CEN econstruction meeting 2004-05-13ΒΆ

Tags: aec

Today we had the final meeting of the CEN econstruction workshop. A project meeting in the morning and the final presentation in the afternoon.

I didn't write down everything, but it gives a good enough impression of what happened. The actual powerpoints will be available at the website. Paperwork before the meeting

CWA1 and CWA2 are publicly available at the CEN website. The other three will be put into the CEN publication machinery in two weeks time.

Apparently, there is a possibility to do a revision. At the least, after two years there will be a check whether the documents aren't obsolete yet. But if there is a formal procedure for the updating has still to be checked.

The rest of the time was spend on a quick walkthrough of the presentations. Now on to the presentation of the workshop results in the formal meeting:

Introduction by Reza Beheshti

The tangible results of the workshop are the CWAs, the Cen Workshop Agreements. So documents which the workshop members say they agree upon.

John Ketchell (CEN): ebusiness standards - moving from e-commerce to interoperability.

John's going to be a bit procovative.

CEN/ISSS is formed in mid-1997 and acts as a focal point for ICT issues within CEN (which is the European standardisation institute). eBusiness standardisation is a core activity. There is a wide range of content in the CWAs, meeting the market needs: technical specifications, guidance/best practice, information.

Some CWAs are available free of charge (like the econstruction ones). They had some 120000 downloads since 2002. Since 1998 they had 47 workshops, 12 are open at the moment.

The econstruction workshop idea came from a number of EU projects and started februari 2003.

eBusiness is a mess.

  • EDI is eBusiness since 20 years. It is still fundamental for those companies that have it (only big ones). There's nothing really better yet.
  • We have seen architectures, frameworks, marketplaces, web services... eBusiness standards galore.
  • IT vendors delight in promoting the eBusiness standards - it sells software...
  • Research projects love creating them.

Slowly, but surely, consensus is beginning to emerge:

  • Classification: beginning of generic principles
  • Cataloguing: work on generic principles
  • ebXML: global framework for business transactions, more user-friendly than EDI
  • Less hype and more realism post-dot-com-bust

But there is still an awfully long way to go.

There are at least four competing standards bodies for web services standards that have overlapping standards. Just because all the big vendors are in all standards processes to prevent the competitors from building a killer application. (RR: See also here, which points to an article by Jim Waldo from Sun). This is a big waste of money.

Cen has a draft report and recommendation on ebusiness. Some interesting other workshops:

  • eProcurement workshop. Started on 14 october 2003, open to all. They have a link with ebXML procurement. They don't want to invent new standards, but have an emphasis on clear "who does what" information, recommendations for gap-filling, etc.
  • eCAT workshop on e-business cataloguing.
  • Electronic invoicing workshop
  • EU framework 6 project "COPRAS": a systematic interface between 6th framework projects and standardisation.

A good link: multi-lingual CEN/ISSS site for ebXML.

Michel Böhms (TNO bouw): European eConstruction framework and architecture

eConstruction deals with the lower-level ICT elements. If you couple it with new/improved building technologies, you allow improved business processes. Followed by innovative products/services and at the end impact for stakeholders. So: there is more than ICT, but ICT enables changes on higher levels.

Innovation goes through phases. Emerging technogy, research, development/integration, take-up, use. A lot of what is presented today is in the research and especially the development phase. The actual use is a big challenge.

An important trend is the shift from paper to model-based. Also important: the open standards trend. Here are some interoperability levels:

  • Closed (no interfaces)
  • Open proprietary
  • Open standards
    • National
    • European
    • World-wide

Especially when looking at this last trend, we can safely say that the web is the communication medium of the future.

So, summarising: model-based and object oriented information management and sharing via open standards over the Web.

Back to the data aspects. There are different kinds of descriptions. There is a difference between definitions (the ontologies) and the specifications (instantiations of the definitions). Specifications can be types (catalogues) or occurrences (i.e. models (in the sense of sets of occurrences, like in a model-based drawing)).

Jeffrey Wix (AEC3): CWA3 eConstruction metaschema

A metaschema is used to make a set of fundamental ideas explicit using a formal language. A metaschema can then be used by extension in the actual creation of a schema or ontology that will do something useful.

Model approach layers:

  • Language: a formal language used to represent modeling concepts from within the universe of discource (like Class, Attribute).
  • Schema or ontology: a formal representation of concepts within the universe of discource (like Door, height).
  • Model/occurrences: an instance of a schema. Model is used as the end-user's understanding of the term "model" (RR: there was some disagreement beforehand about this use of the term "model". To many, a model and a schema are almost the same, so then model should be the word used on the previous level. Later on in the presentation there were also a number of slides with "schema/model" on them, so Jeffrey wasn't too consistent either (or I misunderstood)).

Examples of formal languages (=> the metaschema)

  • express
  • uml
  • xsd (or relaxng)
  • topic maps
  • rdf
  • owl

The second level, the schema, is sort of the scale-model of the later actual data. It can for instance be a cost model, a geometry model, etc.

On the difference between type and occurrence. Type specifies the common information to be captured about all occurrence of that type. Occurrences determines positional and identity data and references the type

12006-2 is an very important underlying metaschema for existing and new classification systems. 12006-2 also provides some standard tables of which many are generally reused.

