Upload
others
View
1
Download
0
Embed Size (px)
Citation preview
Templates for Supporting Conceptual Modeling: a Delicate Balance
Chiara Ghidini, Muhammad Tahir Khan, and Chiara Di Francescomarino
FBK-IRST, Via Sommarive 18, 38050 Trento, Italy.|ghidini|tahirkhan|dfmchiara|@fbk.eu
Abstract. Using templates for conceptual modeling is becoming a fashionable way to guide and facilitateDomain Experts in providing rich and good quality knowledge. A possibility to build templates without startingfrom scratch is grounding them on existing foundational and core ontologies. In this paper we investigate howthese ontologies can be effectively used for the construction of templates able to guarantee a balance betweenusability and rigorousness. We report findings and lesson learned from a survey carried out for the evaluationof templates built from existing foundational and core ontologies in the enterprise domain.
1 Introduction
Effectively involving domain experts in the process of authoring OWL ontologies, making them able not only toprovide domain knowledge to knowledge engineers but also to directly write ontologies on their own or togetherwith knowledge engineers, is increasingly recognised in a number of works (see e.g., [7, 2, 22]) as a crucial stepto make the construction of ontologies more agile and apt to the needs of organisations and business enterprises.
In spite of the recognition of the problem, and of the efforts and progresses made in the area of knowledgeand ontology modelling, involving domain experts in the process of authoring OWL ontologies that provide anappropriate “computational representation of the structure, activities, processes, information, resources, people,behavior, goals, and constraints” [17] of their enterprise is still a challenging task and several efforts in the knowl-edge engineering field are currently devoted to study how to guide and facilitate domain experts in authoring OWLontologies, even if they lack a deep understanding of knowledge engineering techniques and formal languages suchas OWL.
A popular way of involving domain experts in authoring conceptual models is to provide them with graphicallanguages (diagrammatic representations). Nevertheless, as argued in [15] and briefly summarised in Section 2,diagrammatic representations are often adequate to represent only part of the knowledge that needs to be containedin a conceptual model. Templates, usually constituted by tables, forms, controlled language sentence patterns orany other kind of structured textual information, have been proposed in literature as an effective way of collectingthe rich and diversified knowledge that needs to be contained in conceptual models and to complement the onecontained in the graphical part.
While most of the analysis contained in [15] refers to software engineering conceptual modelling languages,such as UML, ER, ORM and Data flow, a similar analysis applies also to ontology based languages such asOWL. In fact, diagrammatic representations are mainly used to represent the taxonomic and membership structureof an ontology (especially these structures involve a limited number of concepts and instances) and to makethem easily understandable to ontology authors, but are not used to render more complex knowledge such asaxioms and properties of relations. Template-based approaches can therefore be applied to “encapsulate knowledgeand modelling recurrent sets of axioms” in ontology engineering, as explicitly proposed in [20] and indeed theyhave been used to better involve domain experts in authoring the entire content of specific domain ontologies inexisting tools such as MoKi [12], COE [13] and Populous [14] where templates are used to complement graphicallanguages and guide domain experts to describe ontology elements in a structured textual fashion. Templates,however, are difficult to build. They have to balance the needs of both domain experts and knowledge engineers,i.e., being usable but also rigorous enough. Moreover, methodologies for the definition of usable and rigoroustemplates need to be put in place, in order to facilitate their adoption and the customisation of tools that make useof them to specific scenarios.
In this report we propose an approach along with a concrete method to build ontology templates based on amixture of foundational, mid-level and specific ontologies, and we evaluate it through a survey conducted withdomain experts and knowledge engineers. The reason to ground templates on existing ontologies, instead of build-ing them from scratch, is that of exploiting the modelling patterns of several key entities of a domain alreadyencapsulated in well crafted ontologies, and of making them available to domain experts to reuse and adapt.
More in detail we investigate the ease of understanding, usefulness, completeness and correctness of tem-plates for modelling key entities of an enterprise ontology based on 23 foundational and core (including mid-leveland domain-specific) ontologies. The main findings of our analysis are the following: (i) grounding templates onexisting ontologies allows the construction of templates usable and rigorous enough; (ii) opportunely groundingtemplates on foundational or mid-level ontologies and characterising them with specialised domain ontologiesallows us to reach the right level of granularity of templates, thus making them more usable and precisely charac-terised; (iii) domain experts perceive template usability more positively than knowledge engineers, thus validatingthe key role of templates in supporting Domain Experts.
To the best of our knowledge, the evaluation performed in this work provides a first empirically rigorousevaluation of the support provided by a wide range of existing ontologies to the definition of structured forms thatcan guide domain experts in ontology modelling and therefore highlight, in a comprehensive manner, the potentialand criticality of using existing ontologies as a base for template-based approaches.
2 Template-based Conceptual Modeling
As extensively described in [15], template-based approaches have been used as an effective method to elicit knowl-edge in the area of software engineering and database in the last 40 years. By templates, we refer here to any kindof structured textual information (such as tables, forms, controlled language sentence patterns) where the mean-ing of the textual information contained in the template can be determined according to its position within thestructure. The first examples where tabular representations for the description of software functions [21], andglossaries for a structured description of terms relevant for a domain used in database design methodologies suchas DATAID [1] and more recently in the KCPM (Klagenfurt Conceptual Predesign Model) template-based mod-eling language [25]. With the advent of object-oriented analysis and design, and of popular graphical modelinglanguages for the representation of use cases such as UML (Unified modeling language), templates remained aneffective way to complement the graphical use case description with detailed information, which usually consti-tutes the biggest portion of knowledge of a conceptual model as graphically represented in Figure 11, as describedin [6] and in more recent approaches such as the NDT methodology [8].
The visible: a graphical representation of knowledge. Examples: classes, associations, attributes, IS-A relation, ...
The hidden: non graphically represented knowledge. Examples: descriptions, pre/post-conditions, exceptions, complex axioms, ...
Fig. 1: The iceberg of knowledge in conceptual models.
In the last few years several works have started to propose templates also to support Ontology Engineeringactivities in order to encapsulate complexity and to better involve domain experts in the activity of ontologyauthoring[20, 12–14]. Among the main advantages of introducing templates in Ontology Engineering listed inliterature are: (i) a decrease in complexity, as templates hide the complexity of the axioms and expose only therequired parameters to be filled in; (ii) an increase in efficiency of modelling, as templates can be reused in the
1 Taken from [15]
2
Fig. 2: A MoKi page.
modelling of several ontologies; and (iii) an increase in reliability, as templates comprise set of axioms developedby domain experts who, therefore become place domain knowledge at the heart of the modeling process.
2.1 Templates in MoKi
MoKi2 is a collaborative MediaWiki-based [9] tool where heterogeneous teams of knowledge engineers and do-main experts, with different knowledge engineering skills, can actively collaborate to the modelling of ontologicaland procedural knowledge in an integrated manner3. MoKi is grounded on three main pillars, which we brieflyillustrate with the help of Figure 2:
– Similarly to most state of the art wiki based tools, MoKi associates a wiki page to each each basic entity of theontology (i.e., concepts, object and datatype properties, and individuals). For instance, the concept Mountainin Figure 2 is associated to a wiki page which contains its description;
– each wiki page describes the entity it is associated with by means of both unstructured (e.g., free text, images)and structured (e.g. OWL axioms) content. This to store both the rich textual and image-based description ofinformal content and the formal description in the same tool.
– a multi-mode access to the page content is provided to support both domain experts and knowledge engineersin the authoring process. In particular, MoKi provides two modes to access the structured formal content: afully-structured access mode, where knowledge engineers can edit the ontology content by means of OWLaxioms, and a lightly-structured access mode, where domain experts, can edit the ontology content by meansof a simplified, template-based, view on the (same) structured formal description.
2 A comprehensive description of MoKi is presented in [11, 5]. The interested reader can also refer to http://moki.fbk.eu
3 Though MoKi allows to model both ontological and procedural knowledge, here we will limit our description only to thefeatures for building ontologies.
3
As argued in [11], one of the advantages of MoKi, and of MediaWiki-based tools in general, is the ability ofeasily customising their form-based user interface. In fact, while the lightly-structured access mode of Figure 2represents the rather general template included in the current version of MoKi, specific customisations of MoKicontain ad hoc templates for the description of different entities.
Data Properties
Full Name
Location
Start Date
End Date
Event type
Event description
Event
Every Event has partecipant
Every Event is hosted by
Every Event is organized by
Every Event is associated to
Object PropertiesOrganization/People
Organization/People
Organization/People
Process
Is aEvery Event is a
Fig. 3: A template.
