06 December 2016

Where is BIM in Education

Recently I went on-line and had a look at the subject offerings of the four university level design schools in my home town. I wanted to see what they were offering in terms of BIM education, and by extension what could I expect from new graduates.

I could not find any mention of BIM at all - not a thing.

Mind you there was little or no detail information on individual subjects, so there may be BIM buried somewhere. Still, I wonder how a prospective student who wants to get a good grounding in BIM chooses where to go.

But I get the distinct feeling there is no, or only cursory BIM. You can tell from the subject offerings.

The "Computing" major at one school is described as for:
"... students who want to develop strong professional capabilities in programming and development of digital artifacts. You will develop strong technical skills in the areas of media computation, data manipulation and visualisation, interaction design, and usability."
Subjects include Calculus, Algebra, WEB technology and Graphics and Computation. Nothing about virtual construction or building data management.

Another major, "Digital Technologies" has:
"a special focus on digital media artefacts such as web-based media, mobile media, and interactive media."
Again, nothing remotely like BIM. Although this major is no doubt useful for presenting "artist's impressions" digitally. Which is, apparently, quite useful for marketing. I'm surprised they don't have social media in there. Perhaps they haven't caught up with that yet, after all its only been around for 12 years.

Another subject at another school, "Design Communications (Digital)" equates CAD, BIM and graphic softwares as achieving equal outcomes:
"... students will be asked to work through 3 software packages: AutoCAD, Revit Architecture and Photoshop."
which is not surprising because the expected outcome are drawings and geometric models:
"Computation and representation; 2 and 3 dimensional computer models of forms – solids, surfaces; number – dimensions, text; geometry – point, line, plane; scale – proportion, composition, reproduction; and material – hatching, texture, shade. Navigation and scale, perspective and point of view; ordering and referencing drawings, simple rendering and lighting." 
There is no appreciation that Revit can do so much more than merely create drawings. The depressing thing is this subject is flagged as a BIM subject, presumably because it includes Revit.

Some schools even have subjects like "Drawing Studio" and "Architectural Drawing". I can't help feeling that there is this belief in academia that the core skill required for architects is to understand drawing conventions. Drawing may be useful, but surely it is just one of the many tools now available to communicate a design to others.
Of more concern is their belief that the sole use for digital technology is to do drawings, whether archaic technical drawings from CAD, or 3D rendered "artists impressions".
The idea that computers could be used to mimic reality, to create virtual architecture, seems to have bypassed the practice of teaching in academia.


The lack of BIM could just be the conservatism of universities, a lack of awareness of anything new. Or the belief that anything new, while worthy of research, doesn't apply to them.

But it seems deeper than that. BIM software is actively resisted.

I came across a blog post at revitpure.com, Should Architecture Students use Revit? It spoke of the US but resonates with what I see in my home town. Some quotes from architecture students:
"The professor who runs the second year is pretty anti-Revit for her students"
"Some professors strictly forbid it. [...] They did mark down projects that would use the base Revit doors, windows and railings."
I didn't realise architecture involved the designing of custom doors, windows and railings. I though it involved the arrangement of spaces, form composition and satisfaction of user requirements (among other things).

The difference between a building drawn with SketchUP, Rhino or 3D Max and one modelled with Revit or Archicad is that the latter is a closer representation of an actual buildable building, while the former is just a representation of an idea.
Now sometimes a representation of an idea is all that is required. It depends on what is being taught, and what is being assessed. And I can understand the argument that BIM software can sometimes get in the way.
Students obviously have limited skill sets, they can't learn everything at once. So a choice has to be made. Either let them explore concepts and ideas and then teach them how to turn them into buildings, or teach them the skills they will require (the craft) first, and then let them explore concepts and ideas that they can apply their craft to.
In the first case you would initially ban BIM software, in the second you would initially accept boring designs with "base Revit doors, windows and railings".
Most schools use the first method, because concepts and ideas are fun, and they don't want to put new students off. Which is fine. The problem is they don't seem to get around to properly teach the craft of architecture, how to turn an idea into a buildable building.

It might be argued BIM is only suitable for advanced students, those in upper years or doing a masters. But the offerings in postgraduate courses are limited to parametric design, including at one school "designing sonic spaces".
The problem is parametric design is not necessarily, and rarely, BIM. Although it is kind of BIM, it uses associated data just like BIM, it is not necessarily BIM - a representation of a buildable building. It maybe of just planes, it may not be supported, how it integrates with the rest of building may be unresolved. A whole lot of other stuff has to be done to turn an algorithmic design into a piece of architecture or a real structure that is buildable.


Let's face the elephant in the room - BIM is boring.

After all, BIM is just another way of doing things that have always been done.

Sure BIM does things a little better - once you know how to use it. Like being quicker, reducing errors and leaving fewer things unresolved. But surely these things can be done just by being more competent, being more careful, spending a bit more time, by "talking to each other more". Why learn (or teach) a whole new way of doing things for such an unexciting outcome?

Yes, BIM, in itself, is boring. The softwares are complex and take time to master. Dealing with rows and rows of data is boring. Having to think about how high things are, what level they are on, what a wall is constructed of when all you want to do is a plan is tedious. Not being able to use a window or a piece of joinery as a door is annoying.

But there is a pay off.

When you use BIM software to model you are creating a building. Not a real building mind you, a virtual building, but still a building. When you use CAD or 3D modelling software you are only creating drawings or geometric models, a representation of what a building might look like. With BIM it is what it will actually look like. Much more exciting. Something I would have thought valuable in an educational context.

I don't understand why educationists haven't jumped on BIM as way of teaching. Take architecture; using it as a tool to show students how buildings fit together, as a way of assessing what a student's design will actually look like as a building (rather than what they imagine it would look like). Likewise structural design; BIM can link analysis with a model of structural elements directly. Change either and the effects are immediately obvious. Similarly for MEP. Surely that is a fantastic teaching tool.

I appreciate the BIM software we have doesn't always make this that easy or that smooth a process, particularly for the engineering disciplines. But where are the people in academia that are excited about the possibilities? Excited enough to actually use it, rather than just write research papers?


The four schools in my town is not a big sample so I did a bit of research looking at papers on BIM Education. It is quite depressing.

Many are just surveys. Australia based NatSpec have an annually updated report, The International BIM Education report which is a compilation of  responses from "a global group of parties". It is a compilation of expert opinions rather than a comprehensive survey, but provides a good over-view.

A recent paper (May 2016), BIM Curriculum Design in Architecture, Engineering and Construction Education: a Systematic Review published by the US based Journal of Information Technology in Construction provides a review of published papers on BIM Education. One of its findings is that:
"the number of AEC programs that offer elective BIM courses is significantly higher than programs that require BIM courses for a degree."
It seems academia is treating BIM as an "extra", not a core skill required by future construction professionals.

Here in Australia there has been a serious attempt to progress BIM Education. Their web site codeBIM.com has a range of resources and the start of a systematic approach. Unfortunately it seems to be in abeyance. References are quite old, and the last active initiative was in 2010 when they had a one off grant from the federal government Office of Learning and Teaching.  They have their own survey of BIM education in Australian schools, but it is from 2011.
The main focus of this group is (was) in the collaborative aspect of BIM, where students of different disciplines do a group project together. While a valid part of BIM education it is not the only aspect, and probably one of the most difficult to introduce due to the required coordination across departments.

The feeling I got from the limited papers I read is that BIM education in academia is going backwards. Initiatives started years ago didn't develop into anything permanent, one-off pilots never seem to go anywhere.

Is it just all too hard?


BIM is a broad term and means different things to different parties. To some it is the process that makes BIM, to some a container of information about a building, to others the data that can be extracted. To an architect BIM is a model, to a facilities manager it is data about equipment.
Not all aspects have to be taught to the same level to everyone, although everyone should be aware of what others require.

In general BIM can be divided into three broad categories:


Virtual Design & Construct is what is sounds like. BIM software is used to create a virtual building, then that virtual building is used to inform and manage construction.
A Revit or ArchiCAD model is a VDC model. A CAD file is not, a 3D SketchUP or Rhino model is not. VDC models not only have data embedded in them; a door knows what its fire rating is, they also have intelligence; a door knows which wall it is in and what floor it is on.
VDC models can be used for analysis. The virtual building can be tested before it is built, from thermal performance to crowd behaviour. And all this testing means designs can be optimised. Instead of just "doing a design", many alternatives can be tried and assessed to get the best outcome.


Collaboration is an ideal rather than a thing. The idea is that if VDC models are shared, or constructed together cooperatively, there will be greater integration of the various skill sets brought to the building by the different parties involved.
Each participant incorporates their work into a VDC model so that others can use the combined information in the VDC model to inform their work.
If the architect can see the mechanical engineer's ductwork in their model it is easier to ensure their ceilings accommodate them, conversely if the mechanical engineer can see the architect's ceiling and structure above they can work towards their ducts fitting within the space between them.

BIM Management

BIM Management covers the management of BIM processes, and the management of BIM data. There is overall BIM Management, the process of coordinating all the participants "collaborating". This extends from BIM Management Plans to novel contractual arrangements like Integrated Project Delivery (IPD).
BIM data management is the managing of data structures, extraction and formatting. The skills needed for estimating and facilities management, to name a few.


