The Knowledge Reengineering Bottleneck

Submitted by Pascal Hitzler on 05/06/2010 - 08:35
Paper Title: 
The Knowledge Reengineering Bottleneck
Authors: 
Rinke Hoekstra
Abstract: 
Knowledge engineering upholds a longstanding tradition that emphasises methodological issues associated with the acquisition and representation of knowledge in some (formal) language. This focus on methodology implies an ex ante approach: “think before you act”. The rapid increase of linked data poses new challenges for knowledge engineering, and the Semantic Web project as a whole. Although the dream of unhindered “knowledge reuse” is a technical reality, it has come at the cost of control. Semantic web content can no longer be assumed to have been produced in a controlled task-independent environment. When reused, Semantic Web content needs to be remoulded, refiltered and recurated for a new task. Traditional ex ante methodologies do not provide any guidelines for this ex post knowledge reengineering; forcing developers to resort to ad hoc measures and manual labour: the knowledge reengineering bottleneck.
Full PDF Version: 
PDF icon swj32.pdf
Submission type: 
Other
Responsible editor: 
Krzysztof Janowicz
Decision/Status: 
Accept
Reviews: 

Review 1 by Bernardo Cuenca Grau:

Minor comments for improvement:

- Page 1 second column: "… that culminated in the now commonplace use of the term to refer to a syntactic construct" : It is unclear to me what do you mean by a "syntactic construct" in this context.

- Page 2 first column: " ontologies are never used as a component of an expressive knowledge based system, but rather as facilitator for knowledge management" . This sentence sounds rather obscure to me. Could you clarify?

- "Themethodologies" ---> "The methodologies"

- In the quote on Page 2, column 2: Which particular tedious, time-consuming and expensive procedure is the quote referring to?

- On page 2 column 2: "These systems are evaluated against a reference alignment, or checked for coherence, but not against a set of instance data". Could you briefly describe how you think such evaluation with instance data should be performed? Where would the data come from? Should it be realistic?

- Page 4, column 1: What do you mean by data reuse patterns?

- The title of the paper explicitly mentions the "Neats" and the "Scruffies", but the main body of the text only mentions them very briefly. What do you think should be the main role of the scruffies in the age of linked data?

Review 2 by Sören Auer:
The position paper describes the knowledge engineering bottleneck, which emerges on the web of data due to the prevalent data-driven knowledge elicitation and representation approaches. The author identifies four particular challenges: data dependency, limited control as well as increasing complexity and importance. The paper is well structured and interesting to read. It is in particular valuable, since it regards the web of data in the light of the traditional knowledge engineering and AI debate. However, from a slightly different perspective the challenges can be also seen as strengths: data dependency is an indicator for ground truth, limited control for organic evolution and community consensus etc. Adding such thoughts to the final version of the paper could increase its value as a description of the current state-of-play of the data web.

Tags: 
Reviewed