Friday, 12 January 2018

Using classifications in naming conventions

One subject that regularly comes up when discussing BIM and classification is how these should be including in naming conventions for labels/titles. This conversation could be around the naming of items such as layers, documents or objects.

I have been keeping an eye on this recent Twitter discussion - as I was tagged into it - so I thought I'd gather a few opinions and do a blog post.

At NBS, when developing Uniclass 2015 throughout the industry consultation it was clear that (a) more depth was required to classify a greater number of items – so double digit separators where required and (b) as a separator the ‘_’ was preferable to ‘-‘  or a space so that it didn’t clash with 1192 standard series naming convention.

On these two points, there was always a risk that this clashes with certain standards that require classification. For example some standards may put a maximum length on a classification and some may require certain separator symbols. However, a new unified classification was required for object-orientated BIM and the naming convention rules for 2D document names should not hold the industry back.

Naming convention rules for documents or layers or objects are primarily there to allow for sorting in an alpha-numerical traditional file system/list view/tree view. In a modern digital system, the data itself should be represented as attributes – parsing file names should not be the primary route. For example, if an object represents a 1700mm bath product – the core data should be in an attribute for length. Equally, If a file has been uploaded to a common data environment (CDE) by the company ABC Architects – this should be primarily recorded as an attribute within the CDE. The core data for BIM Level 2 is COBie and the schema for COBie is around objects and attributes.

In this above example, due to limitation on character length the company ‘ABC Architects’ might be shortened on the file naming – to ‘ABC’ for example. Equally, for long classifications, these can be shortened too. For example, a drawing or a layer showing radiators may just have ‘Pr_70_60’ or ‘Pr7060’ if space and limiters was restricted as the classification in the name. But in attributes associated with it – this would be stored as ‘Pr_70_60_36_73’ and ‘Radiators’. Some may simply put the highest level classification in the file name for example ‘EF_60_40’ to indicate the function of heating and cooling for the layer/document/object representing the radiator.

Two examples of this are shown in the illustrations below. It should be noted that in these examples Uniclass 2015 is used, but similar principles could be followed if it was another system such as Omniclass or Cuneco.

Fig 1 - Document naming (no restrictions) and associated attributes with full primary data
Fig 2 - Document naming (shortened) and associated attributes with full primary data
The most important thing for a project team is that the approach to putting classification into a naming convention is documented and understood in the EIR and corresponding BEP. By making this clear at the start of a project then the team will understand the rules. (Ideally the standards and associated guides should give very clear examples so it agreed industry wide.)

Equally, the important thing for a classification system is that it is not restricted by naming conventions for files/objects/layers. It must be allowed to serve its primary use case which is a being a classification system fit for a modern digital world.

Further reading:
An introduction to Uniclass 2015:
An introduction to BS 1192 naming convention:
Uniclass 2015 usage in layer naming (page 10):
Article from Bond Bryan:

Examples in practice:
In writing this post I picked the brains and gathered opinions from Sarah Delany@DRossiter87@bondbryanBIM and @StewartGH1970 (not sure we agree on everything - but thanks for that gents). Also, thanks to Rob for some screenshots of how they are doing this in practice on real projects (click for larger imagery)...

Fig 3 - Using layers to filter content in model checking software
Fig 4 - Using classifications to filter in model checking software

Fig 5 - Classifications attributesin a COBie spreadsheet

No comments:

Post a comment