BIM Management

BIM Management is probably the most advanced area of BIM education. Students are not going to be going straight into BIM Management roles on graduation so it is a 'need to know' rather than a 'need to do' subject. Economical lectures and reading assignments are generally adequate to cover it. Guest lectures by unpaid industry lobbyists even cheaper.

Because it is so easy to provide there is a danger BIM Management education goes too far. Understanding esoteric standards that no-one actually uses in practice is pretty much a waste of time.
As are the ravings of BIM evangelists on the Utopian future that their brand of BIM will bring.

As this is the only BIM most students are exposed to it is not surprising the majority of BIM research done by postgraduates has little relationship to actual BIM use in the industry, not withstanding the unending LinkedIn requests by students and researchers to fill out BIM use surveys.


Typically Collaboration subjects create multi-disciplinary teams, for example students of architecture, structure and/or construction, and get them to work on a single project or task. For educators Collaboration is the least boring part of BIM, it is just like the real world the students will soon be working in.

Except it is not really about BIM. Collaboration subjects can be run without BIM. There is no reason they can't use CAD, or hand drawings for that matter, and still collaborate.
I suspect this is why Collaboration subjects have gained some initial traction. People ignorant of BIM can see the benefits without having to understand the BIM part of it.

When BIM processes are used - sharing VDC models and data, there are BIM education aspects to it. Learning to create models others can use, and how to deal with the models of others is a necessary BIM skill. But not as fundamental as VDC.

I suspect the failure of  Collaboration subjects to catch on is not entirely due to logistics of coordinating the timetables of several disparate courses (the usual excuse). The lack of competently created VDC models by participants, who would have had very little education in VDC model creation, have a large part to play. Then add in the novelty of working for the first time with other disciples (these subjects are only ever run once). I'm not saying these programs were a failure, just that by themselves they seem to not have inspired enough confidence to progress BIM education.


VDC is the core of BIM. Without VDC models BIM is not possible. And no-one can fully comprehend BIM without an understanding of how VDC models work.
So even if you ignore the fact that future design professionals will need VDC skills to get a job, to teach BIM you have to teach students how VDC models work, and the best way to do that is to get them involved in making VDC models.

It is not about learning a particular software, any proper BIM authoring software (and there are many which claim to be but are not) will be adequate to teach the fundamentals of BIM.

The problem is many in academia treat BIM authoring software as if it is another version of CAD. It is not. It is fundamentally different. CAD is so dumb BIM software can print CAD files just like it can print PDFs or paper drawings.

When you teach BIM using software like Revit or ArchiCAD you are not merely instructing students on the commands required to produce lines or 3D solids, you are teaching them fundamental concepts like what a door is, that a wall is not the same as a roof, that stairs have limits on their dimensions, that if you remove a column the building will fall down.

Academia has to get away from the current practice of teaching BIM authoring software to merely create drawings or geometric models. To stop churning out graphic artists and mathematicians and start producing building design professionals. Professionals with a good understanding of their craft.


Because VDC is so fundamental it can not be treated as an extra subject. It has to be integrated, it has to replace existing programs.

Within subjects where BIM can be used in the real world the teaching of BIM processes should replace traditional processes, not merely added as an optional extra. I find it strange that academia, who should be looking to the future, use the fact that BIM is not yet currently universally used in industry as an excuse not to teach it.

Anywhere that drawing or graphics is taught should be replaced with VDC subjects. "Communication", graphic and drawing subjects need to be replaced with VDC modelling where the emphasis is on creating virtual buildings rather than representations of buildings.
Computer and "Digital" subjects must include skills beyond mere design generation, and the new kid on the block, Virtual Reality (VR) or Augmented reality (AR). Data management skills, using simulation and analysis, are far more useful than just being able to generate parts of an overall design, or represent your design "interactively".

Frank's sketch of Bilbao

Will that mean the end of drawing? No. Drawing is still, and will always be, the quickest way for a person to communicate an idea. But to turn that idea into a building, into Architecture, the quickest way is to model it as a building. Not do more drawings.

Bilbao as a model
VDC software doesn't mean the death of drawing. These softwares easily create drawings. They may be too information rich for some, but they are still drawings.
I'm old enough to have used log tables and slide rulers in exams. By necessity exams questions had to be kept simple because of the time calculations took. Now calculators are allowed and questions can be much more sophisticated, and by extension more complex concepts can be taught. VDC software does a similar thing. By removing the need to draw everything more time can be spent on design issues and analysis. Letting software create the output also standardises results across students, making it easier to assess design ability rather than graphic skills.


As I touched on above, I don't understand why educators are not considering using BIM processes to extend and improve what they do.
The same model checking and BIM processes we in the industry use can be used to assess student's work.

If students submit a VDC model then:
  • it can be viewed from any angle, it can be cut and sliced to easily check for completeness and analyse how it works. 
  • schedules can be created to check the brief has been met: area schedules, room schedules, efficiencies, etc. Also code requirements: ventilation, daylight, minimum door widths, fire rating etc. 
  • model checking can be used to assess code compliance, realistic spatial allowances (e.g. stair widths, wall thickness, structural and duct sizes), etc.
  • analysis used to check energy use, daylight and sunlight penetration, fire performance, crowd behaviour, etc. 

Just like in the real world use of BIM can improve the quality of student's work by making deficiencies more obvious, and as a bonus a lot of checking tedium can be dispensed with.

This is what I don't get. Academia, for example, will do research papers on Singapore's use of BIM for automating building code compliance, or even BIM Based Architectural Design Quality Checking, but it doesn't occur to them to do it for themselves - to use BIM to automate their own processes.

The secret of BIM is you use it for your own purposes, you don't specifically do it to benefit others, because just the act of creating BIM means you create something that others can use.
For BIM to work it has to be evolutionary - each step has to produce immediate benefit for it to progress. That starts with education. Use BIM for educational purposes and the students of that education will be BIM proficient. There would be no need for separate "BIM education".

For BIM to take hold in academia what we need now are some trailblazers to show how BIM can benefit the needs of education.

And for pity's sake stop bugging us with industry BIM surveys, spend some time looking in your own backyard.

30 September 2016

Making BIM Work: Quality Models

BIM processes only work if there is something those processes can act upon.
No BIM models, no process.
And quality matters: without good quality models no matter how good BIM processes or standards are it will be extremely difficult for anyone to do anything useful.

Pretty basic stuff, but all too often ignored.

Ignored because to ensure these thing happen action has to occur at the very, very, beginning of a project. When each design consultant is signed up, because they are the BIM authors, the ones who will be creating the BIM models.
Ignored because owners assume BIM authors will produce adequate BIM models as part of their normal service, even when the contract deliverables are only drawings, schedules and specifications.

Now, one day this will hopefully change. Consultant agreements will by default contain common, widely understood BIM requirements. Consultants themselves will be familiar and comfortable with BIM software and what is necessary for quality BIM models.

But at present, pretty much everywhere, this is not the case. And it is not going to change if we don't start addressing this issue directly.

Why is there a Problem with Models?

The reason it is so difficult to get quality BIM models from design consultants is that they think their job is, for architects, to produce drawings, for engineers, to produce diagrams. So they use their software, whether BIM or not, to only produce drawings. Also they generally don't use their BIM software to produce schedules. As they see it their deliverable is a paper schedule, maybe an Excel spreadsheet. So why use your drawing software?

All this is because drawings have traditionally been their tangible deliverable, and is still the main contractual deliverable even in notionally "BIM" projects. Due in part because the reality is that drawings are still the legal documents that contractors use to construct from.
Mind you there are good reasons why drawings are used for legal evidence. All information on drawings is visible and unchangeable. Explanatory text can be included, status, revision sequence and issued date are all clearly displayed. It is very hard for someone to say "I didn't see it".
One day BIM models may be able to do these things, or things that achieve the same outcomes. There are examples around that do some of these things, or something similar, like Bentley's 'Hypermodel' functionality, or the open source BIM Collaboration Format (BCF).

But presently there are no common, robust, methods that match the certainty of drawings.
So are architects and engineers justified in using their BIM software to just produce drawings?

BIM Software is Designed to produce Drawings

The softwares we use today to do BIM were not originally designed to do BIM. They were designed to produce drawings.
When ArchiCAD came out in 1987 the way it worked was that the building was modelled in 3D up to a point. Once it was decided to move to drawings, plans, elevations and sections were created as separate files from the model and worked over to turn them into drawings.
This is a common approach. SketchUp does the same with its separate LayOut program.
Revit came out in 2000 with a similar functionality, except that plans, elevations and sections remained live. Everything was in the one file, so changes in the 3D model instantly appeared in all views created for drawings. But the purpose of Revit was still to create drawings.
Other software now used for BIM started life as CAD programs, with gradual 3D functionality added to assist drawing production (e.g. Bentley, and the now defunct AutoDesk Architecture).

BIM as we know it today came from the realisation that the integrated 3D model that these softwares produced could be used for other purposes. In practical terms BIM is what these softwares are capable of doing, despite the efforts to extend BIM into the realm of fantasy by the standards wonks and BIM evangelists (see my previous post on standards).