12006-3 is also a metaschema and basically sits above classification systems. 12006-3 allows you translate between concepts. The Dutch LexiCon, the French SDC and the Norwegian Barbi are instances of 12006-3.

IFC, step, episle, plib and bcxml-xtd are further examples of metaschemas.

Some recommendations

  • Languages. We should ideally be able to define our schemas in appropriate multiple forms. Express and owl. Xsd and uml. We need mappings between language forms and also between the resulting data files.
  • We need to add business, trading and procurement standards into our work. It is needed.
  • Communal BC model is what we want. Create an architecture for communal model. Select a metaschema for communal model. Etc.
  • Till now we used express and step mostly. But OWL is a major candidate for future use, which is a clear outcome from this workshop.

Celson Lima (cstb): CWA4 and CWA5 - ontologies and software tools

Main guidelines we tried to keep in mind when we researched ontologies:

  • Purpose What is the ontology created/used for?
  • Promotion Do not re-invent the wheel, but try to promote existing ontologies.
  • Core foundations Support the development of new ontologies by providing a foundation.
  • Harmonisation with the semantic web An absolute necessity.

An ontology denotes a typically shared understanding of a particular domain. (There was some more text in this definition, but I missed it. Ah well, just look in the document.)

Current state of the art regarding ontologies in the building industry:

  • the lexicon (12006-3)
  • bcBuildingDefinitions (lexicon variant used in eConstruct)
  • e-cognos ontology
  • Barbi (norwegian 12006-3)

Lessons learned from e-cognos:

  • Users did not want to handle big ontologies. They did not want 22000 items at the same time.
  • They are prefectly happy if their small (and relevant) taxonomies are in place providing the results they expect.
  • Content is somehow difficult to be standardised. Multiple needs, multiple sources, multiple insights...

Building an ontology is hard work and time consuming. Use computer power to do it, but human "approval" is always required. Sources can be texts, dictionaries, knowledge bases, semi-structured data, etc.

You cannot force people to use ontologies. They are used if the right tools hide the complexity and facilitate use and if they prove their added value for the business. Collaboration among people working in the field is mandatory.

When putting into action an ontology, it is important to evaluate it. For that, a uniform representation is needed. Then you can compare and test ontologies. Owl is the best choice for this.

(Comment somewhere in the sheets that I liked: compliancy with OWL is an added value).

What is missing?

  • Better scenarios to demonstrate the added value of ontologies.
  • A European center to gather all available ontology sources.

As a starting point for the latter, cstb works in the FUNSIEC project which aims at being such a center.

The CWA5 document has a list of software tools in this area.

The biggest problem with the tools is the lack of interoperability, integrating one ontology into the ontology library of another tool. Or merging two ontologies created using different tools or languages.

Panel discussion: Graham Storer, Michel Böhms, Celson Lima, Jeffrey Wix, John Ketchell

Alain Zarli (cstb) was impressed with the results. He missed one thing in the presentations, namely the differing interests of the different stakeholders. Which stakeholders like IFC, which stakeholders need ontologies, which stakeholders have an interest in the metaschema.

Jeff: it is even more complex than that. The same stakeholders have different needs in different parts of the process. We have given some attention to it, looking at views for instance, but mostly it was outside the workshop's scope.

*David Leonard (taylor woodrow): The building industry needs this, but doesn't know it yet. A lot should probably be hidden behind software and service providers. Are they seeing more of this research and implementing it?*

Michel: there is still a lot of work to be done. You need to develop a good model. The ontologies need to be filled (ouch). And existing information providers (like Heinze in Germany) aren't terribly willing to give away their proprietary information.

It is a bit maddening. We need it. But we don't have it. Graham Storer says that the last 10, 12 year the progress has stepped up a few gears. It also starts being a real business issue. It has not bubbled to the surface yet, but it's about to.

John: where is the business case in all of this?

Graham: the building industry is fragmented. Many SMEs (small and medium enterprises). And the bigger ones are SMEs at heart, with many departments and individual engineers. So building is all about information transfer.

Jeff made a prediction that in five years time a company will start to implement all of this and to take a large piece of market share. He's writing a business plan at the moment...

Jeff Stephens (Taylor Woodrow) is encouraged by the rising level of possible interoperability.

Graham: you need some real cathalyst to see any really big change in our industry. Some real treaths. Some real money-saver. Some regulation.

Jeffrey Wix: the biggest civilian client in the USA (GSA) will demand IFC-compliant drawings from 2006 onwards. The panicy shuffling among the contractors was a sight to behold.

Graham starts on monday with an idea of his. He's gotten together some free copies of a good model-based CAD package and is going to put it in schools so that the kids can start designing a few small things with them. Hopefully some of them will have daddies and mummies in influential building industry positions who will start asking questions of their architect: "where is the model" after they've seen the stuff from their kids. Technology push.

Jeffrey: We're dealing with the individual trees, others need to look at the wood as a whole and evangelise that.

Reza: we can be a bit more positive, looking at what Graham said about the uptake was much quicker the last 10 years.

blog comments powered by Disqus
 
vanrees.org logo

About me

My name is Reinout van Rees and I work a lot with Python (programming language) and Django (website framework). I live in The Netherlands and I'm happily married to Annie van Rees-Kooiman.

Weblog feeds

Most of my website content is in my weblog. You can keep up to date by subscribing to the automatic feeds (for instance with Google reader):