||"Our focus is to consider the construction industry as essentially an information processing system. In its ideal form, practitioners (each with an individual internal representation of design intent) interact with other practitioners by first interacting with an information processing system that manages various shared external representation of design intent. The underlying assumption (from an information technologist's perspective) is that design data is held in a sufficiently complete representation, and that changes to this representation are transactions that move the representation from one consistent state to another. We might call this 'enterprise computing' for construction.This ideal of 'enterprise computing' for construction can be compared to the realities of current practice.- Due to its fragmentation, the construction industry generally perceives its use of information technology in terms of multiple discrete 'individual' systems (with the resulting proliferation of discrete documents) rather than as an enterprise systems.- The drawing tradition, which represents building in 2D, with different representations of the same design split across multiple independently editable documents inhibits consistent management of design and the use of analytical tools. While these may be familiar arguments, there are new object oriented and data management tools emerging from key software developer, such as Bentley Systems, that are designed to address the specific needs of a 'construction enterprise', namely geometric generality, multiple application semantics, multi-user access, and transaction management. These systems also address the scalability and reliability issues required for deployment in practice. Again, arguments for (and advantages of) systems of this type have been discussed in the research literature for more than two decades. The difference is that these systems are ready for deployment. But with this prospect for a broader application of 'Enterprise Computing' for Construction, there are associated other significant issues which may concern both the 'strategic' and the 'creative' practitioners, namely:- Semantic completeness: building a sufficiently complete multi-disciplinary representation of design intent- Data integrity: where any intelligent components are used, these should not become 'orphaned', for example, by object ""instance"" data being detached from the definitions of the corresponding class- Data longevity: the integrity of design and other data should be maintained for the life-time of the building, across new hardware platforms and operating systems. Upgrades to the application and any intelligent components should not disrupt or invalidate existing data- Parallelisation of design: individual designers or engineers should be able to work in parallel, and then be able to synchronize their changes to design data with co-workers- Expressibility: architectural design and construction engineering are open-ended domains. Additional intelligent components should be capable of being added on a ""per project"" basis.Within this context, this paper will explore the essential 'tension' that exists within the Architecture and Construction sectors. On the one hand, there is a perceived need by construction managers for computing tools based on clearly defined and agreed schema to control the construction process (thereby giving economic advantage, comparability, etc.). On the other hand, creative designers who are under other competitive pressures, are expecting a different set of computing tools to allow the exploration of new building configurations and construction geometry. While in the former case a standardisation of schema (as the foundation of a traditional ""Enterprise Computing"" system) would appear to be in order, in the later case the essential 'open-ended-ness' of the creative process demands ""extensibility"" as a pre-requisite of any computing system. These differing requirements (and indeed, attitudes) within the user community, presents software developers with interesting challenges. What technologies (for example, object and/or relational) and what 'domain abstractions' are appropriate foundations for solutions for these differing requirements. Or indeed, what technologies and 'domain abstractions' can be used as the basis for broader set of applications whose design is intended to unify across this apparent ""management-creative"" divide…hence the theme of this paper: ""'Extensible Enterprise Computing' for Construction"".Fundamentally, this is not exclusively an issue of technology. We need to address both the technical and cultural issues if we are to realise our collective ambition of providing effective tools with which to support collaboration between the diverse range of interests that occur within the Architecture and Construction sectors."