So if you use BIM software as it is intended to be used it will produce drawings for you. There is no need to "take shortcuts" to produce convincing looking drawings. And if you use the software properly it will be BIM ready, it will not "take more time" to do BIM.

The Benefits of BIM to Authors

Much is made of using BIM models for 4D (construction sequencing), 5D (quantity measuring), 6D (life-cycle management), and other 'D's. They can also be used for analysis and simulations, particularly in engineering - structural analysis, power circuits, mechanical systems etc.

What is often not appreciated is that a BIM model can also be used for quality assurance (QA) purposes. Checks can be utilised that minimise design errors, that ensure the model is in fact a quality BIM model and an accurate representation of what is to be built.
If all you produce are drawings and separate schedules, you can only check drawings and schedules.
If your BIM models are created properly you can use the model to do the checking.

If your walls contain their fire rating as data in a parameter you can colour code those walls, the same for fire rated doors, fire rated dampers etc. Which makes checking that the correct walls and doors are in the right place much quicker than trawling through multiple drawings and cross referencing schedules.

If your doors contain size data you can run a model check for doors with heights or widths below minimum required values. Check concrete walls required to be fire rated to ensure their thickness achieves their rating. You get the idea, the list is endless.

Revit Warnings

These checks can be manual (e.g. use Filters in Revit or the free add-in Color Splasher to color-code  views by object parameter), or automated (e.g. Revit's inbuilt Warnings,  Autodesk Model Checker for Revit, Solibri IFC model checker).

And because drawings and schedules come directly from the model they will be correct if the model is correct.

When I say correct, I mean correct information. Letting software create your drawings and schedules means forgoing some control over graphic representation. But then BIM authors are architects and engineers, not graphic artists. The question they need to ask is not does this drawing look "neat", but can it be misinterpreted? Will the contractor build the wall using the wrong material or in the wrong place because the lineweight is not exactly right, the hatch pattern doesn't align perfectly? Will they mistake a grid for something else because its head doesn't perfectly align with other grids?

The bottom line is that it is possible to be much more thorough when checking a model as compared to checking drawings and schedules. It can also be done much quicker, especially if standardised automated model checking processes are implemented alongside manual checking. All this leads to less errors in documents, meaning less time wasted dealing with mistakes, both internally and on site.

Looks OK in 2D plans and elevations, but an on-site mistake waiting to happen
There is no real excuse for design consultants to NOT produce good quality BIM models. They should be doing as part of their normal duty of care.

Why won't People Share?

Producing good quality BIM Models is one thing, but if these models are not shared BIM processes will fail.

The comments above about checking the model instead of drawings and schedules extends to other people's work. It is much easier and quicker to see if structure is aligning with architecture if both models are linked together and viewed in 3D. And there are softwares that can automate this checking. Revit has a built-in clash detection ability, Naviworks and Solibiri are specialised software for doing this type of checking.

Design consultants, particularly architects, generally don't like giving their native models to anyone (a topic I covered in my post IP - it is not all yours, get used to it). They see them as their property. The justification is that their contractual deliverables are completed drawings, schedules and specifications. BIM models are their "internal working documents".
Design professionals are also generally paranoid about having their ideas stolen, which they extend to the documents they produce.
And some have this view that as initial author of BIM models they have the right to total control of that model including getting paid whenever anyone makes use of it.

None of these justifications are valid. They just need to get over the fact that in the 21st Century drawings are no longer their only deliverable, and that current legal protections easily extend to cover other deliverables.

However there is a mistaken belief (and not just by design consultants) that handing over models means providing an untouched copy of the model, still containing all its housekeeping and drawing creation setup. None of this is required for BIM Uses. It should be removed - as a requirement. No-one wants to trawl through someone else's rubbish, and allowing people to recreate the drawings of others is a legal minefield.

What is a Quality Model?

In simple terms by quality model I mean a model that is:
  • Fully modelled in 3D.
  • Is modelled as it will be constructed.
  • Uses correct categories and types.
  • All objects contain data about themselves.
  • Data is consistent and coherent.
and taking into consideration the fact that drawings and schedules are contractual deliverables:
  • The model must match issued drawings.
  • Data in the model must match issued schedules.

It is not so much about WHAT information is in the model (which most standards seem to concentrate on, including LOD descriptions like the BIMforum LOD Specification), but that the information that is there is complete and can be relied upon.

This is where discussion of BIM becomes confused. Many believe BIM is about extra work and extra data. It is not. It is about data that is produced for normal purposes being in a consistent format.
The final format doesn't even matter. If data is consistent it can be converted from one format to another. If data required for COBie exists in a model it can be converted to COBie format on export. It doesn't have to exist in the model in COBie format.

If someone wants data that is not usually created as part of your normal service then it is an extra cost. An architect may put minimum warrantee requirements in their specification, but if the owner wants the manufacturer's actual warrantee information in the architect's model that is work they would not normally do and so is an extra.

In short a quality BIM Model is one that has been created by people doing what they normally do and using their BIM software the way it was designed to be used.

Not such a big ask.

How to Obtain Quality Models

As explained above it can't be left solely to design consultants to initiate quality models. To be fair that is not all design consultants, there are some who are very good. And not yet, one day it will become standard practice, but for now owners have to be proactive.

There are two places requirements can be spelt out: - consultant engagement agreements and a project's BIM Brief.

Consultant engagement agreements are better because they are contractually binding, whereas a BIM Brief may or may not be, depending on what is in consultant engagement agreements. Also the BIM Brief may not be completed before consultants are engaged (particularly if those consultants are expected to participate in creating the BIM Brief).

Generally best practice is to include generic BIM model requirements in consultant engagement agreements, with specific requirements, and perhaps specific examples of good modelling practice, in the BIM Brief.

It is also best practice to embed BIM requirements within consultant engagement agreements and not simply have a separate "BIM Addendum" or "Exhibit", which can lead to contradictions and perpetuates the belief that using BIM is a separate service.

A good approach is to include BIM requirements in a consultant's project scope. This has the advantage of being easier to understand as a lawyer is less likely to have authored it, and can be tailored to a specific project. Consultant scope is also more likely to be available to those actually working on the project, whereas consultant engagement agreements tend to be withheld as they contain sensitive commercial information.

Current BIM Engagement Documents

A number of organisations have produced contract addendums for BIM.
In the US there is the AIA Document E203-2013, BIM and Digital Data Exhibit by the American Institute of Architects, and in the UK the CIC/BIM Protocol by the Construction Industry Council.
There are also commercial documents available like the US Consensus Docs BIM Addendum.

None of these documents adequately address the issue of model quality. Some are better at addressing model sharing than others, but then introduce unnecessary complications. And some are simply impenetrable for normal humans, those who have to implement them.

The AIA E203-2013 is more like a BIM Brief, or BIM Execution Plan, it describes BIM processes rather than modelling requirements.

The CIC/BIM Protocol is mainly about sharing of models and delivery. That is if you can understand it. You would think by the 21st Century lawyers would have learnt to write understandable English. There is one sentence of 130 words with the only commas dividing up lists of items.
It also has other issues that conflict with standard head consultant engagement agreements; like diluting duty of care, and adding things that may not be in it; like assuming the owner has taken ownership of everyone's IP. Read more on the limitations of the CIC/BIM Protocol in the research paper by Kings college London, Enabling BIM through Procurement and Contracts.

I'm not saying these documents are useless or dangerous (although I'd be careful of using the CIC/BIM Protocol), but they are not enough to ensure quality BIM models.

What to put in Model Author Agreements

By Model Authors I mean anyone who is going to create BIM models. This may include design consultants, sub-contractors, construction consultants, and possibly FM consultants.

I've only discussed model sharing and model quality above, but there are other issues that should be covered within contractual agreements. The minimum that an agreement should cover includes:
  • participation in BIM planning
  • provision of adequate resources to achieve BIM
  • model sharing
  • model quality
  • model use

Some examples:

BIM Planning

The Project BIM Briefing Plan forms part of the building brief and must be complied with.
When requested the Consultant will participate in the process to develop and update all project BIM Management Plans and will comply with these plans.
When requested the Consultant will attend BIM Planning meetings, Coordination and Clash resolution meetings.


The consultant will provide, at their own cost, all software, hardware and training required to comply with their project BIM requirements and responsibilities.
A person experienced in use of the Consultant’s main documentation software will be appointed Discipline Model Manager and be available to attend BIM meetings and address BIM and software related issues raised by other project participants.

Model Sharing

All models the Consultant creates for the project, and all exports from those models, shall be made available without restriction to all other project participants.
The consultant may, and is expected to, remove all elements, options, views, imports etc. that do not contribute to issued drawings and schedules from models before issuing them. They may also remove all titleblocks, sheets and layouts used to create issued drawings.
It is acknowledged that the Consultant retains Copyright of their authored models.
The Consultant will respect the rights of authors of models issued to them and will not use those models, or parts of those models, for purposes not directly required by the project.
The Consultant will not provide to third parties models issued to them by others without the consent of the model author.
The Consultant will not print or export contract drawings or schedules from models provided to them by others. If drawings or schedules are required they must be requested from the original model author.

Model Quality

The Consultant will ensure the information in issued BIM Models exactly matches information issued as drawings, schedules and other related documents. This requirement only extends to information that has been modelled or placed as parameters in model objects. (i.e. excludes 2D details and data linked to the model then used in schedules).