However, defining ad hoc templates such as the one drafted in Figure 3, which contains a structured descriptionof key domain knowledge, that is the concept “Event”, and can be used by domain experts to: (i) refine / adapt thecharacterisation of the concept to the specific domain; (ii) model more specific concepts (e.g., a concept “Productlaunch event”) in a guided manner; and (iii) populate the ontology with specific instances of the concept (e.g., theinstance “Fiat 500 UK launch event”) is a difficult and time consuming task, as templates need to be produced bybalancing usability aspects and rigorousness aspects.
To address this problem we have investigated an approach which builds templates based on a mix of founda-tional, mid-level and domain-specific ontologies, which we present in Section 3 and evaluate in Section 4. Notethat while the work presented in the paper is motivated by the work on MoKi, as described in the previous section,the methodology presented in Section 3 and the evaluation presented in Section 4 have a general character and donot depend upon the tool used. Therefore they can provide a first insight on the potential and problems of usingexisting ontologies to structure templates for ontology engineering.
3 Building Templates for Enterprise Ontologies
In this section we provide an overview of the approach we followed to construct templates for the domain ofEnterprise Ontologies. We selected the domain of Enterprise Ontologies as this provides an important and generaldomain where domain experts are usually involved, but the suggested mehtod that we have followed can providesuitable guidelines for the definition of templates to support the task of ontology engineering in a wide range ofspecific domains.
4
The approach The approach relies on building the templates based on a model extracted from a range of existingontologies, which include foundational, mid-level and domain-specific ontologies. The reason to ground templateson existing ontologies, instead of building them from scratch, is that of exploiting the ontological analysis per-formed and modelling patterns already encapsulated in well crafted ontologies, and of making them available todomain experts to reuse and adapt. The reason to exploit a number of existing ontologies, ranging from foun-dational to mid-level, to specific ones is to evaluate the appropriateness of the different resources to the specifictask.
The Realisation The method that we have followed is composed of the four macro steps illustrated in the remainingof the section.
As a first step we focused on selecting entities relevant to our problem domain from foundational ontologies.The reason to start from foundational ontologies is that they contain a well founded characterisation of the foun-dational ontological categories, and therefore provide a good a starting point for building good quality ontologies.In our case we decided to use the DOLCE ontology [18], which, we believe, provides a good reference ontologyfor defining templates on a wide range of domains which encompass the one investigated in our experiment.
The second step concerned in complementing / specifying the entities selected from the DOLCE ontologywith entities selected from core (mid-level) ontologies. In fact, while grounding the templates on DOLCE helpsus to build templates that provide a rigorous characterisation of the entities they have to describe, we need toproduce templates that guide domain experts in the modelling of specific entities for a given domain. Mid-levelontologies provide the best candidate for this task as they usually describe specific (yet general) concepts for aparticular domain (in our case the enterprise domain). Our strategy was first to review mid-level ontologies thatare produced either as extensions of DOLCE (see e.g., the Ontology of Social roles presented in [19]) or groundedon DOLCE (see e.g., [3]) and to select entities relevant to our domain together with their characterisation.
The third step was to put together the entities selected in the first two steps in a taxonomy of 20 entities,asillustrated in Figure 4. The leaves of this taxonomy constitute the 8 entities to be characterised in the templatesthat we have evaluated in our experiment, shown in Figure 5 . Note that while reviewing the mid-level ontologieswe found out that some of the entities we were interested in are contained both in foundational and mid-levelontologies. In our experiment that was the case for “event” and “process”. In Figure 4 we have decided to depictthe entities in the most abstract type of ontology in which they were found. Thus “event” and “process” are listedamong the entities coming from foundational ontologies. We also found out that several entities are contained inmore than one mid-level ontology. In these cases we decided to merge the characterisation of the entities containedin the different ontologies to produce a more comprehensive one, and possibly to solve some conflicting issues,which were nevertheless extremely limited in our experience.
The final step concerned in evaluating the completeness of the characterisation of the different entities. Evalu-ating completeness is a complex task. Therefore we decided to take a pragmatic approach and aim at a characteri-sation composed of a number of properties ranging between 5 and 20. This, in order to have enough “attributes” todescribe the entity, but not too many to become a burden for the domain experts. To complement the descriptionof the entities we looked at domain-specific ontologies such as The Enterprise Ontology [23] and Schema.org4.The whole process resulted in an ontology of 20 concepts, 86 relations and 159 axioms which has provided thebasis for the construction of the templates for Organization, Actor, Role, Process, Event, Task, and Artefact. Indetail, the ontology Description Logic axioms have been used for the automatic construction of the templates: Fig-ure 3 shows an example of the resulting Event template. A description of the characterisation of the seven entities,together with the documentation of the source ontologies that have been taken from and the final templates aredescribed in Appendix.
4 Survey Evaluation
We are interested in investigating whether templates built starting from existing ontologies can be used to supportdomain experts in the conceptual modeling while preserving the rigorousness and the accuracy characterizing thegood principles of knowledge engineering. These templates should indeed be able to fill the gap between DomainExperts and Knowledge Engineers while balancing usability and ability to capture the design needs. In order toevaluate whether they can suit the different requirements they are supposed to satisfy, we built (Section 3) a set oftemplates grounded on different existing ontologies ranging from foundational to very specific domain ontologiesand related to the enterprise domain, evaluating them through a survey.
4 http://schema.org
5
Fig. 4: A case for Enterprise Model.
Fig. 5: A case for Enterprise Model: Selected entities for Templates.
4.1 Survey Description
The goal of the conducted survey is investigating (i) the usability (in terms of ease of use and usefulness) oftemplates built starting from existing ontologies as perceived by both Domain Experts and Knowledge Engi-neers; and (ii) the capability of these templates to completely and precisely capture the modeling needs of thespecific domain as perceived by both Domain Experts and Knowledge Engineers. The research questions weare interested to evaluate are:
RQ1 Are templates built from existing ontologies perceived as usable by Domain Experts and KnowledgeEngineers?
RQ2 Are templates built from existing ontologies able to completely and precisely satisfy the modeling needs ofa specific domain according to Domain Experts and Knowledge Engineers?
RQ1 aims at investigating the usability of templates by their final users, i.e. Domain Experts and KnowledgeEngineers. RQ2, instead, evaluates the modeling capabilities offered by the templates through their characteri-zations.
In order to answer the above research questions, we implemented the seven templates described in Section 3in MoKi and we asked a group of experts including both Domain Experts and Knowledge Engineers to answer
6
a questionnaire investigating four main factors: (i) ease of understanding; (ii) usefulness; (iii) characterizationcompleteness and (iv) characterization precision for each template. We used the subjective perception expressed bysubjects about the first two factors for investigating whether templates built using well known existing ontologicalresources are actually usable (i.e., for answering RQ1) and the last two factors for providing a subjective evaluationon the feasibility of the usage of ontological resources for addressing the real needs of the users in terms ofcharacterizations (RQ2).
The subjects involved in the survey were 43 among Domain Experts and Knowledge Engineers. DomainExperts are mainly enterprise workers with, on average, high knowledge of enterprises and medium expertisein modeling. Knowledge Engineers are instead both modeling experts and ontologists with average modelingexpertise almost high. We asked each of them to evaluate a set of three (TA) or four (TB) templates. In detail 21subjects (11 Knowledge Engineers and 10 Domain Experts) were asked to evaluate the TA templates (Role,Artefact and Actor), while the remaining 22 subjects (13 Knowledge Engineers and 9 Domain Experts) toevaluate the TB templates (Organization, Task, Event and Process). Table 1 reports the distribution of enterprisetemplates among the different experts.
TA TB
Knowledge Engineers 11 13Domain Experts 10 9
Table 1: Distribution of the enterprise templates among Domain Experts and Knowledge Engineers
Each template has been evaluated by means of a set of questions belonging to two categories: (i) closed-endedquestions, where answers were selected from a 5 point Likert-scale, (ii) open-ended questions, where participantswere allowed to provide suggestions for improving the templates. In detail, we used closed-ended questions and 5point scales for asking experts their subjective perception about template ease of understanding (0=immediate, ...,2=reasonable, ..., 4=complex), usefulness (0=absolutely useful, ..., 2=neither useful nor useless, ..., 4=absolutelyuseless) and completeness (0=totally complete, ..., 2=neither complete nor incomplete, ..., 4=totally incomplete).For the precision, instead, subjects were asked to evaluate, for each template characterization whether they wouldkeep or discard it (yes/no answer). For each template, we used the collected answers to compute, according to theinformation retrieval notion of precision, the subjective precision of each expert, i.e., p(si) = |ACT (si)|
|CT | , whereCT is the set of the provided characterizations for the template T , and ACT (si) ⊆ CT the subset of CT positivelyjudged by the expert si. We then classified the values of each subjective precision in three categories: high (100%precision, i.e., the subject agreed on all the provided template characterizations), reasonable (the precision isbetween 66% and 100%, i.e., more than 2/3 of the characterizations have been judged correct), and low (less than66% of precision). Templates have been made available to subjects on-line and questions provided in the form ofan on-line survey.
