Sunday, April 19, 2020


I keep promising myself to write short, punchy posts so that I can publish twice a week.  Not happening.  Seems like I have been working on this one forever.  It takes the form of a retrospective.  One year in.  Project Notre Dame.  An accident of history, voluntary collaboration, "BIM pencil initiative", journey of discovery.

Just over a year ago I was recovering from an operation. The problem had cropped up while visiting the UK and I had the operation there, while staying with my son. During the course of one weekend, we had visited Winchester Cathedral, a concession to my obsessive interest in Old buildings.

Recuperating back in Dubai I had the idea of making a rapid BIM model of the cathedral. I allocated a single day to the task and my goal was to probe my understanding of "how the building works". Sketching and modelling are ideal complements to live visits, backed up by research, image gathering and periods of reflection.

The process of constructing a diagram with your own hands, forces you into a different mode of attention. You are compelled to ask questions and resolve anomalies.  Put simply, it's "learning by doing"

Revit is my tool of choice, my BIM pencil, the familiar friend which has become an extension of the hand-eye-brain loop, familiar to all artists.  When I was working as a bricklayer, in my twenties, the trowel became part of my body.  Most people have developed a similar relationship with a pencil.  You don't have to think about how to hold it. You just immerse yourself in drawing or writing.  That's why the "BIM pencil" metaphor works so well. 

From the time I got my hands-on Revit I was motivated to use it as a thinking tool to search for the BIM equivalent of a freehand sketch. Thinking on the fly, engaging the intuitive, subconscious areas of my brain.

Then a strange coincidence occurred. Notre Dame caught fire. How odd that a Gothic cathedral should hit the news so soon after I had grappled with the form of another, of similar age and size. I couldn’t resist the temptation. This time I gave myself a full weekend, and I was able to reuse some of the parametric families from Winchester, simplified pointed arches and windows.  Along the way I shared two images to Linked-In, the first of these receiving 11,000 views, which was way more than my previous record.

It’s interesting to recall what I wrote when I shared the first image:  

“Half a day spent massing out Notre Dame in Revit, then linking in the Winchester model from last month. Will try to push it a bit further tomorrow…  It’s just an example of learning by doing.  One model prompted by a personal visit, the other by international news.”

And the second one:

“Still very crude and simplified, but starting to get an understanding of the interior space. #notredame is such a symmetrical and regular design for a Gothic building.  Fascinating.  If people are interested in building a more accurate and detailed model by way of open collaboration, maybe this first roughing out could be used as a basis …”

The fire broke out on a Monday night.  I worked on the model over a three-day weekend (Friday-Sunday) and converted this work into a blog post by the following Wednesday.  Two long-standing friends (Alfredo Medina & Paul Aubin) contacted me around the same time, offering to join the collaboration.  

By the beginning of May I had set up a Slack group and we were a team.  Marcel Rijsmus started working on the roof trusses and Daniel Hurtubise joined in the fun from Paris.  Since then at least 50 people have contacted me showing an interest in contributing.  Some made a start but found it difficult to allocate enough time, others never really got going.  That’s fine.  My enthusiasm often runs far ahead of my ability to deliver.  

Here is a view of the current model with some of the major modelling contributions (far from exhaustive, and thanks for everyone who has chipped in so far)

In this project my role has tended to be that of pioneer: forging ahead to map out the territory that we are hoping to inhabit.  Others have then picked up specific tasks and developed them in more detail.  Daniel has done some modelling but contributed mostly in terms of model management and guidance.  He has been invaluable to me as a “shoulder to lean on”.

Paul and Alfredo made major contributions over the first few months but have since taken more of a back-seat role, under pressure from their day jobs.  We began with a strong North American bias to the team, but more recently the centre of gravity has shifted to France, (appropriately enough.)  Of late Nader has been doing excellent work seeking out information and making connections to other groups.

Over the months I have come to realise that Notre Dame has been modelled many, many times using different tools and materials, with different purposes in mind.  There have been film sets, model kits, a wooden maquette that stood inside the cathedral before the fire, several point clouds (including the well known 2012 version by Andrew Tallon) … You can go on to platforms like Sketchfab and orbit around models of the whole cathedral... and of various parts.  Some of these are based on photogrammetry, others clearly “hand-built” using mesh modelling software packages.

Our Lady is a subtle and complex creature.  We continue to discover surprising subtleties as we proceed further along our adventure.  It’s no surprise that most models miss the way that the Triforium galleries turn the corner at the transepts.  That was one of my earliest "AHA moments" and those four locations have been a constant source of new insights, each of them subtly different in its complexity.

