Two common misconceptions about classification

One of the most confused topics around BIM is the subject of classification. From my experience, two of the most common misconceptions are:

1. Naming convention vs classification
A long text description with separators that show a glued-together list of attributes is not classification. For example:
‘SJHDes\Extng\Proj101\Gym\Flr\Tmbr’

Looking towards a BIM-world, labelling items with this sort of naming convention should never be done manually. If a label such as this is required it should be generated programmatically from the attributes within the object.

A parallel with this outside of the construction industry would be with music files on a home computer.
Would anyone ever think they have to name all of your MP3s and manually add a label such as
‘RollWithIt\Oasis\1993\BritPop\Manchester\3-49s’?

Would they then go through 1,000s of similar files doing the same exercise?

The answer is clearly ‘no’.

The classification from this example here is simply from the classification table ‘Music Genre’ with the value being ‘BritPop’.

This is an exact parallel to objects in the construction industry. A number of objects are part of larger objects. Each of these objects requires its own classification. For example, a ceramic tile will have a classification (a product), this may be part of a wall covering system that will also have a classification. The wall covering system is part of the larger wall construction which has a classification (element). The covering system is also in a space which is classified (say ‘washroom’). The room is part of a building which could be classified as being an education building and so on.

If an organisation would like a label that is built up from all of this information using their preferred naming convention then this can be automatically generated from the attributes inside the objects within the BIM. Manually creating this is not classifying the object in context. If anything, this is breaking the principle of ‘saying things once and in the right place’. As soon as the object is substituted, if a label has been manually created, then it is risks being instantly incorrect. In a BIM-world the author must ‘store things once and in the right place, but may report this in many places’.

A classification system like Uniclass 2 (based on the ISO 12006 international framework) is simple. It has classification tables for objects such as spaces, elements, systems and products. The project-specific relationships between objects are then authored and determined by the information modelling software. The relationships between objects are not classification – these are database links. In a BIM world, any labels adhering to naming convention rules must be automatically generated from the attributes.

Returning to the music example, the power of adding information through attributes is apparent to anyone who uses software such as iTunes. The user can create a music play list any way they wish:

• Play me all songs by ‘Oasis’ 
• Randomly play me songs from 1993
• Play me my 25 most listened to songs from the ‘BritPop’ genre
• Play me my top 10 most listened to songs from by ‘Oasis’ or ‘Supergrass’ with the from 1993

This same power to query BIM will also come to end-users’ finger tips. For example:

• Report on all ‘ceramic tiles’ products in ‘washrooms’ spaces in my ‘education’ buildings
• Report on all ‘wall covering systems’ in my buildings within 100 miles of ‘NE1’ in zones with the tag ‘public access’
• Report on all ‘ceramic tiles’ products across all spaces in Building X.
• etc…

This is the BIM power that classification will bring.

Naming convention is a different subject.

Fig 1 - A automatic text label from object attributes

Fig 2 - The building information modelling rules that are behind this label

2. Terminology vs classification
Potato – ‘You say po-tay-toe, I say po-tar-toe’.
I occasionally browse the CPIC linkedin group. The sort of criticism that Uniclass 2 regularly receives is around terminology:

• Why the name ‘water supply fittings’? – the industry calls them ‘taps’.
• Why the name ‘rolled covering system’? – the industry calls this ‘wall paper’.
• Why the name ‘intruder detection and alarm system’? – the industry calls this a ‘burglar alarm’.
• etc…

This is not a classification issue. This is a ‘default-text-label-against-the-classification’ issue that can be solved reasonably easily with an array of synonyms. (Being pedantic, the examples above from linkedin in aren’t exactly synonyms either – a tap is one of many types of water supply fittings – it doesn’t have a precise one-to-one equivalent.)

However, the great thing about synonyms is that you can keep everyone happy, users can customise the titles themselves, and the classification can do such a good job that it blends into the background and isn’t actually seen. The classification system allows all of the similar items to be grouped together, whatever names are used.

Fig 3 - A classification allows the grouping of objects regardless of the synonym

The final thing to stress in this example in terms of classification is the need to translate to different languages. An item can have one classification, but different text value in English, German or French – phrases from different languages can be treat as synonyms too. This again reinforced the difference between terminology and classification.