One of the sessions at June’s BDVA Summit (where Big Data Europe took part) was on standardisation. Organised by Felix Sasaki of W3C and Asunción Gómez Pérez of UPM, Standards – an essential foundation for the European Data Economy looked at the importance of standards and interoperability for the development of Europe’s data economy, and what needs to be done to encourage their use.
Standards have created a lot of value and innovation. We heard about:
- The Internationalization Tag Set (from Felix Sasaki) which improves the efficiency and profitability of localisation companies by providing a means to mark up text with words and phrases that should and should not be translated.
- DCAT (from Phil Archer) which enables interoperability of datasets between different data portals and is used by the European Commission for their open data, who devised a specific profile to adopt.
- The Linked Data standards BDE Partner Open Phacts uses to integrate data from multiple sources, making it available via simple APIs to reduce barriers to drug discovery in industry, academia and businesses.
- The work of the ISO/IEC JTC 1 Study Group on Big Data that has analysed existing standards relevant to big data and completed a gap analysis (Abdellatif Benjelloun Touimi of Huawei).
However, there is more that needs doing to capitalise on the benefits of standardisation. Key conclusions from the session were:
Standardisation needs funding
- Standards represent the consensus of a community. It takes time, effort and therefore money to work on standards that often only pay dividends once they’re done and proven to work.
Best practices & guidance are just as important as the standards themselves
- It’s not all about new standards. Guidance on using existing standards, training and exemplars are all equally important.
Don’t standardise processes
- Not everything should be standardised. In particular, the processes followed by a particular company should not be standardised as it’s the processes that capture innovation. That has to be down to individual entrepreneurship!
Encourage competition by citing open standards in procurement
- When procuring new equipment, cite the use of open standards as being a prerequistite. This:
- allows companies of all sizes to compete on a level playing field;
- prevents vendor lock-in.
Greater standardisation/predictability needed for APIs
- It’s not just data standards, it’s also APIs that can and should be standardised. For example, a standard API for machine learning and predictive analytics would improve the efficiency of businesses that extract data from natural language documents.
Other aspects discussed included the need for more help in converting data to easy but powerful formats such as JSON-LD, the desire to improve interoperability between XML and Linked Data systems and the absolute requirement that changes in technology are evolutionary, not revolutionary. No one is going to get rid of years of investment to replace it with a whole new tool chain in one go. Each step must be small and show a return on investment.
DBVA is new; it’s a collection of stakeholders both large and small that are all concerned in some way with driving Europe’s data economy, and is a key source of intelligence for the European Commission’s Data Value Chain Unit. Tool chains like Big Data Europe and the standards that underpin them have an important role to play. We plan to ensure that the need for any new standardisation work is recognised and acted upon.
It’s early days yet, but you have an interest in this space, please contact Irina Bolychevsky at W3C.