, explaining how it is executed and what the objectives and expected results are. Sections 4.2 and 4.3 present, respectively, the Skinner Box Experiment described by OWL-S and described by our RALOWS approach.
4.1. The Skinner Box Experiment with Rats The property hasSchedule (lines 7 to 10 of Document 4) from the Resource class has a cardinality of zero or more (not shown in Document 4). The resource can have several IntervalEvent which correspond to a schedule to use the resource for an experiment (event) during a time slot (interval). The interval of the schedule needs to be a valid Interval, according to intervals of availability of the resource, and can not overlap an existing scheduled interval. In this paper we do not make any assumption about the service that schedules experiments. We propose the description of resource availability and resource scheduling by using the Time-Entry sub-ontology. A semantic web service for scheduling may be implemented to use the Time-Entry subontology and schedule the resources that compose an experiment.
3.5
The variation of the Skinner box experiment discussed in this section aims at conditioning a rat, placed at a special cage called Skinner box, to press a bar to liberate water. The rat is, first, deprived of water for some time so that it is supposed to be thirsty. The Skinner box, in this case, has a bar and a water dispenser which releases a water drop upon activation. This activation can be done externally, by a experimenter pressing a water-release button, or internally, by the rat pressing the bar in the cage.5 The overall process occurs by the experimenter choosing the correct moment to release the water drop, so that at, as the time passes, a more specific behavior is demanded from the rat. At first, the rat may be demanded to close to the bar by some distance which; the experimenter controls that the distance required to release water is gradually is reduced so that, at a certain point, water is only released when the rat touches the bar. By then, the rat usually touches the bar with its nose: the experimenter keeps demanding more specific behaviors from the rat so that water is only released when it presses the bar it its paws (one or more times). One variation of this experiment is the use of a large ring instead
Monitoring Experiments
Individual services and, even more, compositions of services, may often require some time to execute completely. A user may want to know, during this period, how is the status of his or her request, in a batch type remote experiment . Also, plans may change, requiring alterations in the actions the software agent must take [9]. Other interesting examples that may require monitoring, during interactive type remote experiments, are: a teacher
4 Also known as the operant conditioning chamber, it has been used by the Professor of Psychology B.F. Skinner for experimental analysis of animal behavior (see http://en.wikipedia.org/wiki/B. F. Skinner) 5 In this example, we do not model negative reinforcement.
795
The Repeat-While construct specifies an iteration and a stop condition (UserNotExit in line 3). The Choice construct indicates that only one process among those listed can be chosen at a time. In other words, the experimenter is allowed to turn the light on and off and to release water, one at a time.
of a bar: the rat is trained to pass through the ring (a number of times) before the water is released. Another common variation is the use of a light bulb which can turned on or off by the experimenter and used so that the water is released only when the the light bulb is on, for instance [14]. To execute the experiment remotely, a Skinner box has been connected to a computer an associated with an application containing a video window and 3 buttons. The video window shows images captured from a camera placed at the front of the cage. One of the buttons activates the water dispenser, another turns the light on and off, and the third causes a sound to be played and can be used to wake the rat when it feels sleep, for instance.
0 1
2 3 4 5 6 7
In this section, the Document 5 to 7 show the experiment description using the OWL-S descriptors without the extensions proposed in this paper. Document 5 shows all inputs, preconditions, effects and outputs of the experiment as IOPE’s in the ServiceProfile. In the document, the only precondition is that the rat is deprived of water (line 7).
4.3. Using RALOWS to describe the Skinner Box Experiment Documents 8, 9, 10, 11 and 12 show the experiment description using the OWL-S descriptors, resource ontology and time ontology with the extensions proposed in this paper as part of RALOWS. The first extension is shown in Document 8. In this document an Instrument is defined to describe a Skinner box as an instrument in RALOWS (an Instrument is shown in Figure 4 and Document 1). Document 8 defines three isAvailableBy properties (lines 4 to 6) for the “SkinnerBox” Instrument. These properties include time periods (described with the Time Entry sub-ontology) in which the resource is available for remote experimentation. In line 7, the hasSchedule property indicates a time slot scheduled for the use of the instrument. Line 8 specifies that this instrument provides the service SkinnerBoxExperiment.
Document 6 contains the OWL-S process for the experiment. This process is composed for a sequence (line 5) of two processes: an atomic process named SignIn and a composite process named ExecuteExperiment (lines 7 and 8). The composite process ExecuteExperiment is described in Document 7. 0 1
8 9 10 11 12 13 14 15 16 17 18
4.2. Using OWL-S to Describes the Skinner Box Experiment
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
0 1 2 3 4 5 6 7 8 9
2 3 4 5 6 7 8 9 10 11 12
In Document 9, the rat is specified resource is an InputResource. This resource does not provide services per se: because it is used as an input for the experiment, it must be scheduled which demands association if the isAvailableBy and hasSchedule properties.
