LOTED2: an Ontology of European Public Procurement Notices

Paper Title: 
LOTED2: an Ontology of European Public Procurement Notices
Isabella Distinto, Mathieu d'Aquin, Enrico Motta
This paper describes the construction of the LOTED2 ontology for the representation of European public procurement notices. LOTED2 follows initiatives around the creation of linked data-compliant representations of information regarding tender notices in Europe, but focusing on placing such representations within their legal context. It is therefore considered a legal ontology, as it supports the identification of legal concepts and more generally, legal reasoning. Unlike many other legal ontologies however, LOTED2 is designed to support the creation of Semantic Web applications. The methodology applied for building LOTED2 therefore seeks to find a compromise between the accurate representation of legal concepts and the usability of the ontology as a knowledge model for Semantic Web applications, while creating connections to other relevant ontologies in the domain.
Full PDF Version: 
Submission type: 
Full Paper
Responsible editor: 
Minor Revision

Submission in response to http://www.semantic-web-journal.net/blog/semantic-web-journal-special-is...

Solicited review by Rinke Hoekstra:

General remarks
* The paper is wel structured, and well written. Alhough the conclusion is clearly an afterthought, listing all the things the authors wanted to remark, but could not fit in the rest of the paper.
* The LOTED2 ontology is clearly the result of careful considerations and study of existing work.
* To what extent do the authors feel that they have succeeded in finding the sweet spot between lightweight vocabularies and axiomatic ontologies?
* Does LOTED2 meet its purpose? In other words, does it really allow better access to public procurement notices? What is the sketch of an application that would use LOTED2, and how would you go about evalutating it? It is clear from the paper that building LOTED2 was very instrumental in a better understanding of the domain, an evaluation of the reusability of other ontologies (in particular LKIF-Core and GoodRelations), but the authors do not really evaluate LOTED2 itself. This is a real pity.

section 1/p1,2
* What evidence is there that computational legal ontologies are playing an increasingly prominent role in AI&Law research? My personal observation is to the contrary, with a shift towards (the always prominent) argumentation.
* what references does [9-4] refer to?
* I wouldn't say that [9] introduces a new trend, but rather that the focus on lightweight ontologies/vocabularies that was always there (cf. the DALOS and LOIS work) is now more prominent than it used to be, compared to axiomatic ontologies. This is in line with the developments in the SW world at large.
* "related to public procurement domain" -> "related to the public..."

* "related works" -> "related work"

section 3/p4.5
* Pellet and Hermit deserve a citation/reference
* Something funny is going on with footnotes 18 and 19
* There's a Word reviewing/track changes artifact on p5 (just below figure 2) and p7
* If you mention AkomaNtoso, it makes sense to also mention CEN MetaLex (or plain MetaLex) which was first in including semantic markup in legal/legislative XML. cf. also the Metalex Document Server.
* Describing the modules of LOTED2 in an indented list does not contribute to readbility
* Consider using more paragraph breaks in the section describing the modules.

section 4
* How useful is an integration with Good Relations, if you had to decide to modify GR for your purposes? Isn't just the reuse of widely used GR concepts what makes the integration worthwile? Modifying GR just introduces another ontology/vocabulary that is not used anywhere.

Solicited review by anonymous reviewer:

The paper presents an interesting discussion of the balance to strike between expressivity and computability of legal ontologies in the framework of linked data and SW applications.
The authors provide also an interesting discussion of the issue of 'locality' vs. 'globality' of legal ontologies and its consequences for linked data. The ontology modules are described in detail.
However, several aspects could be further developed:
- validation: how has the ontology been validated? Did domain experts take part in this process?
- methodology for concept identification: an extensive literature discusses the different methodologies for legal ontology engineering (i.e. bottom-up, enhanced by terminology extractors, manual analysis of legal sources, ...). The paper mentions only briefly the methodology followed for the construction of the ontology. A more detailed description would be desirable; for instance, were any terminology extractors used? How were concepts and term-variants identified and distinguished from each other?
The use of English is correct, Fig.3 should be bigger for publication, the bibliography entries are complete.

Solicited review by anonymous reviewer:

This paper presents the LOTED2 (Linked Open Tenders Electronic Daily) ontology. It is built upon the existing LOTED ontology, providing a more detailed modeling of the domain and a convergency with the Good Relation ontology. The design of the ontology is driven by applications in the Linked Open Data Web.

I am not an expert in modeling legal ontologies so my reviews merely focus on judging the merit of the paper as a research paper.

There are three drawbacks with this paper:
- lack of clear description about the problem domain
- lack of clarity in the methodology
- lack of evaluation

1) The biggest issue with this work is the lack of a systematic description about the legal domain and the scope of issues that are addressed by the proposal. This makes extremely hard to understand why Semantic Web technologies are needed for this domain and why a new version of LOTED is needed. Without concrete examples and a clear presentation of the domain dealt by this work, the motivations remain a bit superficial and the solutions are hard to justify or assess. Most of the motivations can almost equally be applied to any other application domains to which semantic Web technologies have been applied.

2) The methodology description suffers from organisation. In my opinion both sections 3.1 and 3.3 are related to the design decisions. But section 3.1 failed to articulate the decision made for the ontology. It is not clear to the readers what the authors mean by "a compromise between a suitable and rigorous representation of legal domain and a practical way to represent data". The same applies to section 3.3. How did the bottom-up and top-down approach work together in the end? Was there an iterative design, testing process? If so, how was the ontology tested and evaluated throughout the design process?

3) There was no presentation of any Semantic Web applications built using this ontology, nor was this mentioned in the future work. Combined with the aforementioned lack of concrete examples and motivations, it is hard to see how this ontology can be useful. It is strongly recommended that the authors add one section to either walk through an example of how this ontology can be useful or provide a discussion about ongoing or planned implementation of this ontology in a concrete application, or something similar.

A minor note on section 3.4: This section is a bit lengthy and dry. It is really hard to follow all the concepts described in the section, without a clear motivation example or understanding about the methodology applied to design the ontology. Again I would strongly recommend the authors give at least one real example, preferably with a diagram, to demonstrate how the concepts could support a real case.