For many years ICT-related research for BC process innovation has been researching communication technologies that would help the BC industry with the application of computers and computer networks. After 20 years, progress has been made, but real improvements in speed and consistency were lacking. The main reason is that our instruments are not able to understand our technical communications. Humans have to read drawings, extract input for their instruments, let those instruments do their jobs, translate computer output back into humanly understandable information and pass that on. All these transformations do not add value. On the contrary, they are error prone and add to the confusion.

With the arrival of XML/RDF and the Semantic Web a new opportunity arose to really improve human-computer and computer-computer communication. This might potentially be the missing technology the BC industry needed for a long time. What if we could communicate with our instruments in the same technical terms as we use between ourselves? Could ICT communication be elevated to that level? Would that mean that unneeded human activities can be cut out of the information chain and that humans no longer would have to function as translators between the project communications and the computer applications? Would that mean a reduction in communication errors and in cost of failure? Would that speed up the process? What would the consequences be and how should this new technology be implemented?

Could it be that the missing technology is the ability for instruments to participate in human technical communication?

The research was performed at Delft University of Technology and at STABU, the Dutch building Specification institute, so BC ICT in theory and in practice. The first two years took place within the EU research project `eConstruct' which ensured a good embedding in international BC ICT research.

Current first generation ICT support mainly focuses on individual tasks and not on project-related tasks requiring information and knowledge sharing. (Computerised) Classification and Specification systems are widely accepted in the market, but do not offer enough information richness to adequately support information sharing in BC. PDT solutions are either too centralised or effectively cater for the top 5% of the companies only, not for BC as a whole.

The research reported here was mainly aimed at the use of innovative Internet technologies on the intersection of two fields: (a) existing BC ICT, especially product modeling and information exchange; (b) the desired process innovation in existing BC practice.

Internet has created an unprecedented number of new possibilities. The question arose which of those would prove usable in BC ICT to help initiate BC process innovation. Also the question arose as to the manner in which to use and implement those possibilities.

After the first analysis chapters, the research questions were revisited and reformulated. How can state-of-the-art web technology possibly help to realise new, dynamic BC processes that adequately serve the Client's needs (provide value for money) and, at the same time, serve the needs of the BC companies (provide money for value)? How to use the NG Internet to provide both parties (and their application software) with the precise and Semantically rich on-line BC object definitions? How can software vendors take advantage of such a development? What is the possible role of traditional Specifications in newly innovated BC processes?

According to the solution concept, the value adding performance of the BC industry can be much increased by creating an open, freely accessible bcoWeb, meaning: (a) Stimulating the development of domain specific BC Ontologies containing definitions of BC terms structured according to a GARM-based model. BcoWeb's main strength is its focus on the interaction points within BC: the points where demand and supply must be matched. (b) Relating the different BC Ontologies in a national open source bcoWeb. Earlier standardisation efforts tried in vain to provide international standards for information exchange. One of the conclusions to be drawn from the European eConstruct project was that BC is mainly country-specific (inner doors are different in each country), so international standards development can only start after national standards have become available. (c) Stimulating the development of co-operative BC web services that `understand' the BC terms and can be used as references to support communication between all the stakeholders--including the general public--where `smart' computer applications and building documents are also seen as stakeholders.

BcoWeb has many advantages: it serves as a reference vocabulary available fur humans and computers; it is open source, so safe to build upon; it has rich Semantics, provides context-dependent definitions, supports multiple representations, so there are many possibilities for innovative and value-adding applications, amongst others by using web services.

During the implementation phase, a bcoWeb web-based collaborative editor was created which was used to fill bcoWeb. Furthermore, a catalog and an object tree application were made and a simple Specification generation application. The cases showed that such a bcoWeb approach provides the flexible, commonly accessible BC language required for the development of `smart' computer applications that are able to behave as if they really understand the technical language(s) of BC.

BcoWeb and the applications surrounding it have a number of advantages over earlier mainly paper-based mechanisms. Computer applications can use the bcoWeb to share a common view on the project. Providing Internet-based access to the data that underlie the textual documents helps to turn those documents into `smart' applications that are able, like human co-workers, to complete tasks in close co-operation with other co-workers and other smart computer applications, just like they understand the technical BC language. They can complete many tasks regarding information and knowledge management without human interference--though not without human control--thus providing a promising new technology that seems able to drastically improve current BC processes. The primary information process can be best left to the information system, just like in the automotive industry, as that lowers the number of errors.

A contribution of this research: using the GARM results in a kind of Ontology that is much more BC-oriented and that also has a direct economic connection with BC because it simulates the demand/supply process. A second contribution is the emphasis on the necessity of an open source license for the Ontologies that should underpin the information exchange of the whole of BC: that is no place for individual companies' or projects' restrictions. Third, this thesis' use of GARM also allows different views and definitions for the various disciplines. Definitions can be voluntary referenced.

An important contribution is the emphasis on practical applicability for the average BC company--quite unlike research that aims at the top 1% of the BC companies that can afford the big project databases and the most expensive CAD applications.

Specifications will expand their central role in the BC process if they use bcoWeb to become a valuable source of well-ordered and well-accessible information.

What is the recommendation? What should be done? BC should start national BC Ontology developments like demonstrated here with bcoWeb. The start can be small, just a set of Ontologies that together provide enough information richness to enable communication.

An important role is given to information and software providers. Without end-user software, the bcoWeb proposal is just an idea on paper.

Reinout van Rees 2006-12-13