In Document 7 we present the description of the composite process ExecuteExperiment. In line 3, we have defined a condition (Condition in OWL-S). This process is composed of four atomic process (lines 10 to 13). The composition of this process has two constructs: Repeat-While and Choice.
0 1 2 3 4 5 6
796
-->
Document 10 shows all inputs, preconditions, effects and outputs of the experiment as IOPE’s in the ServiceProfile. It differs from Document 5 by including (a) a new input which makes reference to the IntervalEvent (line 3) that should be scheduled for the experiment execution, and (b) two new preconditions specifying that it must exist an IntervalEvent scheduled (line 10) and that this IntervalEvent must be valid (line 11). Another difference is the IntervalEventInvalided effect (line 15) which says that the used IntervalEvent should be made invalid after the experiment is executed. These differences are possible because of our extensions use the Time-Entry sub-ontology. 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
15 16 17 18
By using OWL-S to describe an experiment as a Semantic Web Service, we have imported all aspects of the OWL-S to the remote experiment domain. The fact that RALOWS combine other ontologies and adds temporal concepts to the service and resource description makes it possible to do scheduling and to impose restrictions in the use of the resources involved in the experimentation.
5
Yan et al. [16] propose a double client-server architecture for online experiment systems and the methodology to wrap the functions of instruments into Web Services. The authors pay attention on how the service is implemented and accessed. To do that, they map instrument operations in WSDL operations. In RALOWS we do not make assumptions on how the service will be grounding. Our objective, in this paper, is to make remote experiments available with the semantic descriptors based on OWL-S, Resource ontology and Time-Entry. Hence, we can use what the authors proposed in [16] to make a concrete realization of the services. The service discovery in OWL-S is made by matching IOPE’s from the requester with the ones from the provider of the service. Xie et al. [15] propose the use of natural language for access web services described with OWL-S. This solution can be used together with RALOWS to map user query requests (in natural language) into their corresponding web services described with OWL-S. In Section 3.5 we proposed the use of a monitoring service together with the ProcessModel to monitor remote experiments. The references [2, 3, 4] present proposals of monitoring Semantic Web Services and may be evaluated to use with RALOWS. The authors of the references [12, 16] propose style sheets transformations (using XSLT) to design the web GUI for the instruments. We can use this approach to generate the web GUI for the instrument based on the descriptions of its ServiceProfile.
Documents 11 and 12 correspond, respectively, to Documents 6 and 7 adapted with our proposed extensions. In Document 11, a new atomic process has been included: SelectIntervalEvent (line 8). This process tests if the user has a valid IntervalEvent to execute the experiment. 0 1
2 3 4 5 6 7 8 9 10 11 12 13
In Document 12, the change is in the stop condition. In Document 7 the stop condition occurs when the user spontaneously exits of the experiment. In Document 12 the stop condition occurs when the IntervalEvent scheduled for user expires. 0 1
Related Works
6
Final Remarks
We have proposed RALOWS - Remote Access Laboratory Ontology and Web Service, an approach to make remote experiments available as Semantic Web Services. We have also discussed the use of this approach in a real experimentation scenario. RALOWS reuse other ontologies which are combined in a novel way and extended so as to provide ontology-based descriptions for remote experimentation.
2 3 4 5 6 7 8 9 10 11 12 13 14
797
RALOWS combines the Time-Entry sub-ontology from OWL-Time with the OWL-S ontology (Semantic Markup for Web Services), with its Resource ontology, at the same time that adds temporal concepts to the service and resource description that makes it possible to do scheduling and to impose restrictions in the use of the resources involved in the experimentation. By using OWL-S to describe an experiment as a Semantic Web Service, we have imported all aspects of the OWLS to the remote experiment domain. Aspects like location, selection, composition, invocation, and monitoring of Web Services automatically is applied to the remote experimentation domain. We have extended the Resource ontology to create two types of resources. These resources were created to separate two concepts of resources in the remote experimentation domain: the resources that provide some services and the resources that serves as input for an experiment. The approach proposed and presented in this paper was discussed in one real scenario of experimentation, from which we have previously implemented an ad hoc solution. We plan to further discuss our approach in other scenarios to validate and obtain feedback to evaluate the approach. Also regarding evaluation, we expect to use some OWL inference machine to take some assumptions about the proposed approach.