4.2 Survey Results
In this subsection we report the results of our observations and the statistical analyses we applied to strengthenthese results. In detail, to assess the positivity of the experts’ evaluations, since values are derived from an ordinalscale, we test medians using a one-tailed Mann-Whitney test [26], verifying the hypothesis that F ≤ 2, whereF can be each of the four factors under investigation, F is its median and 2 the intermediate value of the scale.Moreover, in order to evaluate the magnitude of the statistical significance obtained, we computed the effect size,that provides a measure of the strength of the relationship between two variables, by means of the Cohen’s dformula. Finally, we used a one-tailed Mann-Whitney test [26] to compare pairs of templates with respect tothe same factor, thus identifying possible templates performing worse or best with respect to the others. All theanalyses are realized with a level of confidence of 95% (p-value < 0.05), i.e., there is only a 5% of probabilitythat the results are obtained by chance.
Table 2 reports the descriptive statistics (median value and percentage distribution) related to the answersprovided by subjects for RQ1 and RQ2. For the sake of readability, in reporting the percentage distribution, wegroup together the two positive and the two negative values of the three 5 point scale, thus reporting the percentageson a 3 point scale, as for the subjective precision (high, reasonable and low).
7
By looking at the table, it comes out that subjects perceived the templates as overall easy to understand (themedian value is 1, i.e. “easy” on the 5 point scale). Such a overall positive evaluation is also confirmed at statisticallevel (p-value < 0.05) and by the large effect size5. Indeed, by looking at the distribution of the provided answers,more than half of them state the ease understanding of templates, while in only 11% of the cases the tempmlateshave been evaluated difficult to comprehend. Nevertheless, the number of answers that are clearly positive ornegative is not trivial (about 33%). Although reasonable could also be considered a positive evaluation, in theset reasonable answers also fall the ones of experts that did not want to risk taking a clear stand. Consideringthe distribution of these “answers in the middle” per template, we found that the evaluation provided for theRole template are significantly worse (at statistical level) than those provided for the Task, Event and Processtemplates, thus remarking the criticality of this template. By looking at Figure 6 (left), showing the subjectiveevaluation related to ease of understanding provided by the experts for each template, the criticality of the Roletemplate, for example, is quite evident: 62.5% of the subjects expressed a borderline answer for this template. Thehigh number of reasonable answers, hence, can be partially reconducted to this case. On the contrary, the Eventtemplate is the easiest to understand resulting at statistical level easier to understand not only than the Role butalso than the Artifact one. subjects provided an overall positive evaluation also with respect to the usefulness.
Moreover, by further investigating the figure and comparing pairs of template evaluations (related to the easeof understanding), we found that the evaluation provided for the Role template are significantly worse at statisticallevel) than those provided for the Task, Event and Process templates, thus remarking the criticality of this template.
On the contrary, the Event template looks the easiest to understand, resulting at statistical level easier tocomprehend not only of the Role but also the the Artefact one.
RQ Factor Median p-value cohen-d Rate Percentage
RQ1
Ease of Understanding 1 < 0.05 0.87easy 55.3%
reasonable 33.3%difficult 11.4%
Usefulness 1 < 0.05 0.82useful 64.9%
neither useful nor useless 20.2%useless 14.9%
RQ2
Completeness 1 < 0.05 0.6complete 56.1%
neither complete nor incomplete 25.4%incomplete 18.4%
Precision 1 < 0.05 2.43high 82.5%
reasonable 14.9%low 2.6%
Table 2: Statistical measure of the survey results
Subjects provided an overall positive evaluation also with respect to the usefulness. The median value of theevaluation on the 5 point scale is indeed 1, i.e., useful. The result is also statistically confirmed by the Mann-Whitney test and, also in this case, the effect size is high. Templates have been judged overall useful in almost65% of the cases versus the 15% of useless evaluations. In this case, the borderline answers were slightly less(about 20%) than for the ease of understanding, still representing a non-trivial percentage. One of the reasons forthese uncertain answers could be the lack of usage of the templates in a practical setting: subjects were providedwith templates and questionnaire without being required to perform any practical task. This way, unless they hadalready been exposed to similar situations, getting a clear idea of the usefulness of the templates, could havebeen difficult. This explanation is in part confirmed by looking at the distribution of the unclear stands betweenKnowledge Engineers and Domain Experts. Almost 96% of neither useful nor useless answers have beenprovided by Knowledge Engineers. Also in this case, the evaluation provided for the Role template are worsethan the others. Indeed, also at statistical level the usefulness of the Role template has been perceived as worsethan the one of Actor, Task, Event and Process.
We can hence conclude that, except for some critical templates as the Role one, usable templates can bebuilt starting from existing ontologies and hence positively answer to RQ1. Nevertheless, to have a more preciseevaluation, mainly of the usefulness of these templates, subjects should had the possibility to use them in practicaltasks.
The third row of Table 2, reports median, statistical analyses and distribution of the survey answers related tothe characterization completeness. Experts evaluated the characterizations as overall complete: the median value
5 The effect size is considered to be small for 0.2 ≤ d < 0.5 medium for 0.5 ≤ d < 0.8 and large for d ≥ 0.8
8
Fig. 6: Template usability subjective evaluation
is 1, i.e., complete in the 5 point scale, with ˜Compl < 2 (p-value < 0.05); the cohen-d effect size, instead, islower than for the other factors (medium). Indeed, in this case, although more than half of the evaluations werepositive, the percentage of the negative ones is slightly more than for the other factors (18%). Similarly to theusefulness evaluation, most of the 25% of uncertain answers (79.3%) was provided by Knowledge Engineers,who are overall more rigorous about what a complete characterization is. By looking at the figure, the Organizationtemplate seems to be the most problematic one. Indeed, the subjects’ answers related to its completeness result tobe (at statistical level)worse than those of Artefact, Actor and Event, while no other significant differences existamong the other templates.
Finally, the fourth row of Table 2 reports the values related to the precision of the template characterization.Overall the precision is high: the median is 1, i.e., the highest value and P ≤ 2 (with p-value ≤ 0.05). Indeedmost of the evaluations (82%) are high, while only a small percentage(2.6%) is negative. Again, when looking atthe distribution of the subjects’ evaluation on the different templates we found that there are some differences. Forexample the Event and the Process templates outperform the others, while the Role and Actor characterizationsare less precise. The across-template statistical test also confirms this result indicating a worse performance of theRole and Actor templates with respect to the Event and the Process ones.
Overall we can hence positively answer RQ2, although the characterization completeness and precision is notthe same for all the templates.
4.3 Discussion
The analyses of the above research questions revealed that differences exist in the perception of Domain Expertsand Knowledge Engineers, as well as in the evaluations they provide for the different templates.
In relation to the difference between Domain Experts’ and Knowledge Engineers’ perception, Figure 8reports their evaluations for each of the four factors. Surprisingly, by looking at the figure, it seems that DomainExperts judged template ease of understanding, usefulness and characterization precision more positively thanKnowledge Engineers. On the contrary, the completeness of the template characterization has been evaluatedslightly more positively by Knowledge Engineers. This result can be partially explained by considering thedifferent competencies and objectives guiding the two roles.
Domain Experts, indeed, are interested in the practical issues related to the domain they need to represent.They want a template that is useful to describe exactly what they need to describe, without caring whether termscould convey other meanings in other contexts. On the contrary, Knowledge Engineers are more focused on
9
Fig. 7: Subjective evaluation of the template characterizations completeness and precision
the formal part: to be easy to understand a term should be enough detailed and clear so as not to confuse itssemantics; applying the same criteria to the properties of the template characterizations, their perception of thecharacterization precision decreases as well.
Regarding the differences among templates, instead, in order to investigate the reasons behind the criticalitiescharacterizing some of them, we looked more in detail at how they have been built and to the ontologies they havebeen grounded to. Overall, two out of the seven templates are defined not only in domain-specific (DS) and mid-level (ML) ontologies, but also in foundational (F) ontologies and, among them, there is the one (Event) that isbetter perceived with respect to almost all the factors. Table 3 reports for each template, the typology of ontologyfrom which the corresponding entity has been taken as well as the percentage of characgterization attributes andproperties belonging to foundational, mid-level or domain specific ontologies, respectively. Finally it also reportsthe percentage of attributes and properties used to complete the characterization, as well as the percentage ofproperties common to more than two ontologies.