Model Use

The Consultant accepts that models issued by them will be used by others as a source of information for work the Consultant is responsible for.
The Consultant will make reasonable endeavours to ensure their models are adequate for purposes others may want to use their model for. However the Consultant can reserve the right to seek compensation from others if it involves work additional to their normal service.

Example Model Quality Requirements

To be really clear specific modelling requirements can be spelt out. They don't necessarily have to be contractual requirements, they can be presented as expectations, or examples of acceptable modelling practice. After all, we are talking about good modelling practice, things that should be done by competent professionals anyway.

This list is generic enough to apply to all those who author models and the different BIM authoring software they use. It is by no means exhaustive, and can be augmented by specific requirements.

  • All project participants shall use compatible software to facilitate model exchanges. If a particular software is used by multiple participants all shall use the same version and all shall keep it updated.
  • A control model shall be created for levels, grids and shared coordinates. This shall be used by all model authors to establish common baseline information.
  • Grids shall be to nearest 5mm increment apart to 10 decimal places, and shall be absolutely orthogonal to 10 decimal places, or if not orthogonal at an angle with no more than 2 decimal places (exactly 2000, not 1999.099899, nor 2000.00006789, and 90° not 89.99967°, or 32.45° not 32.453678943°).
  • Correct categories shall be used, or layers / types etc. will be named to identify the type of object. For example, beams must be modelled as beams, or identified as beams, and not as floors.
  • All model elements in authoring models shall be in the authoring BIM software format. Imported geometry of a format different from authoring software shall not be used for parts of the building the consultant is responsible for.
  • Modelling shall, where possible, match construction methods. For example walls go between floor slabs, not through them.
  • All 3D models shall be consistent with issued 2D drawings.
  • All parameter data shall match issued schedules. This includes, but is not limited to, Area schedules, Revisions, FFE, Wall types, Equipment schedules.
  • All tags and identifying marks on drawings shall match parameter data within the objects being tagged or identified.
  • Text notes shall only be used for general noting or where applicable to multiple objects. Where notes refer to individual objects tags shall be used.
  • Deliver 3D models as separate files per discipline with the same base point.
  • All 2D/3D drawings/models used as references in issued drawings shall be provided with the host file. Pathing of linked files shall be relative and not absolute.
  • When requested provide any associated databases with the models that are linked to the unique component identifiers (i.e. such as external databases for door schedules or steel part / assembly numbers). Provide information on how to access these databases.
  • When requested editable 3D geometry and data shall be issued in native authoring formats (e.g. RVT, 12da, .DWG, .DGN, Moss Genio, ASCII etc) as well as published formats (ie. .PDF, .NWC, DWF etc).
  • Regular exports shall use pre-configured settings to ensure consistency of output. For example “Export for Coordination” view / settings to show only the elements that are to be shared for coordination purposes.
  • Ensure that the exported models retain unique element identifiers (i.e. that there is a globally unique identifier associated to each element that will not be duplicated by another element in the model).
  • Ensure that all elements are modelled as individual selectable elements rather than multiple elements modelled as one element (e.g. don't model a row of columns as a single column element). Nesting or grouping where individual elements are still selectable is acceptable.
  • Where appropriate typical groups of elements can be grouped and copied around the model. There should be no groups with only one occurrence. 
  • Elements, including groups and nested components, are not to be mirrored where doing so creates a different product. (e.g. a dishwasher with an outlet on the left is a different product to a dishwasher with an outlet on the right). Mirrored versions are to be a completely separate element, group or nested component than the original.
  • Main construction elements (walls, columns, slab edges etc) and setouts are to be perfectly orthogonal or at angles no greater than 2 decimal points (e.g. 31.65°).
  • All dimension entities must be rounded to the nearest 1 millimetre, no higher (or rounding errors may occur in strings of dimensions). Dimension values shall not be overridden.

For those of you who know how to use your BIM software the things listed above will be seem pretty basic and obvious. Hopefully there is nothing that you are not already doing. But sometimes the simplest thing can prevent models from being used.


  • BIM requires quality BIM models.
  • BIM requires BIM models to be shared.
  • BIM software, if used as intended, will produce quality BIM models.

If you are a consultant or sub-contractor who authors BIM models review how you are using your BIM software. If those using it in your office are treating it as a drawing tool rather than a modelling tool then retrain them and introduce processes that ensure quality models are produced.
Accept others require access to your BIM models, and when providing those models make sure they only contain information you would normally be providing anyway.

If you are an owner, or contractor who engages design consultants and/or sub-contractors, review your engagement agreements and include minimum BIM and modelling requirement, and the obligation to share models. Don't rely on those you engage to do it for you unless you are certain that they will. And if they complain it will cost more find another consultant or sub-contractor that knows how to do their job properly.

25 July 2016

What makes a good Office BIM Manager?

Many professional design firms and construction sub-contractors are being forced to become BIM authors, with the expectation they can manage and provide BIM deliverables.
They have to use BIM software, which is only efficient if it is genuinely managed. If used properly many things can be done quicker and with less error, but if not project teams can find themselves trapped in a nightmare of tedious tasks, repeating work and redundant effort. Leading to missed deadlines, error filled documentation and very unhappy clients.

There is gradual appreciation of the need for the skills of an Office BIM Manager, but not much understanding of what the role entails.

The role of Office BIM Manager is different from an FM or construction BIM Manager, who manage BIM coordination rather than BIM creation. Of course they are vital for BIM success, but their role, tasks and responsibilities are different.

Unfortunately not all AEC firms appreciate the need for an Office BIM manager, nor understand the benefits a good Office BIM manager can bring.

Often a recent graduate who is "good with computers" is given the role, or a young drafter who has recently used BIM software in their course. These people may become good BIM managers, eventually, with experience. But for now they have no understanding of the profession they work within; what core services the office provides (unless it is a drafting company drawings are not a core service), what the purpose of deliverables are (what is being communicated), and that the number of people and the time a task takes is important (to profitability and therefore their firm's future).

As with any role there are those who are better at it than others. But what I see at the moment is a lack of understanding about what an Office BIM manager should be, and could be, doing.

BIM is not CAD

It has been common practice to simply change the title of CAD Manager to BIM Manager, without changing the role or responsibilities.
But CAD has only ever been about drawing production. CAD can make drawing production more efficient but can do little to improve accuracy or consistency of information. Whether a drawing is hand drafted or computer generated, it is still a drawing.

Drawing and CAD - same information, just neater

But you can't issue a hand drawn BIM model (or a CAD file as a BIM model for that matter).

BIM contains more information than drawings
Nor is there enough data in CAD for automated QA processes. CAD doesn't manage cross referencing or revisioning. You can't query a CAD file to check if any doors are lower than the minimum allowed under regulations; nor colour code fire rated and acoustic walls, as well as the doors in those walls. And CAD does little for the efficiency and accuracy of schedules, including ensuring consistency between drawings and schedules.

BIM introduces new processes that CAD never had to deal with, and the traditional CAD manager was not involved in. For example QA. You can't give a BIM model directly to a senior designer for them to mark up with a red pen. QA has to be part of the BIM process itself.

CAD Managers have been around for 30 or so years now, so there is a lot of experience. But not all make the transition to BIM. They can in fact be an impediment to BIM as they bastardise BIM software to implement CAD workflows and practices. Introducing complicated workarounds that achieve pointless results, sometimes making BIM processes impossible to implement.

Initially this is seen as a positive. The office, particularly project leaders, designers (including engineers) and directors can all continue working as they have always done. They can can ignore BIM.
But soon it becomes apparent the expense BIM software and powerful new computers the office paid for are not producing the efficiencies they were promised by the BIM evangelists. It seems to take more time to do things, not less. And the documents produced are no more accurate than they were when CAD software was used.

Then the office gets hit with a BIM deliverable. The client wants Navisworks or IFC deliverables. They expect coordination to use clash detection. They expect the to be able to use the model for costing. The client has been told all this is possible if BIM is used.

The office is using BIM software so made claims in their (successful) submission that they use BIM. But the BIM (CAD) Manager is now telling them it will require additional resources to deliver BIM requirements.

Accusations starting flying. The client is unreasonable, the BIM software is useless, BIM is an unnecessary impediment forced onto the industry by inexperienced academics...

But just maybe, maybe, BIM is not being managed properly.


BIM software was never intended to merely produce drawings or 3D models. It was intended to provide a single resource for documenting - explaining and communicating - a designed solution.
If you are only using it to produce drawings you are using a fraction of its capabilities.

Much is made of external BIM requirements; owners using BIM for facilities management, contractors using it for clash avoidance, estimators using it for costing. But there are a lot of BIM capabilities that can be utilised internally, within the office that authors it.

And here is the secret to BIM - if you use BIM yourself, for your own purposes, it will also satisfy external BIM requirements.

If your schedules come from the BIM model then there is sufficient information for owner's FM, if you model in 3D it is suitable for clash detection,  if you include materials for tagging and scheduling it is suitable for costing.

That is not to say owners and contractors won't still make unreasonable demands.