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
Acknowledgments [12]
We would like to thank FAPESP for supporting the authors in the project in which context this work has been developed, the TIDIA-Ae project. We also thank the following organizations: CAPES, for financial support to C´assio Prazeres and to the Graduate Programs in Computer Science at the University of S˜ao Paulo and at the Federal University of S˜ao Carlos; CNPq, for the financial support to Maria da Grac¸a Pimentel.
[13]
[14]
References [1] J. A. D. Alamo, L. Brooks, C. McLean, J. Hardison, G. Mishuris, V. Chang, and L. Hui. The MIT Microelectronics WebLab: A web-enabled remote laboratory for microelectronic device characterization. In World Congress on Networked Learning in a Global Environment, page 7, Berlin, Germany, May 2002. [2] F. Barbon, P. Traverso, M. Pistore, and M. Trainotti. Runtime monitoring of instances and classes of web service compositions. In ICWS ’06: Proceedings of the IEEE International Conference on Web Services (ICWS’06), pages 63–71, Washington, DC, USA, 2006. IEEE Computer Society. [3] L. Baresi, C. Ghezzi, and S. Guinea. Smart monitors for composed services. In ICSOC ’04: Proceedings of the
[15]
[16]
798
2nd international conference on Service oriented computing, pages 193–202, New York, NY, USA, 2004. ACM Press. F. Casati, E. Shan, U. Dayal, and M.-C. Shan. Businessoriented management of web services. Commun. ACM, 46(10):55–60, 2003. M. Casini, D. Prattichizzo, and A. Vicino. The automatic control telelab: a user-friendly interface for distance learning. IEEE Transactions on Education, 46(2):252–257, 2003. J. M. Ferreira and D. Mueller. The MARVEL EU project: A social constructivist approach to remote experimentation. In REV’04: 1st Remote Engineering and Virtual Instrumentation International Symposium, page 11p, 2004. J. R. Hobbs and F. Pan. An ontology of time for the semantic web. ACM Transactions on Asian Language Information Processing (TALIP), 3(1):66–85, 2004. A. M. Khamis, F. J. Rodriguez, and M. A. Salichs. Remote interaction with mobile robots. Auton. Robots, 15(3):267– 281, 2003. D. Martin, M. Burstein, J. Hobbs, O. Lassila, D. McDermott, S. McIlraith, S. Narayanan, M. Paolucci, B. Parsia, T. Payne, E. Sirin, N. Srinivasan, and K. Sycara. W3C member submission 22 november 2004: OWLS: Semantic markup for web services. [Online]. Available: http://www.w3.org/Submission/OWL-S/. Visited on 15/07/2007., 2004. MIT. The challenge of building Internet accessible labs. [Online]. Available: http://icampus.mit.edu/ilabs/. Visited on 15/07/2007., April 2005. F. Pan and J. R. Hobbs. Time in OWL-S. In AAAI ’04: Proceedings of the AAAI Spring Symposium on Semantic Web Services, page 8, Palo Alto, CA, USA, 2004. C. Prazeres and C. Teixeira. A structured documentbased approach for WebLab configuration. In WebMedia ’06: Proceedings of the 12th Brazilian symposium on Multimedia and the web, pages 1–10, Natal, Rio Grande do Norte, Brazil, 2006. ACM Press http://doi.acm.org/10.1145/1186595.1186597. D. Reske and Z. Moussavi. Engineering in medicine and biology. In EMBS/BMES ’02: 24th Annual Conference and the Annual Fall Meeting of the Biomedical Engineering Society, volume 3, pages 1847–1848, 2002. C. A. C. Teixeira, D. Capobianco, C. Prazeres, and M. Barbosa. Processo de modelagem de resposta: Refinando requisitos de software de apoio a laboratorios de acesso remoto. (in Portuguese). In SBIE ’05: Proceedings of the 16th Brazilian symposium on computer science in the education, pages 1–11, Juiz de Fora, MG, Brazil, 2005. F. Xie, H. Gong, D. Deng, S. Wang, G. T. Wang, J. Hu, and P. C.-Y. Sheu. Integrating semantic web services for declarative accesses in natural language. In ISM ’06: Proceedings of the Eighth IEEE International Symposium on Multimedia, pages 201–208, Washington, DC, USA, 2006. IEEE Computer Society. Y. Yan, Y. Liang, and X. Du. Controlling remote instruments using web services for online experiment systems. In ICWS ’05: Proceedings of the IEEE International Conference on Web Services (ICWS’05), pages 725–732, Washington, DC, USA, 2005. IEEE Computer Society.
