Populating the Data Space for Cultural Heritage with Heritage Digital Twins

Round 1

Reviewer 1 Report

SUMMARY

The stated aim of the article is to introduce a novel semantic ontology for the design of a digital space for cultural heritage, based on the Digital Twin as a pivot concept. It outlines the main features of the proposed Heritage Digital Twin ontology and provides some good examples of application. The article advances current knowledge in the field and contributes to the objectives of the European Strategy for Data, at a time when there is significant interest in the digital transformation of cultural heritage, as well as in the creation of a common European data space in this area. However, some of the terminological indications and acronyms deployed in the article may prove confusing to readers, for example, there are varying references to both “data space” and “digital space”, with no distinction made between these concepts.

CONCEPT COMMENTS

The proposed Heritage Digital Twin ontology is described reasonably clearly and seems suitably tailored to the methods and theory of this sector. Aimed at organising and managing digital information about heritage, it is presented as a starting point for building a cultural heritage data space and would appear to have the requisite characteristics, ensuring compliance with existing standards, and enabling reuse and interoperability with previous platforms. As the authors acknowledge in the Conclusion, the ontology will require further refinement and testing in practice.

REVIEW

While this article should be of primary interest to those working in the cultural heritage sector and interested in its digital transformation, it also provides a useful introduction to the concept of cultural heritage for those less familiar with this field, including appropriate references. The authors draw interesting comparisons with the use of Digital Twins in other sectors, building on this concept to extend it to the domain of cultural heritage. Nevertheless, there is a lot of detail, which may distract the reader from the principal focus of the article, i.e., a cultural heritage digital/data space.

In terms of novelty, the article takes the discussion beyond current proposals to use a Heritage Building Information Modelling (HBIM) approach, based on ‘enriched’ 3D, as an entry point to access heritage documentation and as the main constituent of the digital twin. It aims to address the limitations of such an approach, which would be restricted to immovable physical assets and which fails to consider other relevant features of many cultural heritage assets, including intangible value or a ‘born-digital’ nature. The authors develop a semantic model of the digital twin, leading them to arrive at a new definition of the Heritage Digital Twin (HDT) as the digital representation of the complex of knowledge about a heritage asset, organised according to a specific ontology (the Heritage Digital Twin Ontology). It thereby extends the HBIM model and marks an important contribution to the digital evolution of cultural heritage and to meeting the objectives of the Digital Europe programme. Furthermore, the article points to the possibility of exploring the automatic production of parts of the Heritage Digital Twin by means of artificial intelligence.

The introduction to the Heritage Digital Twin ontology and semantics in section 2.2 is aimed at readability also by non-technical but digitally-aware heritage professionals and researchers. However, while it seems scientifically sound, the long narrative description of classes and properties, may assume a certain degree of technical knowledge and could risk losing some of these readers. The system of labels and numbers, and use of many unfamiliar acronyms may also prove confusing.  While the inclusion of real examples to illustrate the classes is helpful, at times these are overly detailed.  The inclusion of examples of services operating on the ontology in section 3.2 is also helpful, but perhaps they could be reduced in number. The semantic graph of the HDT ontology illustrated in Figure 1 could also possibly be enlarged. As an additional suggestion, it could be an idea to move the tables of Class and Property definitions (2.3.1 and 2.3.2) to an Annex, as they seem to break up the flow of the narrative that leads to a discussion on suitability of the ontology.

Considering the intended focus of the article is on using HDT ontology as a foundation for a knowledge base of cultural heritage and as the main constituent of the heritage data space, there is perhaps too great a focus on describing this ontology and not enough attention on how to actually build a data space on this ontology (a discussion of this aspect starts on page 15 and the vision for the data space is only outlined on page 16 – the cultural heritage data space is mentioned for the first time on pp. 8-9).

A final minor comment concerning language: there are a number of grammatical errors (for example an incorrect use of the word “such” on repeated occasions) - it is suggested that the authors review and revise the language as needed.

Author Response

Explanations to Review 1 (reviewer’s comments in italic)

While this article should be of primary interest to those working in the cultural heritage sector and interested in its digital transformation, it also provides a useful introduction to the concept of cultural heritage for those less familiar with this field, including appropriate references. The authors draw interesting comparisons with the use of Digital Twins in other sectors, building on this concept to extend it to the domain of cultural heritage. Nevertheless, there is a lot of detail, which may distract the reader from the principal focus of the article, i.e., a cultural heritage digital/data space.

Actually, we deliberately included details as it is well known that “the devil is in the details”. We wanted to avoid a theoretical discussion that might require further development and sound as an academic exercise, by showing “how it works” in some examples taken from real documentation. In the revised version we explain why some details are included and suggest which parts may be skipped at first reading but are available when further insight is desired.

In terms of novelty, the article takes the discussion beyond current proposals to use a Heritage Building Information Modelling (HBIM) approach, based on ‘enriched’ 3D, as an entry point to access heritage documentation and as the main constituent of the digital twin. It aims to address the limitations of such an approach, which would be restricted to immovable physical assets and which fails to consider other relevant features of many cultural heritage assets, including intangible value or a ‘born-digital’ nature. The authors develop a semantic model of the digital twin, leading them to arrive at a new definition of the Heritage Digital Twin (HDT) as the digital representation of the complex of knowledge about a heritage asset, organised according to a specific ontology (the Heritage Digital Twin Ontology). It thereby extends the HBIM model and marks an important contribution to the digital evolution of cultural heritage and to meeting the objectives of the Digital Europe programme. Furthermore, the article points to the possibility of exploring the automatic production of parts of the Heritage Digital Twin by means of artificial intelligence.

