Modeling Execution Techniques of Inscriptions

The paper presents a small ontology that extends existing models for representing inscriptions with the notions of execution phases, techniques, tools and characteristics of the letters. The authors first discuss the two existing models which already cover this kind of information, namely the EAGLE and CRM-Tex models, and show with examples how one would model execution techniques in them. The result has its limitations, and the authors propose a small ontology that can be combined with the existing models to achieve a more accurate representation. They go into detail in the description of the ontology, giving plenty of examples of its different usages.

Content-wise, there is nothing to criticize. The ontology, being quite small, is extensively exemplified, as are the existing models that it aims to improve, so its relevance in the field of epigraphy seems clear. Also there seems to be nothing wrong with the quality of the ontology, especially considering that the authors extend well-known existing models, including the standard CIDOC-CRM, in a sensible way, using well defined classifications based on epigraphic research.

The negative aspect of the papers is thus not its contents, but its form. Personally, I found myself having to read certain sentences and paragraphs multiple times just to be able to understand the message the authors were trying to convey. Also, some of the examples provided in the paper can be difficult to follow at times, specially those in Section 2, given that they rely on properties and classes that are identified only by their code. The meaning of these codes is provided at some point in the textual body of the paper, but I found it quite cumbersome having to read through the descriptive paragraph over and over, trying to find the definition of each of the properties. Perhaps just listing them in a bullet point list would improve this, or even better, a graphical representation of the existing examples could be provided, as it is done with the proposed ontology.

In summary, the paper seems worth publishing, and all it needs is a bit of rewording to improve its readability.

It is an interesting paper addressing the modeling of execution techniques of inscriptions.
(1) Quality and relevance of the described ontology (convincing evidence must be provided).
The paper contributes with a small ontology that extends the CIDOC CRM standard (ISO 21127:2014) and its extension CRMtex. The scope of the paper although limited to the world of inscriptions is significant since not only improves modeling in this field but also indicates how a standard ontology can be specialized for the needs of a specific field.
In the paper the authors mix the notion of ontology with a hierarchically structured vocabulary and I think it would read better if they distinguish the two. They don't define clearly what are the classes extech:Technique, extech:Tool, extech:Color, extech:ResultType. In my view they are all subclasses of crm:E55_Type (skos:Concept) and if they are modeled this way they will achieve a well organized hierarchy of (eagle) terms. Following this view, ArsSubtractiva, ArsAddictiva, ArsPlastica and ArsMixta would be the toplevel terms in four hierarchies and not subclasses of Technique.
(2) Illustration, clarity and readability of the describing paper, which shall convey to the reader the key aspects of the described ontology.
Some minor suggestions to the authors in order to improve illustration, clarity and readability:
1. The proper way to reference the CRM extension for the study of ancient documents is CRMtex and not CRM-Tex as it is in the paper.

2. Following the CRM way of presenting properties, when a link used is the inverse, it should be indicated by using the inverse symbol (i) in order to read well.
For instance in the first example:
:INS1 a crm:E22;
crm:P108i :Activity1;
crm:P31i :Activity2.
:Activity1 a crm:E12;
crm:P32 eagle:3.

Both P108 and P31 are the inverse properties (P108i was produced by and P31i was modified by)
This should be reflected and changed in all the examples.

3. I don't understand why in the third modeling approach, when the ontology is introduced, the Activities become three instead of two that were in approaches one and two. I assume that the example is always the same (an inscription INS1 which was engraved and then painted). Could the authors elaborate on this?

4. It would help a lot the reading of the paper if the authors add labels in their modeling approaches. The reader doesn't know what is eagle:3 or eagle:10. But if it is presented as eagle:3[a-readable-label-for-3] and eagle:10[a-readable-label-for-10] then it would be better understood.

5. Be careful with the namespaces. Sometimes extech is written ExTech (fourth and fifth modeling approaches).

6. In the paragraph that starts after the fourth example domain is used instead of
range:
… domain of extech:uses_technique …
extech:usesTool will have as domain extech:Tool
extech:hasColor will have as domain a extech:Color.
I believe in all three cases the authors mean range.

7. Figure 1 is almost impossible to read. There is enough space to use a bigger font and also it would be better understood if the subclassing to the CRM is also indicated.

8. Footnote 4 in page 1 should be updated since it refers to a planned meeting on February 2019. Did the meeting happen?

The paper proposes a small scale ontology aimed at modelling the semantics of execution techniques of inscriptions, particularly focused on techniques, tools and letter characteristics. The main motivation of the work is driven by the need to provide specialisations for the types of techniques and tools that relate to the domain of study. The proposed work attempts to address a particular problem in typological specialisation which can be addressed in various ways. For example CIDOC-CRM provides the E55 Type class for implementing this kind of specialisations as this is already acknowledged in the paper when discussing the EAGLE model in section 2. In addition, several other European projects, such as CrossCult ( https://www.crosscult.eu/), Ariadne Infrastructure (https://ariadne-infrastructure.eu/), Iperion CH (http://www.iperionch.eu) and the British Museum project ResearchSpace (https://www.researchspace.org/) have used SKOSified resources to complement the semantics of CIDOC-CRM for addressing the issue of in typological specialisation. Therefore, the paper aims to address an issue which has already been addressed but by looking into the particularities of epigraphy. In this respect, there is room for some original contributions but the paper should be significantly strengthen before it is considered for publication.

In details, there is extensive use of colloquial language and arguments are delivered in convoluted statements which in many cases are hard to follow. The language of the paper must be significantly improved, as at the current state the paper is hard to comprehend. The contributions of the paper are summarised under void statements which claim that the proposed method models the semantics of the domain “better” but the claim is neither justified or is thoroughly evaluated. The paper will be benefited by providing a clear context of the research objectives making a clear case to the need of the proposed work and acknowledging other similar works that addressed the problem of typological specialisation. In addition, the paper should consider using figures and graph diagrams for communicating the proposed semantics as the current examples use code snippets which are not always easy to follow. The paper should be also made self-contained, meaning that key arguments, motivation and background of the work should be included in the paper. Currently, the paper simply points to reference [2] for all the important pieces of information and arguments that support the proposed work. Though, the referenced paper is written in Italian making it impossible for an international audience to follow the proposed work. Section 4 should be expanded to include the necessary context to support the discussion around the given examples. For example why the particular examples are included, how they compare between them and what is the main benefit that we receive when using the proposed semantics. Finally the paper should be expanded to include some form of evaluation or discussion section that highlights main contributions, limitations and future improvements before coming to the conclusion section.