Template Entity Foundational Mid-level Domain Other shared by > 2Ontology specific ontologies
Role ML and DS 0 93.8% 0 6.3% 37.5%Artefact ML and DS 0 75% 8.3 % 16.7% 14.3%Actor ML and DS 23.5% 64.7% 0 11.8% 0
Organization ML and DS 10 % 40% 30% 20% 0Task ML and DS 0 40% 40% 20% 0Event F, ML and DS 7.7% 15.4% 46.2% 30.8% 0
Process F, ML and DS 8.3% 58.3% 25% 8.3% 0
Table 3: Distribution of the enterprise templates among Domain Experts and Knowledge Engineers
The table reveals that differences exist in the template composition (in terms of ontology categories used fortheir construction). Crossing these data with the users’ evaluations we found that the largest part of the characteri-zation of the templates that have been perceived as easier to understand and more precisely defined than the others(i.e., the Event and the Task templates) comes from specific ontologies. This result is also confirmed at statisticallevel. Indeed, we found a strong positive correlation 6 between the percentage of characterizations deriving from
6 We performed a correlation analysis applying the Pearson’s coefficient at 5 percent confidence level.
10
Fig. 8: Domain Experts and Knowledge Engineers evaluations
domain specific ontologies and the average evaluation of both ease of understanding and characterization preci-sion. In other terms, the higher the percentage of specific properties is and the higher the perception of the ease ofunderstanding and of the precision is. On the contrary, a negative correlation exists between the percentage of at-tributes and properties from mid-level ontologies and the average perception of ease of understanding, usefulnessand characterization precision
A second factor that seem to influence the perception of both ease of understanding and usefulness is thenumber of different source ontologies sharing the same property or attribute. Surprisingly, characterizations sharedby more than two ontologies seem to negatively affect the overall perception of the ease of understanding andusefulness of the template. This result can be connected, as well, to the negative correlation between too generalcharacterizations and their usability.
Hence, a good trade off to build templates able to find the right balance between rigorousness and usabilityand satisfy both the needs of Domain Experts and Knowledge Engineers, seems to be starting from founda-tional ontologies concepts and, passing through mid-level ontologies, characterizing the templates with concreteproperties obtained exploiting domain specific ontologies.
4.4 Threats to validity
Despite the rigorousness and the care used for conducting the survey and for analyzing the results some threatscan hinder the result validity. A first threat is due to the lack of the usage of the templates in a real-life task.Nevertheless, the templates were accompanied with a description of a possible scenario, thus helping the subjectsto figure out a possible practical use case for the template usage. A second threat is the subjectivity that has beenused for building the templates. The impact of this threat, however, is partially limited by the definition and theusage of a methodology the construction of the templates. Finally, the specificity of the domain considered is afurther threat to the validity of the study. On the other hand, the enterprise domain is a quite general domain inwhich Domain Experts are often involved.
11
5 Related Work
The use of templates is recently becoming an attractive means to hinder complexity and support domain experts inthe ontology authoring. Works dealing with templates based approaches in the ontology engineering field can bemainly classified in two categories: those focused on the ontology structure and those oriented to both structuraland domain knowledge.
There has been some effort done to generate templates. As in [20], authors presents a approach to supportontology engineers in modelling of ontology templates. They have demonstrated their approach with templatesfor ontology design patterns. This work is different from our approach as it is more focused toward knowledgeengineers and it needs to be extended, the usability aspect have not been studied and it has not been used for toplevel ontologies.
Also in MoKi [12] templates are mainly used for providing structural support to the ontology authoring. Thelightly-structured view, indeed, provide a semi-formal rendering of the OWL content in a predefined template tosupport domain experts access to the content of the ontology. The work of this paper will extend the tools likeMoKi with a set of templates grounded on existing foundational ontology and core ontologies instead of startingfrom scratch.
A slightly different approach is the one of Populous [14], a template based approach designed for domainexperts to gather knowledge that can be used to build ontologies. In Populous the authors advocate the use ofspreadsheets as templates to gather and organise information about concepts and their relationships as it providesa simple and intuitive form fill-in style of user interface. However, the main purpose of this approach is knowledgegathering, the population of the templates at the point of data entry, and not ontology building. In this paper, in-stead, we focus on the complexity of building usable and rigorous templates. Another approach that uses templatesto capture experts knowledge is presented in [13]. In order to acquire knowledge from experts this approach al-lows users to structure their knowledge by using concepts maps or the formalised knowledge structures defined astemplates. The templates in this approach corresponds to commonly used owl structures (e.g., subclass, instance,owl restrictions), they do not help very much in the overall structure of the ontology.
6 Conclusion
The paper shows, through a rigorous evaluation, that templates based on existing foundational and core ontologiesare able to reconcile usability and formality dimensions, ensuring ease of understanding and usefulness whileguaranteeing completeness and precision. As a side effect of our evaluation we provide some insights about therelation between the different types of ontologies used for building templates and their resulting perception byusers. We plan for the future to go ahead with this stream of work, further investigating the impact of different cat-egories of ontologies on template characteristics, thus (i) refining the proposed methodology for the constructionof ontology-based templates; and (ii) devising measures and criteria of their appropriateness to define templatesthat meet certain requirements.
References
1. Albano, A., Antonellis, V.D., Leva, A.D. (eds.): Computer-Aided Database Design: the DATAID approach. North-Holland(1985)
2. Bagni, D., Cappella, M., Pazienza, M.T., Stellato, A.: Congas: A collaborative ontology development framework based onnamed graphs. In: Proc. of the XIth Int. Conf. of the Italian Association for Artificial Intelligence. pp. 42–51. AI*IA ’09:,Springer-Verlag (2009), http://dx.doi.org/10.1007/978-3-642-10291-2_5
3. Boella, G., van der Torre, L.W.N.: A foundational ontology of organizations and roles. In: DALT. pp. 78–88 (2006)4. Bottazzi, E., Ferrario, R.: Preliminaries to a dolce ontology of organisations. Int. J. Business Process Integration and
Management 4(4), 225–238 (2009)5. Chiara Di Francescomarino, Ghidini, C., Rospocher, M.: Evaluating wiki-enhanced ontology authoring. In: Knowledge
Engineering and Knowledge Management: The 18th International Conference on Knowledge Engineering and KnowledgeManagement - EKAW2012, Galway, Ireland. vol. 7603, pp. 292–301. Springer Berlin Heidelberg (2012)
6. Cockburn, A.: Writing Effective Use Cases. Addison Wesley Publ. Comp. (2000)7. Dimitrova, V., Denaux, R., Hart, G., Dolbear, C., Holt, I., Cohn, A.G.: Involving domain experts in authoring owl ontolo-
gies. In: Proc. of ISWC 2008. pp. 1–16. Berlin (2008)8. Escalona, M., Reina, A., Torres, J., Mejıas, M.: Ndt a methodology to deal with the navigation aspect at the requirements
phase. In: OOPSLA Workshop: Aspect-Oriented Requirements Engineering and Architecture Design (2004)
12
9. Foundation, W.: Mediawiki. http://www.mediawiki.org (Accessed on 6 Nov 2011)10. Gangemi, A., Borgo, S., Catenacci, C., Lehmann, J.: Task taxonomies for knowledge content d07. Tech. rep., Metokis
Project (2004)11. Ghidini, C., Rospocher, M., Serafini, L.: Conceptual modeling in wikis: a reference architecture and a tool. In: The Fourth
International Conference on Information, Process, and Knowledge Management (eKNOW2012), Valencia, Spain. pp.128–135 (2012), http://www.thinkmind.org/download.php?articleid=eknow_2012_6_10_60015
12. Ghidini, C., Rospocher, M., Serafini, L.: Modeling in a wiki with moki: Reference architecture, implementation, andusages. International Journal On Advances in Life Sciences 4(3&4) (2012), to appear
13. Hayes, P., Eskridge, T.C., Saavedra, R., Reichherzer, T., Mehrotra, M., Bobrovnikoff, D.: Collaborative knowledge capturein ontologies. In: Proceedings of the 3rd international conference on Knowledge capture. pp. 99–106. K-CAP ’05, ACM,New York, NY, USA (2005), http://doi.acm.org/10.1145/1088622.1088641
14. Jupp, S., Horridge, M., Iannone, L., Klein, J., Owen, S., Schanstra, J., Wolstencroft, K., Stevens, R.: Populous:a tool for building owl ontologies from templates. BMC Bioinformatics 13(Suppl 1), S5 (2012), http://www.biomedcentral.com/1471-2105/13/S1/S5
15. Kop, C., Mayr, H.C.: Templates in domain modeling - a survey. In: The Evolution of Conceptual Modeling. pp. 21–41(2008)
16. Magro, D., Goy, A.: Towards a first ontology for customer relationship management. In: Chbeir, R., Badr, Y., Abraham,A., Laurent, D., Koppen, M., Ferri, F., Zadeh, L.A., Ohsawa, Y. (eds.) CSTST. pp. 637–643. ACM (2008)
17. Mark S. Fox and Michael Gruninger: Enterprise modeling. AI Magazine 19(3), 109–121 (1998)18. Masolo, C., Borgo, S., Gangemi, A., Guarino, N., Oltramari, A.: WonderWeb deliverable D18 ontology library (final).
