Web services8.6expose application data via the Internet, as discussed in Chapter 4. It is a standard way to allow applications to interact with other applications via the Internet. This opens up a lot of extra possibilities and it gives computer applications a more active, more value adding role in the BC process.
Documents in BC are currently created after-the-fact and then passed on to the next phase in BC, without much possibilities for change or interaction. A good example is the building Specification: a static document, signalling the end of the design phase. This passive crop of documents should be replaced by a more active type of document, a smart document that `understands' its purpose and its content and is able to see that it gets things done. This vision in fact means that smart documents, like robots and intelligent manipulators, are also treated as stakeholders in the process.
In the same way, applications can be treated more as participants, instead of just as tools used in the background by the actual human participants. If applications are conceptually given a more participating role, they can provide a lot of value to the process. BC web services provide a way for documents and applications to become participants.
Web services for BC can take on various forms. Project databases (like big IFC Express databases) can be opened up to Internet access. Calculations can be done on-line, for instance strength calculations based on an IFC model. Service can be provided, by allowing building Specifications to be edited on-line, based partly on input from a project database for instance, with the Specification data in its turn being used as basis for other applications. Rule checking is another promising area, checking solutions against regulations or best-practice databases.
These different types of web services make integration an a priori possibility. Many of the examples mentioned above are now done by hand and after the fact. When the building design is finished, there are three days left for making the building Specification. Only then can a real cost estimation be made. A posteriori integration, as analysed in Chapter 2, limits BC greatly.
Web services are part of the solution concept, as they make integration a possibility early on in the process. Web Services--on a conceptual level--state that every computer program can be integrated into the BC process. This `yes, it can be integrated' enables a greater interaction earlier in the process, leading to less friction (and friction costs) later in the process. Current practice means one-off integration afterwards, mostly by hand--leading to high friction costs.
An important part of the concept is to take a deeply distributed approach at information storage. In principle, information can be stored everywhere, as long as the information items are available using the Internet. Web services, as intended here, also include a simple putting on-line of data files as the simplest way of web services.
Tightly coupling applications and databases is impractical, as the coupling might never be used again. One-way referrals from one application or document to the needed item in another application or document are much more practical. Identify what is needed in order to be able to search for updates to this information.
Existing applications can be fitted out with a web service interface, for the common desktop applications this will mostly allow them to consume data, not make them into an information source themselves. But, for instance, TNO's IFC browser [97] could be made into a web service with relatively little effort, exposing the rich IFC data for consumption by other value adding applications.
Also project extranets (see 2.3) could be turned into an information provider. It is quite natural to start swapping small pieces of functionality of a project extranet for more semantic alternatives: one at a time. Such a swap is a good measurement point, too: it should be better, not worse, than the original solution.
Much additional benefit will be possible if the web services use the bcoWeb Semantics, as without common Semantics the use that can be gotten out of web services will still be limited. The solution concept assumes the use of bcoWeb in combination with web services.
Entirely new applications are made possible by web services. The value-price-cost model [19] [17] needs frequent re-calculations during the process in order to be practical. Applications that perform such calculations are made possible in a web services setting.
The next section introduces the knowledge sharing possibilities that are also newly made possible.
Web services lower the costs (in time and effort) of integration of applications in the BC process. This allows an increase in computer-provided information and--more specifically--knowledge.
Chapter 4 analysed some of the knowledge representation possibilities in the NG Internet. These are also the possibilities that can be integrated into the BC process using web services. The simplest form is knowledge contained in the BC Ontology, like the common parts of a bridge or stadium. Or the different kinds of roofs. More involved is restriction or rule-based knowledge. For instance, the fact flat tarred roofs must be re-done every 10 years.
Often, a (theoretical) subdivision is made in data, information and knowledge. Data is the `raw' data; information is data tagged with meaning so that it is usable; knowledge is `active' information, information that is being used, information that is consciously acted upon. Using this common subdivision for this thesis, Ontologies are used to tag data with meaning, turning it into information.
Knowledge goes one step further--this chapter provides two ways. On the one hand, the use of the FU/TS-mechanism as the Ontology structuring mechanism brings the dynamic supply/demand, question/answer, wish/possibility back-and-forth directly into the Ontology, which facilitates the expression of knowledge. On the other hand, web services interacting with the information and the represented knowledge can themselves provide knowledge by acting on it.
Knowledge provision specific to web services is the knowledge that requires access to multiple information sources, coupled with calculations, lookups in huge databases, etcetera.
This specific kind of web services based knowledge provision opens up huge possibilities for parties in BC that have a lot of knowledge. Examples include standardisation institutes, research institutes (like CSTB, CSIRO, TNO, universities) and, for instance, Contractors that have made knowledge formalisation a priority.
The knowledge present in those institutions is currently only transferable either by consultancy or by paper-based documents. Web services open up the way to make the knowledge available more widely by allowing it to be integrated more directly in the BC process.
Compared to paper-based documents, smart documents also allow a more detailed and comprehensive transfer of knowledge. Paper-based documents are restricted in size because they must be manageable by humans and because they must be economically viable. Web services based electronic knowledge support of the BC process knows no such limits.
The integration of knowledge into the process opens a knowledge market in BC, allowing also smaller participants to offer their--possibly highly innovative and specialised--knowledge in an economically viable manner.
Web services open up new possibilities for value adding in BC. They open up project databases for generic application access, allow service providing, rule checking and a much increased knowledge integration opportunity. The knowledge integration possibility opens up a new knowledge market in BC, making it a more value adding industry.
Human actors can receive much more support during the process, as web services provide a priori integration instead of current after-the-fact by-hand integration. Web services automate what can be automated and support what has to be looked into further by human actors. In both cases, however, the feedback must be speedy.
Web services provide application integration and support knowledge integration. In combination with bcoWeb's network of BC Ontologies--allowing the applications to talk the same language--a big step to an increase in BC's value adding capacity is taken.
Reinout van Rees 2006-12-13