Agile data modelling for parliament(s?)- using Domain Driven Design and RDF
Author(s): Anya Somerville, Michael Smethurst, Silver Oliver
Full text: submitted version
Abstract: The offices that make up the UK Parliament have largely evolved independently, and over many years. Organisations with co-evolving parts are complex, making it difficult to design common data representations. At the UK Parliament, this has resulted in disjointedness in the data, internal tools, and the public website. This case study outlines our use of tools to address this problem. The approach acknowledges and addresses areas of organisational complexity and follows an agile methodology. We have had success pairing the flexibility and expressiveness of the RDF data model with domain driven design. The outcome mitigates many of the previous modelling design issues associated with parliamentary projects.
Keywords: linked data; ontology design; domain driven design; agile data modeling; parliament
Review 1 (by anonymous reviewer)
The paper does not conform to the LNCS style. Also, the paper does not go into technical detail, but rather expands on a number of considerations for the authors to choose RDF as data model for capturing UK parliament data. There is no evaluation of this choice, nor an implementation report.
Review 2 (by anonymous reviewer)
The paper's Introduction opens with "We think there are weaknesses with the standard modelling approaches being used by the UK Parliament..." Unfortunately, the paper does not deliver examples of either the poor modelling approaches or the exemplar modelling approaches they desire (i.e. by lacking pointing to the precedents outlined in previous work or literature.) Section 2 claims that they data models the have created have resulted in an "improved open data platform" and "more effective business facing tools". Both claims should have evidence that indicate how these improvements were measured... or cite separate works or surveys that does. (I realize the paper is focusing on the Agile modelling process.) The section continues by describing the current complex landscape but uses colloquial language that would benefit from diagrams or some form of illustrations. For example: - "Bits of parliamentary business... Sticking out at strange angles" The strongest sections of the paper describe Agile and domain driven design. But even in this section, the authors make statements about the creator of DDD and what he recognizes and encourages, but the authors fail to cite the work. The sections on subject matter expertise are written in a narrative manner. In my opinion, the section is overshadowed by the phrase "answering our dumb questions". I appreciate the honesty and insight in describing the scenario, but it raises questions about how much preparation was given to the subjects prior to the meetings (which is probably the point). Likewise, the statement about avoiding technical language so as to not alienate public servants and librarians is problematic. The section on documenting the results seems to gloss over seemingly important details -- especially regarding an RDF data platform. Overall, the paper is successful is explaining many of the complexities of dealing with UK Parliament and the approach using domain driven design to evaluate the space. However, it is difficult to ascertain what was accomplished through this process. In its current state, this paper is simply too casually written and researched for ESWC standards. The references and footnotes show no citations pointing to published literature about RDF or data modeling. Because of this, it is difficult to assess how much experience and background information is being brought into approaching this use case. This is compounded by the lack of examples in the paper. The prose is probably a personal preference for me, but there are simply too many "we" and "you" throughout the paper.
Review 3 (by Monika Solanki)
In this paper, the authors present the approach they have taken towards the design of an ontology for parliaments using the the paradigm of domain driven design. The paper reads more like a report than a technical contribution to a conference. The authors talk about an Agile approach to ontology development, however fail to reference some of the past work in this area. The literature review in general is weak. The paper is more about how DDD, RDF and building small, modular ontologies are good tools for designing and serialising domain models, rather than how these models have been actually used, what was the impact of these models on the information sharing infrastructure within the parliament offices and more importantly what lessons were learnt in the process of implementing a solution based on Semantic technologies. The merits of modular ontologies are well known and well documented elsewhere in the literature, however these too have not been well referenced. There is no detail provided on the domain models themselves or the process of refining them into concrete data models. This I believe is quite important in deciding on what should be the appropriate level of abstraction in moving from a domain model to its concrete serialisation. I also missed competency questions, which are an integral part of any ontology design. In general the paper just scraps the surface of what could have been a good contribution. There is not enough depth in the content presented which could be used for future learnings by a community looking towards reusing or even extending the proposed parliament ontologies. BTW, the paper does not follow standard llncs format  SAMOD: an agile methodology for the development of ontologies  Modular Ontologies - Concepts, Theories and Techniques for Knowledge Modularization
Review 4 (by Anna Tordai)
This is a metareview for the paper that summarizes the opinions of the individual reviewers. The reviewers agree that this paper does not meet the criteria to be included in the conference. They mention that it lacks detail, depth, and coverage of literature. The paper is casually written and does not conform to the style guide of the conference. Laura Hollink & Anna Tordai