An interview on structured metadata with Vassilios Peristeras, SEMIC.EU's Project Officer
Vassilios Peristeras has recently taken over the responsibility for the semantic methodologies and Linked Open Data actions in the ISA Unit. A seasoned expert in these fields, he has joined as a seconded national expert from Greece the Commission after several years of research and practical experience in organizations like the United Nations, CERTH/ITI, the Greek National Centre for Public Administration and Local Government, and Digital Enterprise Research Institute in Galway, Ireland (DERI).
In this interview, Vassilios describes the usefulness of SEMIC.EU?s Core Person, comments on the more general Core Concept idea in the greater context of eGovernment interoperability - and makes the case for expanding SEMIC.EU in two directions: a) creating, hosting, and maintaining a selective library of harmonized, generic metadata schemas and b) creating the infrastructure for a federated portal for eGovernment metadata schemas.
SEMIC.EU's Core Person was developed in a "shared development process" by some 100 experts. It is now a mature interoperability asset. Where do you see its greatest potential as a concept to enable seamless data exchange?
Vassilios Peristeras
It is really hard to foresee and define the scope and applicability of this concept. The Core Person specification is designed to fit into any usage scenario where information about persons is documented in information systems (ISs). You can easily see that this covers a huge number of existing ISs. Core Person is for us a first nice example of a Core Concept. It is interesting to see that there are already mappings of 14 different national "person" concepts to our Core Person specification. There are two main ideas behind the core notion here: a) the core is highly reusable: the specification is simple and captures basic and generic characteristics of a person, regardless of the context this entity is used, b) the core is extensible: domain specific specializations can be drafted on top of the core representation (e.g. voter, passenger, employee). This approach guarantees a minimum of cross-domain interoperability while it provides domain-specific communities with a common starting point for drafting their own specializations by adding metadata to the core.
This is a simple solution which could lead us forward in promoting eGovernment interoperability. My first and foremost concern is that interoperability assets remain simple, in order to become practice-ready, i.e. (re-)usable, and at the same time powerful in making data and information exchange possible in different contexts. Further to that, I think that this approach increases the possibility for adoption, as complexity and overspecification usually result in diverged views and Babel towers where consensus is hard to reach, especially amongst 27 countries with different languages, cultures, etc.
For those specifications that make it to the "conformant" state like foreseen for the Core Person: What should be the next step to make it mandatory?
Conformance means that we are convinced that an asset is not only ready for use or reuse but that it is the single most recommendable piece of interoperability enhancement for a given purpose or context. In order to claim this, we should involve important stakeholders in the process or at least keep them and their communities in the loop during the design exercise. This includes not only the 27 EU Member States as our direct audience, but also experts from the academia as well as standardization communities and bodies. The idea is to reuse whatever is available (and there is a lot) and define new things only when this is absolutely necessary.
Now for the adoption ? The ISA programme cannot make anything mandatory. This contradicts to fundamental EU principles. But also taking into account the experiences from national initiatives, I personally think that this wouldn't even be effective. At the ISA programme we have a clear target: we want to recommend the usage of common data models and if possible to promote these specifications in the national interoperability strategies and/or metadata framework and repositories. We try to increase our possibilities for success by a) involving our audience in the design process and b) keeping the specification simple and thus relevant and applicable in all different countries.
We intend to seek for broad consensus on the Core Person and on other Core Concepts. We will utilize for this purpose amongst others our ISA Working Groups and Committee where we have representatives of all MSs. The goal is to populate a small library of simple Core Concepts for basic entities (e.g. Person, Vehicle, Company, Building, Project, Software etc) and promote their reuse to our MSs. There, our business ends. It is up to the standardization bodies, if they find some of the work we are doing relevant to their domain and interest, to further promote some of these specs for endorsement in their communities.
You are also an expert in the field of Linked Open Data (LOD). How do you think the concept is applicable in the case of the Core Person?
Core Concepts could become available as Linked Open Metadata. Our intention is to make available our Core Concept library as an eGovernment Linked Metadata Cloud. You can find more details on this here.
Data on persons is surely not the only field that has such an impact across all domains, countries and levels of government? Are there other assets you think should soon undergo a shared development process? Suggestions have already been submitted by the community.
SEMIC.EU currently stores over 500 semantic assets. Is it the only repository to be consulted when someone wants to find reusable assets? The answer is no. SEMIC.EU is not one-of-a-kind and this is rather good news. There are other international and national initiatives that index and make available semantic assets. It is also common for national administrations to document their own eGovernment metadata schemas. Currently, it is not possible for someone to have an overview on all possible assets that s/he might have an interest. SEMIC.EU has a clear goal on its roadmap - to provide an infrastructure to allow this kind of repositories federation. One way to do this, while preserving the autonomy of each repository, is to draft and agree on a metadata schema that generically describes a "semantic asset". This is what we call ADMS, the Asset Description Metadata Schema. The work has already started. We try to bring together representatives from our MSs, semantic assets repository owners, metadata and vocabulary experts and standardization bodies to draft a simple but highly reusable schema, by building on top and reusing existing work in the area (e.g. DCAT, DCMI). Once it becomes available, ADMS could play the role of a common language for metadata repositories, enabling federation, cross-querying and improving the discoverability of the assets regardless of where and how they are stored. By providing this advanced querying mechanisms, SEMIC.EU will extend its scope to become not only an asset repository and clearinghouse but also a semantic asset portal and registry. This is an ambitious however feasible and of great value target to be achieved.
Related information: eGovernment Linked Metadata Cloud