Although the data for a COBie deliverable for FM may be within your model, creating the COBie output is not part of designer's core work so is extra. Modelling every bolt and nut, every penetration smaller than 25mm, or concrete construction pours is unreasonable. Including the Quantity Surveyor's cost codes in your model is you doing their work for them.

But if all your core deliverables are being produced using BIM processes these extras are easy to identify, and to justify as extra.


Another thing about BIM software is that is was not designed to produce BIM outputs for others. They were designed to increase the efficiency and accuracy of the user.
BIM wasn't on anyone's radar when ArchiCAD was developed in the 1980's, even when Revit was developed in the late 1990's BIM wasn't talked about (Revit is an amalgam of "Revise it" - software to make revising a design easy). BIM became the rigeur de jour only after a critical mass of users existed and the collaborative possibilities began to be explored (and AutoDesk, then buildingSMART, started using it as a marketing tool).

So at its core the BIM software you have is designed to make your work more efficient and with less error (unlike BIM standards - but that's another story).

But software is just a tool (or in the case of BIM software a suite of tools). Tools used incorrectly or inappropriately will not perform as promised on the box, and can be downright dangerous.
And it is not just the tool that needs to be used properly, the environment it is used in must be appropriate. Using a chainsaw while on the top rung of a ladder sitting in a muddy puddle on the side of a hill can be catastrophically inefficient. Like using the wrong tool for the circumstances:

handing a man hanging from a branch a saw.

Just as it is for BIM software used within an environment designed for CAD.

An opportunity often overlooked is to take advantage of what BIM software can do. How the power of BIM software can be leveraged to make your office more efficient. To do more with less, to offer more services, to produce a better product.

A good Office BIM Manager doesn't just have technical knowledge of how the software works, they organise its use to improve office work practices and work flows. They mould the environment the software operates within.

This means a good Office BIM manager must be involved in more than just technical support. They must also be involved in advising management. And not just in things like the office "CAD Manual", training, hardware and software selection. They need to be included in resource allocation, task allocation, deliverables scope, deliverables timetable, consultant appointment, consultant coordination, and most importantly QA (Quality Assurance).

In short an Office BIM manager should be viewed as a CIO or CTO, not head of software support.
And an Office BIM manager's KPI should include measurable efficiency and quality gains within the office.


An Office BIM manager does the usual things, for example;
  • Supervise technical teams and provide project support as necessary.
  • Assist Project Directors on technical delivery.
  • Development/Management of the BIM standards, protocols and templates.
  • Liaison and consulting across IT teams, systems administrators, clients and contractors. 
  • BIM training and compliance for junior members of the team.

but what a does a "good" Office BIM manager do?

A good BIM manager understands BIM.
  • Treats the model as a real world representation rather than a 2D representation.
  • Leverages BIM models as a communication tool both between those working in a model, and the recipients of the output of that model.
  • Recognises BIM models are created by a team of people working together, not individuals performing tasks.

A good BIM manager structures a team to leverage BIM.
  • Ensures no-one works in a silo.
  • Sets team roles based on responsibility, not tasks.
  • Forces people to take ownership; make them responsible for complete, not partial, work.
    (e.g. the person responsible for modelling walls is also responsible for wall tagging, wall details and wall schedules).

A good BIM manager is realistic about the capabilities of their workforce.
  • Doesn't expect people employed for their expertise and skill in building to also be experts at using particular software.
    (The reality is architects, engineers and construction professionals will never be fully proficient at the software they use).
  • Tailors work practices to the abilities of those who do the actual work.
    (Don't put someone in charge of facades if they struggle with simple tasks like wall creation).
  • Doesn't try and get designers to use particular software if it makes their primary task - designing, less efficient.
    (Getting designers to provide hand drawn sketches to those modelling is usually more efficient than getting designers to model properly).
  • Doesn't think "more training" is the only solution.

A good BIM manager recognises one size doesn't fit all.
  • Retains flexible workflows so unusual situations can be accommodated and innovative work practices are not stifled.
  • Doesn't enforce "universal standards".
    (an approach that is fundamentally flawed; it is not possible to predict every possible permutation of what needs to be done on every project).
  • Supports different work practices for individual projects based on complexity of the project and ability of staff working on it.

A good BIM manager involves themselves in real projects.
  • Maintains skills and intimate knowledge of how the office operates by actively engaging in projects. 
  • Is involved in setting up every project in the office.
  • Periodically audits all projects.
  • Steps in when required to assist, and uses it as an opportunity for training others.
  • But NEVER works full time on a single project.

A good BIM manager doesn't merely react to specific requests, they question those requests.
  • Assesses a request against the real world outcome it is trying to achieve.
  • Offers solutions that are workflow and work method based, not just technical solutions.
  • Gauges how long a request takes against the value of the result.
  • If appropriate suggests alternatives that achieve the same outcome.
  • Averts tasks that are done for no reason other than "that's the way it is always done".

A good BIM manager is proactive.
  • Uses the opportunity of introducing new software functionality to improve approaches to problem solving and service delivery.
  • Provides fearless advice, but accepts their view may not always be adopted.
  • Listens to others. (as they might just have better ideas).
  • Involves themselves in industry wide BIM issues.


The position of Office BIM manager is a relatively recent phenomena.  Despite what I said above the position does have similarities to the CAD manager role (and many CAD managers do move in to the role easily). Only now, with BIM, computer technology has much greater importance.

I.T. has become critical to the operation of AEC firms. Just as has happened with many other industries (a bank CEO famously once said he didn't run a bank, he ran an I.T. company).
As there is not a tradition of having a CIO or CTO equivalent in AEC firms (except for the very large) the role of Office BIM manager is well suited to filling this gap.

The Office BIM manager must be a part of all decision making processes. That is not to say they should be THE decision maker, just that their advice be sought and considered for all processes within the office, not just for the creating of drawings. They should be involved in practice management, project teams and job submissions. And be given responsibilities beyond just I.T., things like office QA.

However selecting the right person for the job is not enough.

Directors, designers and project leaders have to stop pretending they don't need to change the way they work, that it is only their underlings that need to learn new ways.
Those responsible for managing how the office, projects and output are done must also change the way they work for their office to benefit from BIM processes. Just checking drawings is no longer a viable QA approach.

After all even the most experienced and proficient Office BIM manager can only do so much if they have no influence over what half the office does.

BIM, and the benefits BIM can bring, don't happen by themselves. Like any process, if not properly managed it can be an impediment rather than an advantage. And a good Office BIM Manager is a vital part of getting BIM to work.

26 May 2016

How Usable are BIM Standards?

This month the UK Chartered Institute of Building (CIOB) published an article in their BIM+ blog.
it is a series of interviews with people involved in having to use BIM.

It is a timely and interesting article, and as regular readers of my blog will appreciate, close to my heart. A sample:
“People like me are invited on seminars and conferences and sent papers on BIM, but the information isn't easy to navigate,” he says. “It is made to sound more complicated than it really is and I'm having difficulty understanding what it is I have to do that I am not already doing."
Equally interesting are the comments on a LinkedIn group discussion about the article. Many thought those complaining about BIM standards simply didn't get BIM, and furthermore don't want to. Discussion on whether the contents of current BIM standards are good or bad seems to be not only avoided, but shut down.

Do I think BIM standards are unnecessarily complex? You bet. I also believe they are inadequate.

What makes a good Standard?

It is not enough to just have a standard, to provide something for managers to tick off, they must also serve a purpose for those that use them.

At a basic level standards should:
  • create consistency.
  • reduce industry effort.
But these aims won't be met if no-one, or insufficient people, follow the standards. For standards to work they must:
  • be useful for the creators of information.
  • be useful for the users of information.
To work in the real world standards must be evolutionary, not revolutionary. Each consumer of a standard must find it useful to them. Just as each evolutionary change has to be useful for it to survive in a population and so be passed on.

At the moment in BIM standards only "create consistency" is being considered. A tick box for project initiators and managers. When it comes to individual standards there is little consideration of reducing effort, or assessing cost benefit. When the additional effort required to comply with a standard is questioned it is dismissed as immaterial when considered against the overall savings of using BIM.

The bottom line is that although you can try and force people to adopted standards, they will only actually be used if they are useful to those who have to follow them.

BIM Standard Inadequacies

Talk about BIM standards always revolves around the need to have them. Of course we need them. It is not worth discussing the point. What is more relevant is how adequate are they? Are up to doing what is expected of them?

I'm not an expert on standards, nor do I claim to have an intimate knowledge of all BIM standards. But when I do investigate particular BIM standards I always find inadequacies. I don't do enough of an in-depth investigation to find all deficiencies, but when you find one you start to wonder if there are more. Here are some examples.

PAS 1192-2

The underling principles of PAS1192-2 are probably OK, but it is hard to tell. It is so overly prescriptive with poor explanation of objectives.
The section at the beginning titled "Fundamental Principles" is completely opaque to anyone without pre-knowledge of BIM. The problem is it explains principles in terms of BIM processes, rather than construction and operation processes.
Although the section titled "Scope" is easier to follow, PAS1192-2 gives the impression it is about a totally new discipline rather than a more efficient way of doing things that are already being done.

