Pattern/GlobalLocal

GlobalLocal Pattern Family

A high level socio-technical pattern family addressing recurring issues in aligning technical and and distributed human systems

1. Problem statement

In many contexts, such as ontology development, system design, or standards development, one size does not fit all, and the assumptions shaping particular approaches may change significantly – even during the design phase itself. This is increasingly an issue in large scale, collaborative, heterogeneous and distributed systems such as extended enterprise systems, distributed supply chains.

This is a particular difficulty in the context of ontologies for Health Grids given the vision portrayed of seamless data sharing and the reality that groups within the same disease domain construct very different ontologies. https://wikis.nesc.ac.uk/mod/Main_Page.

FORCE 1 Benefits of Scale and Interoperability

Scaling technical systems requires use of interoperable standards and common infrastructure. Perceptions of cost benefits from this approach are drivers for use of common infrastructure, even where this is at expense of local requirements, usability or changing demands.

FORCE 2 Benefits of Localisation: Fitness for Purpose, Usability,Currency,Innovation

Scaling human systems is hard across distributed communities with different requirements, contexts, objective. Scaling can be counterproductive as (i) sensitivity to evolving local and emerging environmental factors drives uptake, local governance of quality currency and security, sustainability and performance [Refs] (ii)innovation is dependent in part on diversity 'at the edge' [Sawhney & Parikh;von Hippel]

2. Suggested Solutions

Family of emerging solutions to 'square the circle' are based on separation and/or alignment of these two competing forces in different ways. Metaphor here is Unity in Diversity, not One Size Fits All. Each solution offers a different balance of costs and benefits, depending on the purpose and priorities of the user. It suggests a spectrum of approahces from a technocentric 'one size fits all, through ABM to a user/context/purpose centric model such as as Motta's faceted ontologies

CORE_LOCAL Agree on a core requirements - reflecting areas of stable consensus and minimal variance Outsource (formally, transparently) responsiblity for local requirements (think MySpace, folksonomies) Identify a means of harvesting benefits from local niche developments

Known users:

Phil Murray Petrotechnics; Praxima; in standardising safety compliance software for oil and gas installations Pete Edwards, PolicyGrid, NCeSS www.ncess.ac.uk Shantenu Jha, Cactus Grid example, Programming Abstractions Theme. NeSC www.nesc.ac.uk

MEDIATION Technical, human or socio-technical middleware in different combinations

..still pasting..

ABM Agent based modelling

FACETED Motta

still pasting......

3. References and acknowledgements

EPSRC PACCIT project Lloyd A. D., Ure J., Pooley R. J., Dewar R. G. and Cranmore A., (2002) “Designing Enterprise Systems: leveraging knowledge in a distributed pattern-building community”. Presented to the International Society for Production Enhancement (ISPE) Collaborative Engineering Conference CE2002. Published in Advances in Concurrent Engineering, Jardim-Goncalves R, Roy R., and Steiger-Garcao A., (eds) Swets and Zeitlinger. ISBN 90 5809 502

---

-- JennyUre 2007-07-06 14:03:59