Tech. rep. (2003)19. Masolo, C., Vieu, L., Bottazzi, E., Catenacci, C., Ferrario, R., Gangemi, A., Guarino, N.: Social roles and their descriptions.
In: KR. pp. 267–277 (2004)20. Parreiras, F.S., Groner, G., Walter, T., Staab, S.: A model-driven approach for using templates in owl ontologies. In:
EKAW. pp. 350–359 (2010)21. Ryszard, J., Parnas, D.L., Zucker, J.: Tabular representations in relational documents. In: Hoffman, D., Weiss, D. (eds.)
Software Fundamentials – Collected Papers by David Parnas, pp. 71–85. Addison Wesley Publishing Comp (2001)22. Tudorache, T., Falconer, S.M., Noy, N.F., Nyulas, C., Ustun, T.B., Storey, M.A.D., Musen, M.A.: Ontology development
for the masses: Creating icd-11 in webprotege. In: Cimiano, P., Pinto, H.S. (eds.) Knowledge Engineering and Managementby the Masses - 17th International Conference, EKAW 2010, Lisbon, Portugal, October 11-15, 2010. Proceedings. LectureNotes in Computer Science, vol. 6317, pp. 74–89. Springer (2010)
23. Uschold, M., King, M., House, R., Moralee, S., Zorgios, Y.: The enterprise ontology. The Knowledge Engi-neering Review 13, 31–89 (1998), http://reference.kfupm.edu.sa/content/e/n/the_enterprise_ontology__241225.pdf
24. Vieu, L., Borgo, S., Masolo, C.: Artefacts and roles: Modelling strategies in a multiplicative ontology. In: Proceedingof the 2008 conference on Formal Ontology in Information Systems: Proceedings of the Fifth International Conference(FOIS 2008). pp. 121–134. IOS Press, Amsterdam, The Netherlands, The Netherlands (2008), http://portal.acm.org/citation.cfm?id=1563953.1563967
25. Vohringer, J., Mayr, H.C.: Integration of schemas on the pre-design level using the kcpm-approach. In: Nilsson, A., Gustas,R., Wojtkowski, W., Wojtkowski, W., Wrycza, S., Zupancic, J. (eds.) Advances in Information Systems Bridging the Gapbetween Academia & Industry, pp. 623–634. Springer, Heidelberg (2006)
26. Wohlin, C., Runeson, P., Host, M., Ohlsson, M.C., Regnell, B., Wesslen, A.: Experimentation in software engineering: anintroduction. Kluwer Academic Publishers (2000)
7 Appendix A
Now, here we will provide a brief description of the template. The structure of the template is based on theontological building blocks. The subclass section represents the top level category to which concept belongs andif there is some part of relation exists then it will have a section for representing all the possible part of relations athat concept can have. Finally there is a property section to represent relation of that concept with other concepts.The tables below presents the characterization of templates along with the sources from where it is derived. Thereare columns stating the reason for modification if the property has been modified from the one existing in thesource ontology.
Organization: According to [4], Organizations are complex social entities, created and sustained by human agents.In other words organization is a complex entity linked to a group of people that are able to perform complexactivities which otherwise could not be accomplished by non coordinated individuals. Entities that populate theorganisation are: the agents who become part of organisation by accepting to play certain roles that are assigned tothem by organisation to act on its behalf, teams and departments. organisations arrange the agents into departments,teams and assign them the tasks to perfom.
13
14
Org
aniz
atio
nSo
urce
Ont
olog
ySo
urce
Ont
olog
ySo
urce
Ont
olog
yPr
oper
tyR
easo
nSo
urce
Ont
olog
yPr
oper
tyR
easo
n
Stru
ctur
alIn
form
atio
nR
ange
has
part
Team
Prel
imin
arie
sto
aD
OL
CE
onto
logy
ofor
gani
satio
nsha
stea
mTe
ams
Team
sar
e‘p
lura
lent
ities
’w
hich
exis
tins
ide
orga
nisa
tions
and
they
are
defin
edin
the
desc
ript
ion
ofth
eor
gani
satio
n.
has
part
Dep
artm
ent
AFo
unda
tiona
lO
ntol
ogy
ofO
rgan
izat
ions
and
Rol
es
hasp
artD
epar
tmen
t
orga
niza
tion
type
Cho
ose
orga
niza
tion
type
Stru
ctur
e,B
ehav
ior
and
Perf
orm
ance
inFo
r-Pr
ofit,N
onpr
ofit
and
Gov
ernm
ent
Org
aniz
atio
ns:A
nE
mpi
rica
lIn
vest
igat
ion *
haso
rgTy
peO
rgan
izat
ion
Type
Prop
ertie
sfu
llna
me
stri
ng(P
leas
eSp
ecif
y)Sc
hem
a.or
gL
egal
Nam
e-Te
xtT
heite
nsio
nto
rena
me
from
lega
lena
me
tofu
llna
me
was
toen
able
dom
ain
expe
rts
toun
ders
tand
the
inte
nded
mea
ning
behi
ndth
isat
trib
ute
e.g
FIA
T’s
full
nam
eis
Fabb
rica
Ital
iana
Aut
omob
iliTo
rino
15
Org
aniz
atio
nSo
urce
Ont
olog
ySo
urce
Ont
olog
ySo
urce
Ont
olog
yPr
oper
tyR
easo
nSo
urce
Ont
olog
yPr
oper
tyR
easo
n
Est
ablis
hed
onda
te(P
leas
eSp
ecif
y)Sc
hem
a.or
gfo
undi
ngD
ate-
Dat
eIn
the
liter
atur
e,th
efr
eque
ntly
used
term
fors
uch
prop
erty
was
Foun
ded
onor
Est
ablis
hed
on.S
ow
ech
oose
tore
nam
eit
toes
tabl
ishe
don
.The
inte
nded
sem
antic
behi
ndth
ese
term
sar
eth
esa
me.