Almost everyone stumbles with the setting out of the bays around the apse.  Thirteen bays on the outside, centre of arc offset from the grid, zig-zag vaults to handle the transition from two bays one side to three on the other.  The primary buttresses are straight, radiating out from the 6 columns that define the 5 inner bays that curve around the east end of the choir.  Intermediate buttresses change angle half-way in.  This is one of several discoveries that haven’t yet made their way into the model.  Much still to do.

But let’s go back to the beginning. Sitting here, confined to my fifth-floor apartment in a suburb of Dubai, April 2019 seems a lifetime ago.  Before embarking on Project Notre Dame, I had spent more than three years recreating the Bank of England, as it was 200 years ago when Sir John Soane retired.  This lost masterpiece had been the subject of a competition which so gripped my imagination that I couldn’t stop working on the model when the allotted period expired.

My first blog post for Project Soane documented the approach I intended to take.  For me, drawing and modelling are extensions of my brain, activities that nurture understanding. The building was still a mystery to me, a black box.  I had plans, sketches, photographs … but how to make sense of them? What was the logic behind the layout of the building, the internal courtyards, the subtle angular shifts between different zones? I knew that Soane was the third architect to work on the building complex, but what was the sequence?  How did the plan evolve?  Who did what?

My strategy is always to jump in feet first, engage with the materials, build something, stand back and review, embark on a second wave of building, reflect some more, and so on.  Certain phrases from that first blog post stand out.  “Broad Brush” … “giving it my best guess”

Project Soane was a seminal experience for me.  I had built Revit models of many historical buildings before, with the same goal of expanding my understanding of how they worked, but this was a much deeper dive.  Also, for the first time, there was an element of collaboration.  This was the initial premise of the competition, and continued throughout (although not to the same extent as Project Notre Dame)

In January 2019 I wrote a post called “Ellipse in Residence” which records work I was doing to flesh out the interiors of the Residence Court. Sadly, some of this work has been lost in the bowels of BIM360.  It would probably take 2 or 3 weeks to restore the lost work, but for the past year, Project Notre Dame has taken priority.  One day …

During January and February, I was on sick leave and took a break from Project Soane. It seemed like a good opportunity to explore classical ornament, something which is quite challenging to represent within a Revit model.  I did some hand sketching of acanthus leaves and used these develop some new strategies for generating Corinthian capitals.  All this is recorded in blog posts.  

Some quotes:

“What will we do when robots and AI take over our day jobs? I would devote my energies to exploring the history of “the way we build” using my BIM pencil”

“Can we narrow the gap between fast, intuitive sketching and the cumbersome world of BIM?”

Then came the fire. 

One year on, and why are we still doing this?

For me it’s partly habit.  But beyond that … I continue to learn.  Just a few days ago, I discovered that the spacing of the timber roof trusses is much closer than the spacing of the masonry columns. As I write this post, I am watching a two part documentary on YouTube.  The channel belongs to Deutsche Welle (German Wave) an independent public broadcaster. But the production is credited to Program33 a French company which describes the film as “historical docu-fiction.”

How much of the story is “poetic licence?”  That’s an interesting question to hold in the back of my mind as we continue our work. I suspect that a generous dash of “human interest” has been added to the story line but it remains an impressive achievement, well worth watching.

I love the short clips of real artisans working with the techniques of the time.

By the way, as we discovered a couple of months ago, the North Tower should be about a metre wider than the South Tower … resulting in 8 statues in that portion, not 7.  I find it oddly reassuring that even lavishly funded enterprises make mistakes when modeling a building of this size and complexity.

The other contributors have chosen areas of interest, and devised their own approaches to the project.  Levels of activity go through natural cycles.   

Another quote from my Linked-In bulletins about Project Notre Dame.  

“I try to apply my BIM pencil as a Renaissance sculptor might apply her chisels … as I look for the natural mode of expression, hiding inside my chosen tools.”

The image below was created by sequential use of three of my favourite chisels (Revit, Enscape3d, Photoshop)  It tells a story about changes that were made, three or four generations into the building process.  As the gothic style progressed in time, builders become bolder, windows larger, light more pervasive.  There is a note of timidity here.  The end bays of the nave were left in their original state, with smaller windows above a portion of the original roof.  

There is another layer to this story.  Bay no 8 is next to the transepts, where the triforium galleries wrap around.  On the inside of each corner there are 3 façade bays, with round windows that give borrowed light to the roof voids.  Six dark roundels then on the two corners that form the West flank of the transept.  

But on the East, those six roundels allow light into the cathedral, directly from the outside.  It seems that this arrangement can be credited to Viollet LeDuc, who renovated Notre Dame in the nineteenth century.  But I  still have doubts about the state of these bays before he modified them.  Were there roundels on the East and blank panels on the West?  Were there full-height windows on the East and blank panels on the West?  The documentary suggests full height windows on both sides.  But why did he change these?  

These conundrums have a habit of resolving themselves over time.  A luta continua.  The struggle continues.