The introduction to the Heritage Digital Twin ontology and semantics in section 2.2 is aimed at readability also by non-technical but digitally-aware heritage professionals and researchers. However, while it seems scientifically sound, the long narrative description of classes and properties, may assume a certain degree of technical knowledge and could risk losing some of these readers. The system of labels and numbers, and use of many unfamiliar acronyms may also prove confusing. 

There are some technicalities that unfortunately cannot be avoided, for example the use of labels and numbers, which are however accompanied by the full self-explanatory names to facilitate reader’s understanding. The full description of classes and properties has been moved to an Annex that can be used for reference or ignored if the reader is not interested in it.

While the inclusion of real examples to illustrate the classes is helpful, at times these are overly detailed.  The inclusion of examples of services operating on the ontology in section 3.2 is also helpful, but perhaps they could be reduced in number.

The details of all the examples are intended to address the diverse interests of heritage professionals who may ask “does this include all what I wish to be included?” All the information is stored in the data space, and users may consider only the parts that are relevant to their needs. We wanted to give the impression that all communities may benefit from the proposed organisation of the data space as they find what they are looking for. This obviously totals to a large amount of data, part of which may be superfluous for a community but very important for another one.

The semantic graph of the HDT ontology illustrated in Figure 1 could also possibly be enlarged.

Unfortunately, the page design of the Journal reduces the width of figures to predetermined narrow limits. We however redesigned the figure to try to improve readability.

As an additional suggestion, it could be an idea to move the tables of Class and Property definitions (2.3.1 and 2.3.2) to an Annex, as they seem to break up the flow of the narrative that leads to a discussion on suitability of the ontology.

As already mentioned, we did so.

Considering the intended focus of the article is on using HDT ontology as a foundation for a knowledge base of cultural heritage and as the main constituent of the heritage data space, there is perhaps too great a focus on describing this ontology and not enough attention on how to actually build a data space on this ontology (a discussion of this aspect starts on page 15 and the vision for the data space is only outlined on page 16 – the cultural heritage data space is mentioned for the first time on pp. 8-9).

We think that describing how to build a data space on the ontology could be the subject of another paper (or perhaps several papers) as it requires further specifications, for example: should it be server-based (as the present Europeana system), or cloud-based, as the present ARIADNE system for archaeology and the envisaged collaborative cloud for cultural heritage? Which community should it serve? Who should manage it and on behalf of whom?

Therefore, we only suggested some possible technical solutions to show that such a system is feasible, reserving a full-fledged analysis to future work.

A final minor comment concerning language: there are a number of grammatical errors (for example an incorrect use of the word “such” on repeated occasions) - it is suggested that the authors review and revise the language as needed.

Absolutely true! With a scrupulous language review, we removed 90% of the occurrences of “such”.

Reviewer 2 Report

The article presents the definition of a data model for cultural heritage (in a broad view that includes them all) and their digital twins. It also presents the definition of a cultural heritage data space.
The paper is interesting and well written, but the structure needs some adjustments:
- paragraph 3.2 presents content that seems marginal: it would be helpful to explain why the paragraph is included (or shorten/remove it)
- paragraph 3.4 describes in detail the Pafos Gate: explain why it is useful in the article or shorten it.
The paper does not include the state of the art and related works.

Some minor errors:
- Figure 4 is not referenced in the text. Apparently not useful figure in the text (as well as figure 5): explain the meaning in the paper;
Figure 9 by mistake is referred to as figure 5.
lines 854-925 are a bit misleading as a graphic presentation: it should be improved.

I suggest accepting it with minor revisions.

Author Response

The article presents the definition of a data model for cultural heritage (in a broad view that includes them all) and their digital twins. It also presents the definition of a cultural heritage data space.

The paper is interesting and well written, but the structure needs some adjustments:

- paragraph 3.2 presents content that seems marginal: it would be helpful to explain why the paragraph is included (or shorten/remove it)

We shortened paragraph 3.2. The value of this part consists in showing that the Heritage Digital Twin proposed here goes beyond the mere organization of documentation in a Data Space, since it may also enable automatic actions as do Digital Twins in industry (examples are proposed). This approach anticipates requirements by the community, as for example envisaged by a forthcoming European call on the use of sensors and digital services relying on the characteristics of heritage data as documented in the Data Space. In this way the Data Space will not only store documentation but also behave as a dynamic resource.

- paragraph 3.4 describes in detail the Pafos Gate: explain why it is useful in the article or shorten it.

The example shows how different information may be organized in a Heritage Digital Twin: art and architectural history, documents from the past, structural information, and information on intangible values may be all hosted in the Heritage Digital Twin, from where users may extract what they need. This is now briefly explained in the example description.

The paper does not include the state of the art and related works.

We summarized related work in the introduction. As far as we know, this is the first ever investigation on using digital twins in the domain of cultural heritage documentation, so the state of the art is nil.

Some minor errors:

- Figure 4 is not referenced in the text. Apparently not useful figure in the text (as well as figure 5): explain the meaning in the paper;

Figure 9 by mistake is referred to as figure 5.

lines 854-925 are a bit misleading as a graphic presentation: it should be improved.

They have been corrected