has
size
stri
ng(P
leas
eSp
ecif
y)
http
://w
ww
.ak
tors
.org
/ont
olog
y/p
orta
l
has-
size
-O
rgan
izat
ion-
Size
Allo
wth
epe
ople
toch
oose
from
asp
ecifi
cse
tofs
izes
but
the
obje
ctiv
efo
rren
amin
gis
toal
low
user
sto
ente
rthe
irow
nsp
ecifi
edsi
zes
has
loca
tion
stri
ng(P
leas
eSp
ecif
y)A
nor
gani
zatio
non
tolo
gySc
hem
a.or
glo
catio
n-pl
ace
orpo
stal
addr
ess
The
sepr
oper
ties
are
usin
gC
lass
esto
desc
ribe
the
loca
tion
ofan
orga
nisa
tion
butw
eha
vead
apte
dto
use
data
type
prop
ertie
sfo
rrep
rese
ntin
gsu
chre
latio
nsbe
caus
ew
ear
eno
tcon
side
ring
DO
LC
Equ
ality
and
Abs
trac
tcat
egor
yin
oura
ppro
ach
has
site
—si
teT
his
prop
erty
isun
dert
heca
tego
ryof
Loc
atio
nso
we
cons
ider
edus
ing
gene
ral
prop
erty
Loc
atio
nan
dth
esa
me
rule
ofda
taty
pepr
oper
tyap
plie
she
re
has
dire
ctor
Pers
onA
nor
gani
zatio
non
tolo
gyht
tp://
ww
w.a
ktor
s.o
rg/o
ntol
ogy
/por
tal
foaf
:Age
nt-h
ead
of-O
rgD
irec
torh
ere
repr
esen
tsth
epe
rson
who
isth
ehe
adof
anor
gani
satio
n
Org
–hea
ded
by–
Affi
liate
dpe
rson
has
cont
act
pers
onPe
rson
Self
driv
en
crea
tes/
prod
uces
Art
efac
tFr
omPh
ysic
alA
rtef
acts
toPr
oduc
tsA
rtef
act-
crea
ted
by-a
gent
orpr
oces
sA
gent
here
also
cove
rsth
eSo
cial
Age
nts
like
Org
anis
atio
ns
16
Org
aniz
atio
nSo
urce
Ont
olog
ySo
urce
Ont
olog
ySo
urce
Ont
olog
yPr
oper
tyR
easo
nSo
urce
Ont
olog
yPr
oper
tyR
easo
n
actt
hrou
ghA
ctor
Prel
imin
arie
sto
aD
OL
CE
onto
logy
ofor
gani
satio
nsA
ctby
-Act
ors
Act
ors
play
ing
cert
ain
role
sin
Org
anis
atio
n
prov
ides
/sel
lsPr
oduc
t/Se
rvic
e
Tow
ards
anO
ntol
ogic
alFo
unda
tion
for
Serv
ices
Scie
nce
Serv
ice
prod
uctio
n-pe
rfor
med
by-s
ervi
cepr
oduc
er
Serv
ice
prod
ucer
can
bean
Org
anis
atio
nor
aSi
ngle
phys
ical
agen
tand
serv
ice
can
beco
nsid
ered
asa
prod
uctf
oran
orga
nisa
tion
uses
Art
efac
tse
lfdr
iven
part
icip
ates
inE
vent
sD
OL
CE
D18
End
uran
tPa
rtic
ipat
ein
Perd
uran
t
Act
oris
anE
ndur
anta
ndE
vent
isa
Perd
uran
t
empl
oyes
Pers
onse
lfdr
iven
carr
you
tPr
ojec
t
The
Con
cept
of‘P
roje
ct’:
APr
opos
alfo
raU
nify
ing
Defi
nitio
n
Org
-man
ages
-Pr
ojec
tA
proj
ecti
sa
wor
ksy
stem
desi
gned
togo
outo
fexi
sten
ceaf
terp
rodu
cing
apa
rtic
ular
prod
ucto
rapr
ojec
tas
anor
gani
zatio
nalu
nitt
hats
olve
sa
uniq
uean
dco
mpl
exta
skba
sed
onth
ese
desc
ript
ions
we
thin
kor
gani
saito
nsar
ere
spon
sibl
efo
rcar
ryin
gou
ttho
sepr
ojec
ts.
Fina
nces
Proj
ect
self
driv
en
17
Event: In DOLCE, Events are subclasses of perdurants and they are the things that happen in time and extendin time by accumulating different temporal parts. The relation between Endurant and Perdurant is that of ’par-ticipation’ as an endurant ’lives’ in time by participating in some perdurant(s).Events have a start time end timeand the major participants of events are objects and actors. Each event has a associated process attached to it e.g.,Conference itself is an event and the associated process is the arrangement of conference.
In a organisational setting there are different events that occur. Let’s suppose that an Organisation ”FIAT”arranges a Product launch event ”Fiat 500 UK launch event” at the London Eye on Monday 12-Dec-2012 and itwill end on 14-Dec-2012. The event type of this event will be a organizational event, more specifically a ProductLaunch event. In order to explain to people what it is about and attract more people to this event a description isprovided to the participants. To arrange this event successfully there is a process specified by the FIAT to followand it also designate the people who are responsible for organising event. These people play the role of Eventmanagers in the organisation. They are responsible for planing and executing this event. In this events there is aevent host to explain the functionalities of this new product and guide the participants through different stages ofthe event.
18
Eve
ntSo
urce
Ont
olog
ySo
urce
Ont
olog
ySo
urce
Ont
olog
yPr
oper
tyR
easo
nSo
urce
Ont
olog
yPr
oper
tyR
easo
n
Prop
ertie
sR
ange
full
nam
est
ring
(Ple
ase
Spec
ify)
Self
Driv
en
has
loca
tion
stri
ng(P
leas
eSp
ecif
y)Sc
hem
a.or
ght
tp://
ww
w.
akto
rs.o
rg/
onto
logy
/po
rtal
loca
tion-
post
alad
dres
s,Pl
ace
The
sepr
oper
ties
are
usin
gC
lass
esto
desc
ribe
the
loca
tion
ofan
even
tbut
onth
eot
her
hand
we
have
adap
ted
tous
eda
taty
pepr
oper
ties
for
repr
esen
ting
such
rela
tions
beca
use
we
are
notc
onsi
deri
ngD
OL
CE
qual
ityan
dA
bstr
act
cate
gory
inou
rapp
roac
h
hasl
ocat
ion-
Loc
atio
nSa
me
asth
eot
herr
easo
n
star
tdat
eda
te(P
leas
eSp
ecif
y)Sc
hem
a.or
gSt
artd
ate–
Dat
een
dda
teda
te(P
leas
eSp
ecif
y)Sc
hem
a.or
gE
ndda
te–D
ate
even
ttyp
est
ring
(Ple
ase
Spec
ify)
Sche
ma.
org
Eve
ntty
pes
Eve
ntpa
geha
sse
ctio
nfo
rsh
owin
gSp
ecifi
cE
vent
type
sev
ent
desc
ript
ion
stri
ng(P
leas
eSp
ecif
y)Se
lfD
riven
19
Eve
ntSo
urce
Ont
olog
ySo
urce
Ont
olog
ySo
urce
Ont
olog
yPr
oper
tyR
easo
nSo
urce
Ont
olog
yPr
oper
tyR
easo
n
has
asso
ciat
edPr
oces
s
Tow
ards
anO
ntol
ogic
alFo
unda
tion
for
Serv
ices
Scie
nce
SUPE
RPr
ojec
tD
eliv
erab
le1.
1Se
rvic
epr
oces
sim
plem
ents
serv
ice
serv
ices
are
tem
pora
lent
ities
(eve
nts)
and
Serv
ice
proc
ess
isa
Proc
ess.
Soea
chev
enth
asa
asso
ciat
edpr
oces
sth
atim
plem
ents
it.T
here
isan
othe
rw
ork
bySU
PER
proj
ect
Del
iver
able
1.1,
that
prop
oses
Upp
erPr
oces
sO
ntol
ogy
(UPO
).A
ccor
ding
toth
isw
ork
the
conc
eptp
lan
ishi
ghle
vel
abst
ract
ion
ofpr
oces
san
deq
uiva
lent
toa
proc
ess
inB
PMO
.Thi
son
tolo
gyst
ates
the
prop
erty
ofas
soci
ated
even
t.pa
rtic
ipan
tPe
rson
/Org
aniz
atio
nD
OL
CE
D18
Tow
ards
afir
stO
ntol
ogy
forC
usto
mer
Rel
atio
nshi
pM
anag
emen
t
End
uran
tPa
rtic
ipat
ein
Perd
uran
t
Act
oris
End
uran
tand
Eve
ntis
aPe
rdur
ant
acto
rpla
yrol
ein
even
t(m
aybe
aC
ompa
ny)
acto
rtha
tpla
yed
the
role
ofve
ndor
inth
eev
ent(
may
bea
Com
pany
)
even
thos
tPe
rson
Self
Driv
enor
gani
zers
Pers
onSe
lfD
riven
20
Role: In an organisation a role can be played by different entities or a same entity can play different roles. Theroles are defined by the norms as they regulate the behaviour of actor playing the roles, dictating what they areallowed to do (directly or indirectly), what they are obliged to do. This means that the actor who decides to becomepart of an organisation must agree to undertake all the rights and duties connected to the role that (s)he will playwithin the organisation [4]. As mentioned by [16] roles not on themselves decide what tasks to perform but theyare assigned to them by other roles e.g., an employee can be given instruction to perform a task by its director.
Accordig to [4][19] a role can specialise another role like ’Production manager’ is a spacialisation of the role’manager’ and there is a precise hierarchy of roles , a forced path to follow in order to reach certain position orplay a certain role. It means that in order to a role the actor must have played a role that is required by the currentrole. Being a ’Production manager’ a person has certain right (e.g., request information from other roles, staff thedepartment, access organisation information and data ), fullfill different duties (e.g., compile the production plan) and carry out the assigned task (decision on production material, order materials and supplies, plan and developnew products or production processes).