I have other criticisms, see them in my post Procuring BIM - PAS 1192-2 and acif PTI


I have to say I don't understand COBie.
Why require an email address for a product rather than a URL?
Why insist "N/A" be against all fields where there is no data without distinguishing whether the data is not applicable, not available, or not known yet?
I could go on. Read more in my post to COBie or not to COBie

But I've kind of given up on COBie. I'm not a facilities manager, I don't know how they work or think. If they find COBie inadequate for their purposes they should speak up.

But there is another issue with the use COBie that impacts on standard compliance and implementation.
COBie is from the US. It was developed by Bill East for the US Military. It subsequently became a US standard (NBIMS). The UK government (under advice) decided to base FM data delivery on the COBie standard. The UK developed their own COBie template files and made them publicly available.

The problem is the UK Templates don't exactly follow the US COBie standard. There are some spelling differences (critically important if computers are going to be used for processing), and examples in the template that contradict the US COBie. I've seen questions like this a few times in LinkedIn discussions:
"Would someone be able to confirm whether the COBie UK 2012 standard follows the NBIMS V3 and exclude certain Ifc types, such as Walls and Slabs? BS1192;4 (Fulfilling employers information exchange requirements using COBie) refer to FM Handover MVD and NBIMS V3 so should exclude these items, however the UK COBie example includes them, are they wrong?"
The original US COBie specifically excludes a building's fabric like walls and floors because they are not a "managed asset". The (sensible) basis of this is that is facility management don't have a remit to alter walls and floors so why include them in their data? (surface treatments to walls, which may come under their remit, are treated differently in COBie).

Now it could be a mistake. An overzealous, inexperienced minion added walls to the example template. But whenever this issue comes up it is vigorously defended on the basis that a facility manager "might" want to include walls. Under that logic COBie could include absolutely everything in the construction model. Which kind defeats the purpose of having a standard.
For more on how the UK is misunderstanding COBie see my post COBie is not what you think it is

NBS National BIM Object Standard

Where do I start. Every time I re-read this "award winning" standard I am in awe of how unhelpful it is.
I make a lot of BIM components. My last job was creating components for a pre-fabrication system. Yet there is nothing in the NBS National BIM Object Standard that I find useful, that would help me standardize the components I make.

It contains methods that can not be done in the most popular BIM software.
The standard insists it own parameter names be used, so instead of Revit's built in parameter 'Fire Rating' the name 'FireRating' must be used. The standard suggests mapping 'Fire Rating' to a new custom parameter  'FireRating' .
But you can't.
Firstly you can't create formulas for wall parameters in Revit (because they are not a loadable component), secondly it is not possible to use text parameters in formulas. Now these may be deficiencies of Revit, but the fact remains anyone using Revit can not follow the NBS standard. Why produce a standard most people can not comply with? Is it arrogance or ignorance?

The NBS don't even follow the standard when naming components in their own BIM Object Library  (see how in my post NBS BIM Object Standard - Where is the Impact Statement?).

Of course the reason they are not following their own standard's naming convention is because it hinders efficiency. But they refuse to change the standard because the naming convention comes from another standard - BS 8541:1 Clause 4.3.2.
To see how the NBS try and justify their approach have a look at this LinkedIn discussion.

Which is a fundamental problem with the current approach to standards. Rather than directly addressing the problem at hand (in this case one of naming) the "correct" approach is to always refer to another standard. It is one of the reasons standards are filled with references to other standards, making them incomprehensible to normal reading. Whilst there may be good reasons to refer to another standard rather then re-invent the wheel, it seems to be happening with no assessment of whether the referred standard is appropriate. Find a standard with a similar purpose (e.g. way of naming files) and then use it without question.


Before BIM there existed building classification systems. In the US Omniclass was created by combining a number of related classification systems (Uniform, MasterSpec etc). In the UK Uniclass was developed. These systems were mainly used by specification writers and estimators.
When BIM came along they seemed like a good way to classify objects in a BIM model. Autodesk added the ability to add classification numbers and descriptions to Revit objects. They also created data files of Omniclass values. But what they found is the existing classification system was not deep enough to be able to give every object that may be used in a Revit model a unique number. They had to add an extra 3 levels of numbers.

So what everyone in the industry assumed was a way of uniquely identifying every element in a building project actually couldn't.

This is an example of what I call the 'Delusion of Standards'. The delusion that a standard does what the authors and promoters think it can. And they maintain this delusion by not testing the standard in the real world, and shutting down any criticism. After all, it is less effort to convince people something is true than to produce evidence that it is true.

In the UK they realised the original Uniclass was inadequate for BIM use. Mainly because of the overall structure and lack of consistent structure between tables. To their credit they are revising it, creating a new Uniclass2 (now called Uniclass2015, I think, I haven't checked lately).
The emphasis is on 'revising' - it is not complete. This is another issue we in the industry have to cope with. Being told to comply with standards that are incomplete.

For more background on classification systems read my post Classification - not so Easy


IFC is at the core of BIM standards. Fundamentally it is a way of structuring digital data that describes buildings. Specifically data for computer programming. It was never intended for building professionals to use directly (if you think it is have a look at this example).

However the IFC structure (or 'schema' as it is called) can be used to structure data us mere mortals interact with. COBie is an attempt at this. The usual COBie deliverable is a spreadsheet file. The data is structured to follow the IFC schema and uses IFC names for things. It is touted as "human readable", but is only just. If directed and instructed adequately anyone can fill in the data, but it requires someone with deep knowledge of IFC to do the instructing.

So whereas IFC is fine for structuring computerized BIM processes it is not suitable for humans. Unless you are a computer programmer requests to "comply with IFC" are a nonsense. The most we can due is use software that claims to be IFC compliant.

Where most of us interact with IFC is with IFC files. That is BIM files in an IFC format (there is more than one). This is promoted as an "open format" that "any software can export and import".
Not because IFC can be exported and imported by all softwares successfully, but because that is the aim of IFC, or specifically buildingSMART, the not for profit and mostly volunteer organisation that promotes IFC.

It is a funny situation. A standard is created, and when particular softwares don't interact with that standard particularly well it is always the software's fault. On the one hand we have softwares actively being used by thousands (millions?) of people to do real world things, and on the other we have a standard artificially created to do theoretical things (there are no authoring softwares that natively use the IFC format). I don't understand why IFC is so sacrosanct.
For more on IFC refer to my post IFC, What is it good for?

But there is another issue with IFC that is not widely known. It is incomplete.

Last year I was upgrading my door library and I thought I would make them IFC friendly. That is, ensure they have enough parameters to support a compliant IFC export.
After some searching I found where buildingSMART keep their IFC specifications. First problem there are two versions, IFC 2x3 and IFC 4. The latter is the most current but not widely supported. Yet. Even though it has been out since March 2013. I decided to go with IFC 4.
I found some parameters (called "properties" in IFC) to do with doors. Mostly concerning geometry, which Revit already has native parameters for. But I couldn't find anything to do with door hardware (locks, latches, hinges etc.).

I though this can't be right. Nearly all buildings have doors, and all doors have hardware. So I asked the LinkedIn IFC group.
What surprised me was the attitude, the immediate assumption that IFC was faultless. Irrelevant other standards were suggested, and helpful suggestions that I develop my own IFC door hardware dataset. Someone offered the list of parameters NBS created for their BIM Object Library as a 'standard'.
But how can it be a standard if different groups create their "own IFC fields" as one commenter suggested?

So no, there are no IFC definitions for door hardware (or window hardware for that matter).
Which means it is not possible to use IFC to issue a standardized construction door schedule.

Do we bother with standards?

BIM standards do not make a pretty picture. Certainly not the utopia BIM Evangelists promote.

To be fair most are still being developed, and predominately by unpaid volunteers and inexperienced academics. The standards are young and untested.
The problem is they are being treated like some kind of dogma that can not be questioned. That the basis for assessment is wholly within the world of standard creation and other standards, not the real world of construction and facilities management where real things happen.

But standards are fundamentally a good idea. The computer industry heavily relies on standards, we wouldn't have all our e-devices without robust standards.

The solution is not in how we rid ourselves of these troublesome standards, but in how we make them useful.

For my two cents I see two fundamental problems.

Lack of Clear Objectives

High level standards like PAS 1192-2 seem to assume they must be as prescriptive as a standard for door hardware (for example, if such a thing existed). They don't, different processes can achieve the same results. For example you don't HAVE to use IPD contracts to get digital FM data.

High level standards should follow a similar format as the Building Code of Australia (and many other standards):
  • Objectives 
  • Criteria to meet objectives 
  • Requirements that are deemed to satisfy 
This structure means that if the objectives are demonstratively achievable any process can be used, but still provides prescriptive processes for the unimaginative.

As long as Objectives are be based on real world outcomes, not objectives wholly internal or in reference to other standards, like this from PAS1192-2 Fundamental Principles:
"application of the processes and procedures
outlined in the documents and standards indicated
in Table 1; "

Lack of basic information standardization

The second is that there is not enough work being done on low level standardization. Like IFC properties for door hardware.

Manufacturers data needs to be consistent, so different manufacturers provide the same data for the same products. It would also be helpful if construction data like door schedules were standardized across all projects.

It seems perverse that we have highly prescriptive standards on processes that manage non-standardized data. An elaborate mechanism to ensure the delivery of door data where this is no standard to say how that door data is to be structured.

