Enseñando Modelado Conceptual en Humanidades y Ciencias Sociales

Cesar Gonzalez-Perez
cesar.gonzalez-perez@incipit.csic.es
CSIC
Patricia Martín-Rodilla
patricia.martin-rodilla@incipit.csic.es
CSIC

Comunicación larga
Humanidades digitales – pedagogía y currículo


Motivation and Context 

Over the years, at Incipit CSIC we have observed that archaeologists, historians, anthropologists, architects and other specialists working on cultural heritage often develop complex information models about the reality they study [2] [3]. However, these models are usually highly informal and expressed in natural language or very loose formalisms; it is the case of e.g. Harris matrices [7] in archaeology or lexical thesauri in most branches of the digital humanities. Simple modelling needs may be satisfied by approaches like these, but the ever increasing challenges of today’s interdisciplinary research projects and large-scale collaborations often mean that very complex fragments of reality are to be modelled; in these situations, humanities and social science specialists need to collect, transform and manage information of such a complexity that more advanced technologies are necessary. Conceptual modelling has been used in software engineering and related disciplines to develop models of highly complex portions of reality with great success [12], even in the particular domain of cultural heritage [1] [6].

Developing a conceptual model helps us understand the portion of reality we are dealing with by removing the unnecessary detail and allowing us to focus on what is relevant at each moment. Thus, we can explore complex realities more easily through simpler and more manageable models. In addition, conceptual modelling helps us communicate our understating of a portion of reality, especially when people of different disciplines and backgrounds are involved, by creating a common shared ontological space where meaningful discussion can take place. Unfortunately, conceptual modelling has been historically appropriated by software engineers, despite the fact that the connection between the two is more accidental than essential. We believe that any humanities or social sciences professional should be capable of creating their own conceptual models if given a good enough modelling language and the necessary training, and with this premise in mind we developed the ConML [8] conceptual modelling language.

ConML was designed to be affordable to non-experts in information technologies, and to specifically address modelling needs that are rarely considered in natural sciences but are however crucial in the humanities and social sciences, such as subjectivity, temporality or vagueness [4]. Also, ConML is oriented towards the creation of people-oriented conceptual models rather than computer-oriented implementation models like other languages such as UML [10]. ConML has been used extensively inside Incipit CSIC; for example, to design the Cultural Heritage Abstract Reference Model (CHARM) [5] [9], and is starting to be used by external independent parties as well [11]. The following sections describe our experiences using ConML as an infrastructure to teach conceptual modelling to humanities and social sciences specialists.

Teaching Approach 

Given the success that we observed in our internal use of ConML, we soon decided to teach others to use it, and in 2010 started designing an education programme on conceptual modelling and cultural heritage through CSIC’s Postgraduate School. The underpinning hypothesis was that it is possible for humanities and social science specialists with no previous exposure to software or knowledge technologies to acquire operational skills in conceptual modelling in just one week.

The first course took place in Santiago de Compostela (Spain) in May 2011 over 5 days, taking 30 hours of contact teaching to cover basic object-oriented modelling aspects such as the concepts of object, class, attribute, association and generalisation, as well as more advanced topics such as the modelling of vagueness, modularity and model refactoring. This course was targeted to cultural heritage specialists with no previous knowledge of information technologies, and gathered 19 students with backgrounds in architecture, geography and archaeology. A similar course took place in 2012. Also in 2012, slightly customised versions of the course were taught in Vitoria-Gasteiz (Spain) and Olavarría (Argentina). In 2014 and 2015, the course was taught as part of a Master’s degree in Archaeology in collaboration with the University of Santiago de Compostela. Additional contents were introduced in newer editions of the course, such as modelling patterns, modelling process, or temporality and subjectivity modelling. The following table summarises the course editions so far.

EditionPlaceDatesNumber of students
1Santiago de Compostela, SpainMay 201119
2Santiago de Compostela, SpainApril 201210
3Vitoria-Gasteiz, SpainJune 201212
4Olavarría, Argentina August 20129
5Santiago de Compostela, SpainFebruary-April 20148
6Santiago de Compostela, SpainFebruary-April 201510
Total68

The following section describes this as well as other results obtained.

Outcomes 

For every course edition, students were evaluated through participation, a mini-project they developed during the course, and a final quiz, usually weighing 10/35/55 respectively. Scores were given on a scale from 0 to 10, with the pass at 5. The following table shows the minimum, average and maximum scores achieved by students for each course edition.