21
Rol
eSo
urce
Ont
olog
ySo
urce
Ont
olog
ySo
urce
Ont
olog
ySo
urce
Ont
olog
yPr
oper
tyR
easo
nSo
urce
Ont
olog
yPr
oper
ty
Rea
son
Prop
ertie
sR
ange
full
nam
est
ring
(Ple
ase
Spec
ify)
self
driv
en
has
righ
tsst
ring
(Ple
ase
Spec
ify)
Prel
imin
arie
sto
aD
OL
CE
onto
logy
ofor
gani
satio
ns
Soci
alR
oles
and
thei
rDes
crip
tions
has
righ
ts
has
dutie
sst
ring
(Ple
ase
Spec
ify)
Prel
imin
arie
sto
aD
OL
CE
onto
logy
ofor
gani
satio
ns
Soci
alR
oles
and
thei
rDes
crip
tions
Task
Taxo
nom
ies
for
Kno
wle
dge
Con
tent
D07
has
dutie
s
repo
rts
toR
ole
An
orga
niza
tion
onto
logy
(foa
f:A
gent
org:
Post
)Rep
orts
to(f
oaf:
Age
ntor
g:Po
st)
org:
Post
whi
chre
pres
ents
som
epo
sitio
nin
the
orga
niza
tion
that
may
orm
ayno
tbe
curr
ently
fille
d.Po
sts
enab
lere
port
ing
stru
ctur
es.
Post
sca
nre
port
toot
herP
osts
.is
role
ofA
rtef
act/
Act
orA
RT
EFA
CT
SIN
FOR
MA
LO
NTO
LO
GY
Soci
alR
oles
and
thei
rDes
crip
tions
Prel
imin
arie
sto
aD
OL
CE
onto
logy
ofor
gani
satio
ns
Art
efac
t–Pl
ays
the–
Rol
eA
gent
play
sro
le
has
task
Task
Task
Taxo
nom
ies
forK
now
ledg
eC
onte
ntD
07
Act
orha
stas
kta
skA
ctor
sar
epl
ayin
gro
les
inth
eO
rgan
isat
ion
soth
eta
sks
are
assi
gned
toth
ose
role
sSp
ecia
lize
byR
ole
Prel
imin
arie
sto
aD
OL
CE
onto
logy
ofor
gani
satio
ns
Soci
alR
oles
and
thei
rDes
crip
tions
Rol
eSp
ecia
lize
Rol
e
requ
ires
Rol
ePr
elim
inar
ies
toa
DO
LC
Eon
tolo
gyof
orga
nisa
tions
Soci
alR
oles
and
thei
rDes
crip
tions
Rol
eR
equi
res
Rol
e
22
Artefact: Artefacts are the entities created for some purpose or use and they are the result of agent’s intentionality.They are constituted by non Agentive Physical objects or amounts of matter. Artefacts can also play roles e.g., Ahouse is an artefact which can play the role of being someones home. The key element that distinguishes productsfrom artefacts is the exchange purpose[24]. In an organisation, artefacts (Products & Documents) are producedby agents playing certain role or by the project carried out by organisation, or as a output of a some processin organisation. For example, consider the document ’Product description’ describing the product details can beconsidered as an artefact produced by an agent who is playing the role of ’technical writer’ in the organisation.
23
Art
efac
tSo
urce
Ont
olog
ySo
urce
Ont
olog
ySo
urce
Ont
olog
ySo
urce
Ont
olog
yPr
oper
tyR
easo
nSo
urce
Ont
olog
yPr
oper
tyR
easo
n
Prop
ertie
sR
ange
full
nam
est
ring
(Ple
ase
Spec
ify)
Self
Driv
en
has
Idst
ring
(Ple
ase
Spec
ify)
Self
Driv
en
purp
ose
stri
ng(P
leas
eSp
ecif
y)
AR
TE
FAC
TS
INFO
RM
AL
ON
TOL
OG
Y
Art
efac
t—Pu
rpos
e—(R
ange
notd
efine
d)
Con
side
rthe
agen
t’s
inte
ntio
nsun
derl
ying
the
crea
tion
ofan
arte
fact
.The
rear
etw
oas
pect
s:th
ein
tent
ion
toge
tan
entit
yfo
rso
me
purp
ose
orus
e,an
dth
ein
tent
ion
toph
ysic
ally
mod
ify
orpr
oces
sso
me
pree
xist
ing
entit
yto
“pro
duce
”th
ear
tefa
ctha
sde
scri
ptio
nst
ring
(Ple
ase
Spec
ify)
Beh
avio
rofa
Tech
nica
lArt
ifact
:A
nO
ntol
ogic
alPe
rspe
ctiv
ein
Eng
inee
ring
has
desc
ript
ion—
beha
vior
alan
dfu
nctio
nal
desc
ript
ions
has
role
Rol
eA
rtef
acts
and
Rol
es:M
odel
ling
Stra
tegi
esin
aM
ultip
licat
ive
Ont
olog
y
From
Phys
ical
Art
efac
tsto
Prod
ucts
AR
TE
FAC
TS
INFO
RM
AL
ON
TOL
OG
YPl
ays
the–
Rol
e
deve
lope
dby
Pers
on/
Org
aniz
atio
nFr
omPh
ysic
alA
rtef
acts
toPr
oduc
ts
AR
TE
FAC
TS
INFO
RM
AL
ON
TOL
OG
Y
crea
tedb
y–ag
ento
rpr
oces
sT
hede
finiti
onof
Age
ntal
soin
clud
esso
cial
agen
ts;
Org
aniz
atio
npr
oduc
edby
Proj
ect/
Proc
ess
The
Con
cept
of‘P
roje
ct’:
APr
opos
alfo
raU
nify
ing
Defi
nitio
n
From
Phys
ical
Art
efac
tsto
Prod
ucts
Prod
uce
duri
ng–
Proj
ect
Art
ifact
sdo
cum
entin
gag
reem
ents
ina
proj
ect:
Req
uire
men
tspe
cific
atio
n,pr
ojec
tpla
n,pr
oduc
tar
chite
ctur
e,m
inut
esfr
omst
eeri
ngco
mm
ittee
mee
ting,
crea
tedb
y–ag
ento
rpr
oces
s
24
Actor: Actors are the physical agents in DOLCE ontology having some affiliation to an organization by playingsome role in it . They are the owners of some processes and perform activities designated to them.
25
Act
orSo
urce
Ont
olog
ySo
urce
Ont
olog
ySo
urce
Ont
olog
ySo
urce
Ont
olog
yPr
oper
ty
Rea
son
Sour
ceO
ntol
ogy
Prop
erty
Rea
son
Prop
ertie
sR
ange
full
nam
est
ring
(Ple
ase
Spec
ify)
Self
Driv
en
acto
rid
stri
ng(P
leas
eSp
ecif
y)
Self
Driv
en
invo
lved
inPr
ojec
tT
heC
once
ptof
‘Pro
ject
’:A
Prop
osal
fora
Uni
fyin
gD
efini
tion
peop
leIn
volv
edin
Proj
ect
peop
leco
rres
pond
tope
rson
and
pers
onar
eA
ctor
sin
our
dom
ain
Dep
artm
ent
hasm
embe
rro
les
Rol
esar
epl
ayed
byA
gent
s.So
we
reco
rdth
ein
form
atio
nat
the
leve
lofA
gent
sno
tatR
oles
mem
bero
fTe
am/
Dep
artm
ent
Prel
imin
arie
sto
aD
OL
CE
onto
logy
ofor
gani
satio
ns
AFo
unda
tiona
lO
ntol
ogy
ofO
rgan
izat
ions
and
Rol
es
hasm
embe
r-ag
ent
depe
ndon
the
role
(s)p
laye
dby
thei
rmem
bers
has
role
Rol
ePr
elim
inar
ies
toa
DO
LC
Eon
tolo
gyof
orga
nisa
tions
AFo
unda
tiona
lO
ntol
ogy
ofO
rgan
izat
ions
and
Rol
es
Age
ntpl
ays
role
affil
iate
dO
rgan
izat
ion
Prel
imin
arie
sto
aD
OL
CE
onto
logy
ofor
gani
satio
ns
Age
ntiv
eph
ysic
alO
bjec
t—A
ffilia
ted—
Org
Act
oris
aA
gent
ive
obje
ct
resp
onsi
ble
for
Proc
ess
The
ente
rpri
seon
tolo
gyA
ctor
resp
onsi
blef
orA
CT
IVIT
Y
We
cons
ider
Act
ivity
asa
Proc
ess
actf
orO
rgan
izat
ion
Prel
imin
arie
sto
aD
OL
CE
onto
logy
ofor
gani
satio
ns
phys
ical
agen
tac
tsfo
rorg
exec
ute
Serv
ice/
Task
Tow
ards
anO
ntol
ogic
alFo
unda
tion
for
Serv
ices
Scie
nce
Task
Taxo
nom
ies
forK
now
ledg
eC
onte
ntD
07
Age
ntex
ecut
esse
rvic
ean
agen
tis
expl
icitl
yco
mm
itted
togu
aran
tee
the
exec
utio
nof
som
ety
peof
actio
n
Act
orha
stas
kta
sk
26
Act
orSo
urce
Ont
olog
ySo
urce
Ont
olog
ySo
urce
Ont
olog
ySo
urce
Ont
olog
yPr
oper
ty
Rea
son
Sour
ceO
ntol
ogy
Prop
erty
Rea
son
part
icip
ate
Eve
ntD
OL
CE
D18
End
uran
tPa
rtic
ipat
ein
Perd
uran
t
Act
oris
anE
ndur
anta
ndE
vent
isa
Perd
uran
t
part
icip
ate
Proc
ess
DO
LC
ED
18E
ndur
ant
Part
icip
atei
nPe
rdur
ant
Act
oris
anE
ndur
anta
ndE
vent
isa
Perd
uran
t
prod
uce
Art
efac
tFr
omPh
ysic
alA
rtef
acts
toPr
oduc
tsT
heC
once
ptof
‘Pro
ject
’:A
Prop
osal
fora
Uni
fyin
gD
efini
tion
AR
TE
FAC
TS
INFO
RM
AL
ON
TOL
OG
Y
Art
efac
tcr
eate
dby
agen
tor
proc
ess
27
Task: Tasks are courses used to sequence processes or other perdurants that can be under the control of a planner.Organizations act by deputing tasks to roles performed by physical agents[10].