Admittedly there is work being done in this area, but not nearly enough, and not fast enough. The UK government would have got more bang for their buck (pop for their pound) if they focused on funding and enforcing standardizing manufacturer data rather than untested theoretical BIM processes.

In fact it appears governments have to get involved looking at the failure of standardizing manufacturer data in the US. Bill East made this comment in a LinkedIn discussion:
"The conclusion reached during the SPie project [in the US] are that "If you build it, they will NOT come" (see movie Field of Dreams for quote). The bottom line is that the integration of product and equipment manufacturer data into the construction supply chain is a very, very hard problem. Publishing a list of product templates does not mean that anyone will actually use them. It has been tried over 4 times now in the US with national projects. Two have been attempted with the authoritative product data publisher, once by NIBS, and once by NIBS (under the SPie project). Despite significant development work and and participation by companies such as General Electric, there has been zero effective use by the supply chain."
We should, we need, to bother with standards. But we need to get them right.
In the meantime how do those of us on the ground, those having BIM standards thrust at us, deal with this unsatisfactory situation?

Don't worry about Standards

I'm not saying ignore BIM standards, just don't take them too seriously. Because BIM standards are not the most important thing you need to understand when utilizing BIM.

The most important thing you need to understand is how your BIM software works.
For designers like engineers, architects, sub-contractors your BIM authoring software, for contractors your BIM federating, estimating and scheduling softwares, for facility managers your BIM capable facility management software.

BIM may be a process but it is a process of managing software. If that software is used inefficiently or inaccurately it doesn't matter how good the management process is, the result will still be a disaster. The problem is not that people don't understand the BIM standards, it is that they don't know how to use BIM software properly.

There is no point a prospective taxi driver learning the streets of the city if they don't know how to drive a car. For managers, knowing the best places to distribute your taxi drivers around the city won't bring work in if none of them know how to drive.

Learn the Software, not the Standards

Unlike standards BIM software is made in a competitive market where the customer matters.
Unlike standards if their product is not useful they will do something about it (if only to the degree that it out-competes the competition).

Good quality BIM software (not CAD with a BIM add on) is designed to do the things you do. Unlike CAD which is for generic drawing BIM softwares are designed for specific disciplines. You will be surprised at how many of your processes are already built into the software. For example Revit has methods for doing area plans, sun studies, energy analysis, managing revisions, managing cross referencing, and many others. ArchiCAD has similar functionality.

But you have to use BIM software the way it is designed to be used. You can not simply force it to mimic the way you have always done things.
A lot of smart people have put a lot of thought into BIM software work processes, a lot of other people are using them, and those processes are likely to be BIM standards compliant.

Use the introduction of BIM software to review existing practice, develop new processes and retrain staff. When I teach Revit I do more than just show how to use the software. I introduce new ways of doing things. More efficient, more accurate ways. Like changing door parameters (to keep the door analogy going) instead of working through a door schedule spreadsheet, colour coding different door types, like fire doors; escape doors; disable access doors, so it is easy to check the right doors are in the right places.

BIM is, and should be, useful to everyone. Work out how to make BIM useful to you. How you can use your BIM software to make your processes more efficient, your output higher quality, to reduce your uncertainty and risks.

If you do that you don't need to comply with BIM standards, because you will be doing BIM.

When it comes to standards compare the work processes you have developed for your purposes against BIM standards, and see how they can be interpreted to match your needs. As I've shown above they are so full of holes it shouldn't be that hard. And even if your interpretation is not strictly legit it is unlikely there is anyone who can follow those standards well enough to realise.

So don't worry if BIM standards appear too complex, don't seem that useful. Forget about them. Concentrate on getting the most - for you, out of your BIM software. Once you do that everything else will fall in to place.

31 March 2016

COBie is not what you think it is

When BIM is talked about mention of COBie is never far away. What I don't understand is why COBie has reached such a privileged position. Sure it is (pretty much) mature, sure it has been used in the real world (although far from ubiquitous). But it is only a small part of BIM, a small part of the whole process of establishing, building and operating facilities.

Part of this seems to be coming from the UK and the furore to understand what they call "Level 2 BIM". But we see it here in Australia as well. Clients and owners who place requests for COBie deliverables that on closer inspection are not actually COBie at all.

I suppose COBie is tangible, you can download COBie spreadsheets and so tick the COBie box on your BIM checklist. But I feel COBie is a bit like Quantum Mechanics - most people have heard of it but very few actually understand it.

So it is very likely your understanding of COBie is wrong.

Not that COBie is as complicated as Quantum Mechanics. In fact COBie is probably far simpler than you think it is.

I'm no COBie expert, I'm an architect, not a facilities manager. Bill East is THE expert. He developed COBie while at the US Army Corps of Engineers (USACE), shepherded it in to the National Building Information Model (NBIMS-US) standard, and is still heavily involved in COBie, including the latest update.

Bill is co-manager of a COBie LinkedIn group, and is active in discussions. It is fascinating reading (for a BIM geek). Rather than provide my own commentary I've used quotes from discussions to clarify what I see as the most common misunderstandings of what COBie is (and is not).

COBie the Acronym

COBie, according to the buildingSMART alliance, is an acronym for 'Construction Operations information exchange'.

Bill East adds an important rider:
"...COBie requirements consistent with it's name the 'Construction (to) Operations Building information exchange' format."
COBie is about exchanging construction information to operations. That is, information that already exists for construction purposes is 'exchanged' for operation purposes.
"COBie is the list of scheduled assets found on drawings and in existing contractor O&M related deliverable. The sole focus of COBie is the capture of assets that need to be managed following construction." 
Therefore COBie is NOT about construction and does not include sufficient information for construction purposes.

And COBie is NOT a method to embed information required for Operations only within construction data or models.

COBie is only for Operations

The meaning of  'Operation' in the context of COBie is limited. Again from Bill East:
"COBie should include 'Managed' assets. Managed assets are those assets which;
- requires management
- requires (considerable) on-going maintenance
- has consumable parts requires regular periodic inspections"
"COBie only defines requirements for information about the spatial containment of managed assets -- these are manufactured products that have tags or serial numbers. These items appear in drawing schedules."

Yet there is a misunderstanding that COBie should contain anything that may even remotely be referenced. Bill East:
"A bit confused about the discussion of Walls since the COBie specification EXPLICTLY excludes walls, beams, columns, foundations, and all other structural members." 
"I hope that this effort will clarify the distinction between requirements for facility and asset management handover (which is COBie) and other needs such as carbon and life-cycle costing (which is not what COBie was designed to do). While other needs may be critically important, delivering them through COBie is likely not to work. Why? because COBie was not designed for that job."

Don't call it COBie

Not that Bill East is saying standards and methods to exchange information outside of COBie can not be done. Just don't call them COBie, call them something else:
"COBie is specifically for the purpose it was designed. If you are trying to make it do something else then it is no longer COBie." 
"Real Estate information is not required to solve the problem of eliminating boxes of paper in the boiler room -- i.e. "information about pump 5 in room 3." As a result, real estate information is not explicitly represented in COBie.
... if you change the purpose or content of COBie you will need to call it something other than COBie according to it's creative commons licencing terms."
And the warning is not just for those attempting to create a different COBie standard. If you use COBie on a project and add things beyond what COBie includes you also should not call it a COBie deliverable:
"If you use COBie in a way which violates the COBie specification you are no longer meeting the COBie requirement. You are doing something else that is not COBie."

COBie is not equal to IFC

There is a fair bit of confusion over the relationship between COBie and IFC. In simple terms COBie is not IFC, but follows IFC standards and protocols.

Ian Hamilton provided a good general description:
"IFC is a data format, well 2 actually: STEP (ISO 10303) and xml (ifcXML).
COBie is a list of things. It can be in a spreadsheet, in IFC or in other appropriate formats.
Bill East was more specific:
"In point of fact, COBie an IFC Model View Definition (http://docs.buildingsmartalliance.org/MVD_COBIE/)."

A Model View Definition (MVD) describes the things in a BIM model that are required for a particular purpose. In IFC those "things that are required" are labelled "information exchanges" (more on that below). From the buildingSMART website:
"An IFC View Definition, or Model View Definition, MVD, defines a subset of the IFC schema, that is needed to satisfy one or many Exchange Requirements of the AEC industry."

Stephen DeVito made it even clearer:
"COBie is a subset of IFC, an IFC Model View Definition, and comparing IFC to COBie is like comparing the whole assortment of fruit in a basket with only the apples. This is a most basic fundamental misunderstanding which occurs constantly in the industry ..."

One of the causes of confusion is that the IFC used by design and construction software is also an MVD. As Bill East explains:
"BIM authoring tools (up to this time) produce IFC files based on the Coordination Model View Definition. The purpose of that MVD is to express the geometry of all physical objects in the project for purpose of collision detection.
The MVD inside COBie has a much more modest goal, simply to deliver information about managed and maintained facility assets. As such, COBie data in any presentation format (IFC, ifcXML, SpreadsheetML, COBieLite) will be smaller than that of the Coordination View."

