Auckland w78 conference (day 1)¶

This is a summary of the first real day of the 2003 cib w78 conference in Auckland. (Yesterday was a workshop day, so that was kind of day 0)... Claude Bedard and Hugues Rivard: two decades of research developments in building design

An overview of 20 years of research in w78 by looking at their own reasearch group. In 1981 they organised a symposium with the following big points:

• CAD drawing.
• Integration of specialists in the design process.
• Allow computerised help as early as possible in the design process.
• Concern about the low penetration of computers (remember: this was 1981, before the PC!)

There were two distinct events that stimulated research 20 years ago:

• The advent of the PC. Cheap and easy to use. (Reinout: There are complaints now about the research not being practical enough, just try to imagine the complaints when first you had to go out and buy a mainframe!)
• Some artificial intelligence was shown to be usable in construction. You didn't have to rely on just some computational mechanics. And that's good, as design isn't something to be just calculated.

They had an early prototype that had the nice function of using building codes to calculate some parameters for design. (Reinout: wouldn't it be nice to make something like this interfaceable to/from a specification system? In a specification the building codes that are demanded are (should be) written down, so... It might even be an incentive to start early with writing down some information in the specification instead of letting it be the last-minute-job it often is.)

The conclusions drawn were:

• The last 20 years there have been enourmous advantages in hardware and software. Especially the hardware...
• Much research attention is now on integration of disciplines and on integration of projects. (Reinout: *A specification system should allow this integration. Probably a two-way road, being integrated in other systems/disciplines and integrating information from other sources itself.)

The big research targets of the future as expected by Bedard:

• Enabling greater integration.
• Augmenting the capabilities of the human designer.
• Designing for sustainability (environmentally friendly).

Nash Dawood: Usage of space in time during construction

On a building site there's almost always a space problem. There has to be enough free area for moving around. You need space for the scaffolding. The building cranes need operating room. Supplies need to be stacked, etc.

The central idea is to see space as a resource. Just like manpower, crane capacity, etc. And as such, it can be planned alongside the other resources in MSproject. At least that way, problems can be seen and spotted.

For more automated support of the planning phase a tool for calculating the various possible ways of allocating the space was made.

Marja Naaranoja and Paul Clarke: Information and knowledge

To start with a question posed at the end of their presentation: "there are a lot of definitions of knowledge, what is yours?". Answer: "knowledge is organised, processed and understood information".

Information is good, but more information is not necessarily better. You can drown in it. It is important to convert information into knowledge. Or better: information must result in knowledge. And in the end, knowledge should result in action.

• Individuals in isolation can utilise experience, perform conceptual thinking and imagination in order to learn.
• Isolation limits experience, and therefore limits the above.

When you use information systems (can be a lot of things, from outlook to complete content management systems) in your organisation, there are three dimensions on which you can look at it:

Economical

e.g. you need to have the money for the information system and you must be willing to spend it.

Organisation

e.g. are you going to prescribe a new system or do you want it to land more softly?

Technical

e.g. is your company used to the information system? Can they use it? ("Oh, there's an agenda function in outlook... Didn't know that...")

Reinout: this was a really interesting presentation. I'm going to do some more reading in the paper and I think I'll be back on this one. Link to pdf paper

Mary Loo Maher: Agents for data sharing

On the one hand you've got a 3D object CAD drawing (like from archicad), on the other hand you've got your own hand-build 3d virtual world which you use in a research setting to co-operate in with other persons. How do you handle the interaction between the two environments?

The chosen solution is the use of agents. Every object in the environment has its own agent. Every agent is rational, that means thinking

• It looks how the world in its surrounding is looking (through sensors).
• It figures out what it wants to have changed in its surroundings.
• It takes action through effectors.

Using agents in this way is a remarkably nice way of coupling the building model and a virtual world.

Moloney and Thabet: 3D game engines

Two presentations about the same subject: using the readily available 3d game engines (doom, quake, etc.) for nicely visualised environments in which to display buildings and in which you can "walk around" with multiple people.

The second presentation showed the use of the system as a digital mock-up of the actual building in order to walk through it with people from various disciplines, trying to find out problems beforehand.