When a process is run, tasks go through one of these states7:
– Waiting means that the Task is not yet Ready.– In Process means that a actor has acquired the Task.– Done means that a actor has completed the Task.– Skipped means that the Task will never be Ready, but tasks that depend on the Skipped task can become
Ready.– Activated means that the Task has been automatically started without being associated with a actor. This state
applies only to Tasks with subtasks.
7 http://download.oracle.com/docs/cd/E13182_01/elink/bpo/bpo11/bpwrkflw.htm
28
Task
Sour
ceO
ntol
ogy
Sour
ceO
ntol
ogy
Sour
ceO
ntol
ogy
Prop
erty
Rea
son
prop
ertie
s,R
ange
full
nam
est
ring
(Ple
ase
Spec
ify)
Self
Driv
en
has
desc
ript
ion
stri
ng(P
leas
eSp
ecif
y)Se
lfD
riven
due
date
Dat
e(P
leas
eSp
ecif
y)B
usin
ess
Proc
ess
Des
ign
Ove
rvie
wD
ueda
te-d
ate
prio
rity
stri
ng(P
leas
eSp
ecif
y)B
usin
ess
Proc
ess
Des
ign
Ove
rvie
wPr
iori
ty-S
trin
g
stat
usst
ring
(wai
ting,
done
,...)
Bus
ines
sPr
oces
sD
esig
nO
verv
iew
Stat
us–d
iffer
ents
tatu
ses(
Rea
dy,w
aitin
g)ha
ssu
bta
skTa
skTa
skTa
xono
mie
sfo
rK
now
ledg
eC
onte
ntD
07Ta
skde
com
pose
din
toSU
BTA
SKS
ista
skof
Rol
eTa
skTa
xono
mie
sfo
rK
now
ledg
eC
onte
ntD
07A
ctor
hast
ask
task
sequ
ence
Proc
ess
Task
Taxo
nom
ies
for
Kno
wle
dge
Con
tent
D07
SKIi
ngw
ithD
OL
CE
:tow
ard
aSc
ienc
eK
now
ledg
ese
quen
ced-
by:T
ask
Task
sar
eco
urse
sth
atar
e(m
ostly
)use
dto
sequ
ence
activ
ities
spec
ializ
edby
Task
Task
Taxo
nom
ies
for
Kno
wle
dge
Con
tent
D07
task
spec
ializ
esta
skor
soci
alob
ject
spec
ializ
edby
soci
alob
ject
Task
sar
ea
kind
ofso
cial
obje
ct.
29
Process: Process is defined as transforming input elements in to some new output element. In organizations eachprocess has a responsible actor playing a role in it and organizes tasks into a structure that determines when eachtask is to be performed.
30
Proc
ess
Sour
ceO
ntol
ogy
Sour
ceO
ntol
ogy
Sour
ceO
ntol
ogy
Prop
erty
Rea
son
Prop
ertie
sR
ange
full
nam
est
ring
(Ple
ase
Spec
ify)
Self
Driv
enha
ssu
bpr
oces
sPr
oces
sA
nA
ppro
ach
toth
eD
efini
tion
ofa
Cor
eE
nter
pris
eO
ntol
ogy:
CE
O
SUPE
RPr
ojec
tD
eliv
erab
le1.
1ha
ssu
bpl
anSU
PER
proj
ectD
eliv
erab
le1.
1,th
atpr
opos
esU
pper
Proc
ess
Ont
olog
y(U
PO).
Acc
ordi
ngto
this
wor
kth
eco
ncep
tpla
nis
high
leve
labs
trac
tion
ofpr
oces
san
deq
uiva
lent
toa
proc
ess
inB
PMO
.Als
ose
eth
etr
ansf
orm
atio
nse
ctio
nby
CE
Ose
quen
ced
byTa
skTa
skTa
xono
mie
sfo
rK
now
ledg
eC
onte
ntD
07
SKIi
ngw
ithD
OL
CE
:tow
ard
aSc
ienc
eK
now
ledg
e
sequ
ence
d-by
-Ta
skTa
sks
are
cour
ses
that
are
(mos
tly)u
sed
tose
quen
ceac
tiviti
es
uses
Res
ourc
esT
heR
ole
ofFo
unda
tiona
lO
ntol
ogie
sin
Man
ufac
turi
ngD
omai
nA
pplic
atio
ns
AC
ore
Ont
olog
yfo
rBus
ines
sPr
oces
sA
naly
sis
Bus
ines
sA
ctiv
ityU
ses
Pers
iste
ntE
ntity
Proc
ess
isa
subc
lass
ofB
usin
ess
Act
ivity
and
Pers
iste
ntE
ntity
can
becl
assi
fied
asPh
ysic
alen
tity,
Non
phys
ical
Ent
ityan
dA
gent
s.In
Foun
datio
nalO
ntol
ogie
sin
Man
ufac
turi
ngD
omai
nth
eyha
vein
trod
uced
the
conc
epto
fres
ourc
esth
atco
rres
pond
sto
the
Age
nts
and
phyi
cale
ntiti
tes
buti
nou
ropi
nion
non
phys
ical
entit
eslik
eso
ftw
are
can
also
beco
nsid
ered
asre
sour
ce.
asso
ciat
edto
Eve
ntTo
war
dsan
Ont
olog
ical
Foun
datio
nfo
rSe
rvic
esSc
ienc
e
SUPE
RPr
ojec
tD
eliv
erab
le1.
1Se
rvic
epro
cess
impl
emen
tsa
serv
ice
serv
ices
are
tem
pora
lent
ities
(eve
nts)
and
Serv
ice
proc
ess
isa
Proc
ess.
Soea
chev
enth
asa
asso
ciat
edpr
oces
sth
atim
plem
ents
it.T
here
isan
othe
rwor
kby
SUPE
Rpr
ojec
tD
eliv
erab
le1.
1,th
atpr
opos
esU
pper
Proc
ess
Ont
olog
y(U
PO).
Acc
ordi
ngto
this
wor
kth
eco
ncep
tpla
nis
high
leve
lab
stra
ctio
nof
proc
ess
and
equi
vale
ntto
apr
oces
sin
BPM
O.
Thi
son
tolo
gyst
ates
the
prop
erty
ofas
soci
ated
even
t.pa
rtic
ipan
tA
ctor
DO
LC
ED
18E
ndur
ant
Part
icip
ate
inPe
rdur
ant
Act
oris
anE
ndur
anta
ndit
Proc
ess
that
isa
Perd
uran
t
prod
uce
outp
utA
rtef
act
From
Phys
ical
Art
efac
tsto
Prod
ucts
Art
efac
tcre
ated
byag
ento
rpro
cess
we
choo
seto
mod
ify
the
crea
ted
byto
prod
uce
outp
utbe
caus
eas
apr
oces
sha
sa
inpu
tand
itpr
oduc
eou
tput
.The
outp
utof
apr
oces
sca
nbe
apr
oduc
tora
Doc
umen
tand
thes
eco
ncep
tsar
eco
nsid
ered
asar
tefa
cts.
has
resp
onsi
ble
Act
orT
heen
terp
rise
onto
logy
Act
orre
spon
sibl
efor
AC
TIV
ITY
We
cons
ider
Act
ivity
asa
Proc
ess
31