EditionMinimumAverageMaximum
146,69
257,19
357,910
467,08
557,310
>667,38
Overall47,110

In addition, on the last day of every course, an evaluation questionnaire was distributed to students so feedback about the course could be obtained. The questions included were as follows:

  1. Contents are interesting
  2. Contents have a high academic standard
  3. Explanations are clear and sufficient
  4. Communication from teachers is good
  5. Visual support (whiteboard, projection) is properly used
  6. The pace of the course is suitable
  7. The duration of classes and breaks is adequate
  8. Theory is adequately illustrated by examples and applications
  9. The exercises are appropriate to understand the theory and acquire the target skills
  10. The exercises are adequate in number and difficulty level
  11. Teachers provide good orientation, guidance and supervision
  12. Assessment mechanisms are appropriate and fair
  13. The course is about what I expected

Students were asked to mark on a 4-point Likert scale whether they strongly agreed (4), agreed (3), disagreed (2) or strongly disagreed (1) with each question. The following table shows the minimum, average and maximum scores for each question across course editions.

QuestionMinimumAverageMaximum
123,54
223,54
323,54
413,74
523,54
623,44
723,24
823,34
923,34
1023,44
1113,44
1223,24
1312,74

Finally, a call is made to students on the last day of the course to keep us informed if they apply the skills they have acquired during the course to their projects or future work. So far, we have collected evidence of 6 students doing this out of 68 (8.8%).

Discussion and Conclusion 

Academic results over the six editions of the course show that, in general, students successfully acquire the intended modelling skills. Only 1 student out of 68 (1.5%) ever failed the course, average scores are stable above 7, and most editions yield students hitting top scores of or above 9. This clearly supports the hypothesis that, in fact, it is possible for humanities and social science specialists with no previous exposure to software or knowledge technologies to acquire operational skills in conceptual modelling in just one week.

In addition, the course seems to be very well received by students, who systematically evaluate it above 3 (agree) for all questions, with the exception of question 13 “The course is about what I expected”. It is indeed difficult to adequately convey what the course is about to potential students, given the large disciplinary differences between their backgrounds and the contents of the course. This is an area on which we are working towards future editions of the course.

Another area of improvement is that of the actual incorporation of the acquired skills to the repertoire of practices that are deployed by cultural heritage professionals at work. We trust that the course “helps to organise your own mind” when dealing with information, as one student put it on the feedback questionnaire, and this alone is of great value. However, specific tools and techniques are needed to facilitate adoption and productive application, not only for the sake of individuals, but also for the benefit of work groups and interdisciplinary teams that are becoming more and more prevalent in the digital humanities.

References 

[1] CIDOC, 2011. The CIDOC Conceptual Reference Model (web site). Accessed on 26 November 2012. http://www.cidoc-crm.org/

[2] Gonzalez-Perez, C., 2002. Sistemas de Información para Arqueología: Teoría, Metodología y Tecnologías. BAR International Series. Vol. 1015. Oxford (UK). Archaeopress.

[3] Gonzalez-Perez, C., 2012. A Conceptual Modelling Language for the Humanities and Social Sciences, in Sixth International Conference on Research Challenges in Information Science (RCIS 2012), C. Rolland, J. Castro, and O. Pastor, (eds.). IEEE Computer Society. Valencia (Spain). 396-401.

[4] Gonzalez-Perez, C., 2013. Modelling Temporality and Subjectivity in ConML, in 7th IEEE International Conference on Research Challenges in Information Science (RCIS 2013), R. Wieringa and S. Nurcan (eds.). IEEE Computer Society. Paris (France).

[5] Gonzalez-Perez, C., P. Martín-Rodilla, C. Parcero-Oubiña, P. Fábrega-Álvarez, and A. Güimil-Fariña, 2012. Extending an Abstract Reference Model for Transdisciplinary Work in Cultural Heritage. In 6th Metadata and Semantics Research Conference (MTSR 2012). Metadata and Semantics Research. Springer. Cádiz (Spain). 190-201.

[6] Gonzalez-Perez, C. and C. Parcero Oubiña, 2011. A Conceptual Model for Cultural Heritage Definition and Motivation, in Revive the Past: Proceeding of the 39th Conference on Computer Applications and Quantitative Methods in Archaeology, M. Zhou, et al., (eds.). Amsterdam University Press. Beijing (China). 234-244.

[7] Harris, E.C., M.R. Brown III, and G.J. Brown, 1993. Practices of Archaeological Stratigraphy. London. Academic Press.

[8] Incipit, 2013. ConML Technical Specification, version 1.4. ConML 1.4.

[9] Incipit, 2013. CHARM White Paper, version 1.0.3. Incipit, CSIC. http://www.charminfo.org/Resources/Technical.aspx

[10] OMG, 2006. Unified Modelling Language Specification: Infrastructure, version 2. formal/05-07-05.

[11] Parthenios, P., 2012. Using ConML to Visualize the Main Historical Monuments of Crete, in Computer Applications and Quantitative Methods in Archaeology (CAA) 2012. Southampton, UK.

[12] Pastor, O. and J.C. Molina, 2007. OO-Method: Conceptual Model-Based Automated Software Production. Springer-Verlag.