The idea is that the design and construction team's IFC Coordination MVD can have the COBie MVD extracted from it.
This is fine in theory, the reality is not quite so simple. Different softwares export varying qualities of IFC, and not all items included in the IFC Coordination MVD are always modelled (because they are not needed for the particular project).
So whilst in theory you should be able to extract COBie from an IFC export, currently it rarely works without a lot of unnecessary extra effort, if at all. One day this might be practical, but generally it is easier and less work to extract straight to COBie from design softwares.

Another misunderstanding is that COBie, although not currently containing a 'full' definition of IFC, will be further developed so it includes a greater range of definitions.  As Bill East says:
"COBie-UK is clearly called out as a stepping-stone to 'full' building information modeling in IFC. In my view UK should just stop calling what they are doing COBie and simply get on with requiring IFC with all disciplines, trades, geometries, entities, properties, etc..."

To sum up, COBie is NOT a general delivery method for everything in IFC.

COBie doesn't need to do everything

The 'ie' in COBie stands for 'information exchange'. From the buildingSMART alliance web site:
"The requirements [of information exchange projects] are defined in an 'Information Delivery Manual.' The IDM clearly defines the problem to be solved and makes clear who is involved, what information is needed, and when that information is needed. These requirements are translated into a 'Model View Definition' that provides the technical description of which parts of the Industry Foundation Class Model (IFC) found in ISO 16739 are needed to solve the problem."
COBie is only one of a number of information exchange projects. Other projects listed by the buildingSMART alliance to date are:
  • BIMSie - BIM Service interface exchange
  • BAMie - Building Automation Modeling information exchange
  • BPie - Building Programming information exchange
  • Sparkie - Electrical System information exchange
  • HVAC information exchange (HVACie)
  • LCie - Life Cycle information exchange: BIM for PLM
  • QTie - Quantity Takeoff information exchange
  • SPie - Specifiers' Properties information exchange
  • WALLie - Wall information exchange
  • WSie - Water System information exchange

All of these follow the same structure as COBie - they are subsets of IFC definitions, with their own MVD.

So you can see the intent is that information specific to a purpose has its own 'ie'. There is no need to expand an existing 'ie', instead a more appropriate 'ie' is used, or another project is started.

For example COBie doesn't define the format manufacturers provide their information in, SPie does that. However when manufacturer's information is required by COBie it has to follow the SPie format.

Rather than have one big standard the idea is to break it down into more specific standards that can reference each other. From Bill East:
"The sole focus of COBie is the capture of assets that need to be managed following construction. The system "ie's" include COBie for that discipline (i.e. the Components), but also include the assemblies of those components and the connections between the components. These system ie's provide the full geometry at least as far as fabrication and can be included in construction contracts as a better statement of as-builts than any attempt at having someone do a walk through of the project after the fact..."

So in the example above where Bill East is talking about Real Estate information (in response to some-one asking why that information isn't included in COBie), the way to do it is to create a Real Estate information exchange:- REALie, or if the data is actually about land titles:- TITie.
This information could be delivered in spreadsheet form, even within the same file (in its own worksheet) as a COBie deliverable, just don't call it COBie.

Anyone can start or get involved in an 'ie' project. From the buildingSMART alliance website:
"To participate in an existing project simply contact the point of contact identified on that project page." 
"If the project you need isn't in the list above, you can start your own project by joining the buildingSMART alliance."

There is more than just COBie

There has been enormous focus on COBie but it is not necessarily the only data exchange that will improve productivity.

Specifiers' Properties information exchange (SPie) is an interesting case. Basically it is meant to standardize the way products are specified, leading to standardizing how manufacturers define their products. So they all use the same names for the same things, and include the same information as each other. You would think this would be an easy task. Apparently not. Bill East admitted:
"The conclusion reached during the SPie project [in the US] are that "If you build it, they will NOT come" (see movie Field of Dreams for quote). The bottom line is that the integration of product and equipment manufacturer data into the construction supply chain is a very, very hard problem. Publishing a list of product templates does not mean that anyone will actually use them. It has been tried over 4 times now in the US with national projects. Two have been attempted with the authoritative product data publisher, once by NIBS, and once by NIBS (under the SPie project). Despite significant development work and and participation by companies such as General Electric, there has been zero effective use by the supply chain."

However the UK may have better results:

Carl Collins:
"Is the failure of SPie in the US a function of who created it? Have the suppliers themselves been an integral part of the process? I'm guessing not, as they are often unaware of SPie when I have spoken to them." 
This is the difference that the CIBSE started Product Data Templates (http://bimtalk.co.uk/bim_glossary:pdt) has; we are defining them with the Manufacturers and Suppliers and asking for sign-off from their Trade Associations, so there is buy-in from the outset.
Our starting point for each template is the SPie template, but we have found that they are not detailed enough to adequately describe the product and have too many project specific parameters that don't really belong on the Product side, but should be detailed on the Project side. This allows a Manufacturer to complete a template once for each product line and may be used for any project."

This is really critical. Producing COBie mainly involves manually transferring manufacturer data into COBie format, whether done directly into COBie spreadsheets or into a BIM model. Enormous productivity gains will happen when this data can be pulled in directly.

Rather than just mandating COBie specifications and contractor's supplier contracts should be required to demand suppliers provide standard format product data. I'm sure transferring a little money from manufacturers' marketing budgets would more than cover the cost.


The UK government has a BIM mandate. Does this mean COBie is a required deliverable for all projects?

Rob Jackson:
"COBie will be mandatory for all centrally procured UK Government projects from January 1st 2016 [now 4th April 2016]. The private sector can do what they want but even most of them will align with recognised standards."

And COBie is not necessarily a requirement if UK standards are followed. COBie is merely one method that may satisfy requirements.
Charlotte Brogan (Gray):
"COBie is one way of transferring data from a 3D environment to an facilities management software, however depending on clients in the UK will depend on how the data is handed over. This is why it is not mandatory in the UK, the standards state that a single source of asset information is to be produced and handed over to the client upon project completion. Therefore with the use of document management systems and links this can be achieved without the use of COBie. One day when clients are up to speak COBie may become more prominent in the UK AEC Industry."

Is COBie-UK different from COBie?

The fact the UK have developed their own COBie templates tends to confuse many. Some believe there is a UK COBie and a US COBie, which makes a farce of the idea of COBie as a standard.
COBie allows for regional customization so the fact COBie-UK templates exist is not evidence there are two standards. For example COBie-UK mandates UniClass for classification. But this does not necessarily mean it is non-compliant as the choice of classification system is not defined in COBie. Bill East:
"According to the standard, classification is required but arbitrary. The choice depends on regional, national, local, owner convention. Ultimately, the best choice is the one that serves the recipient of COBie data."

Although COBie-UK does have some language differences which can cause problems for computerized processes. Rob Jackson:
"The U.K. COBie-UK-2012 template have Moveable and Fixed in the standard Picklist for AssetType. The US standard as Bill points out uses Movable. This is one of the minor differences either side of the pond."
But more of concern is Bill East's view that COBie-UK is pushing COBie beyond its intended purpose.
"COBie-UK efforts have resulted in unrealistic expectation that (for example): the Coordinate sheet should be required (it is, in fact, junk and will likely be removed from the next iteration of the COBie standard); the insistence that COBie can successfully model steel structures; and the expectation that every possible permutation of room finishes -should- be included in COBie." 

This may just be due to overzealousness, or perhaps a lack of understanding the concept of MVDs and information exchanges as envisioned by buildingSMART.

But I agree with Bill East. If you don't think COBie is adequate for your purposes then use, or develop, something else. Don't mess with an existing standard. It just confuses everyone.

For example if the UK want a single deliverable for all information exchange data create a container for COBie and any other 'ie' that a client/employer/owner may want. And call it something like "UKie", or "HMGie", after all the mandate is only for government projects.

Use COBie properly

From Bill East:
"Owners need to answer three critical questions if they want COBie data they can use. Why? because COBie is only the format to deliver handover data. COBie can't possible predict the specifics of an individual Owner or project. Here are the questions: 
  1. What assets do we manage? The Owner should look at what they actually maintain over time. The default position of getting "everything" distracts the team from the Owner's real needs. 
  2. What information do we need? If Owners need the fan belt size for fans, say so in writing. Without such specifics Owners can expect to get whatever is given, and like it. 
  3. How will it be organized? Campus/Installation owners with a consistent Classification method for COBie.Space and COBie.Type will be able to mine data."

COBie is only meant to contain information called for in design and construction contracts. As Bill East says:
"This topic keeps cropping up, so I thought I would remind everyone of the following rules about COBie and product data. First, COBie product and equipment properties at design must match what appears on design schedules. Second, COBie equipment properties at construction must match what is in the existing non-COBie contract specs (typically equipment nameplate data). Because of this, there is -no- additional cost of "doing COBie" since the information being required is no different from what is currently in design and construction contracts."
Therefore if your COBie deliverable contains data that is NOT required for design or construction, it is beyond the scope of COBie. And that extra data is extra work for those creating the deliverable. Don't expect it for free.

(Although I disagree with Bill about there being no additional cost for "doing COBie". If that were the case we should be able to deliver our documents in Spanish, or Chinese, - same information, just different format.)

But my favourite comment from Bill East says it all:
"what COBie repeatedly has been about is the 'art-of-the-possible' not the 'art-of-the-aspirational'."