A 2012 Taxonomy of Drawings: Paper to Vector

A 2012 Taxonomy of Drawings: Paper to Vector

A Taxonomy of drawings and why I hate paper, whether I’m looking at it on a table or on a screen

Despite all the hype about BIM, I seldom see a contractors’ trailer with computer screens instead of paper drawings. In fact, despite working on recent projects by architects such as Foster + Partners, KPF, Gensler etc. the only real change in the scenery on site is the equivalent of sheet drawings sent by electronic transfer of files. In other words, the drawings are still 99% sheets; it’s just that they arrive as a .pdf or a .tiff. instead of as a sheet of paper.

Believe it or not, as recently as this year we still experienced couriers delivering paper “logs” and cases of binders, so we had to beg the contractor to send us electronic drawings, since we no longer have the desk space to spread out 400 paper sheets. I am much happier when the drawing arrives as a vector-based, editable file, because both I and my staff have worked on large and / or dual screen setups for years and so it is much easier for us to store, share, view and mark up electronic drawings. It is also easier to save, sort and retransmit them.

Categories of drawings, incremental categories of value

So let’s say it is 2012 and we work in the office of a contractor or trade. We don’t receive paper drawings anymore, but we do receive the following four types:

1)      A scan of a paper drawing

You have to see it to believe it, but 25% or more of drawings we received for new projects, to the middle of 2011, were scans of paper drawings. This means that while the architect produced an electronic drawing, the folks upstream printed it, fed the print in to a scanner and then sent us the scan as an image file. This is instead of just sending us the original electronic drawing file. Apart from the waste of trees, the big problem with scans is that, unlike a .pdf or a .dwg, all the vector information is absent, so the only way to use the drawing for design development (other than as a sketch pad) is to re-build all the data, line by line. An additional problem is that the files are seldom large enough to capture small details, so when you blow them up you get pixels where there should be mullions. Sadly this practice is most common with government and large contractors.

The dumbest part is that a lot of useful information is lost and it actually took work to lose it. This is what I refer to as a negatively productive process; like paying a contractor $20,000 to put Angelstone on a Victorian house, thereby depreciating its value.

The information value of scans is of course very low.

2)      A vector-based 2D drawing

As mentioned before, many architects process the final drawing output from a 3D program like REVIT in AutoCAD, or they use REVIT or AutoCAD directly, generating drawings that are saved and transmitted as a vector-based file. The electronic file is sent as is, in the form of a .dwg, or it is converted to a .pdf, which looks and handles better. We still see a lot of light yellow lines on white backgrounds, (or black lines on black backgrounds) but generally vector files are legible. The best part is that the data includes conveniences like layers, dimensions and fixed relationships between lines which can be extracted and used. Also the drawing will be legible regardless of magnification.

The bad part is that most .dwg or .pdf drawings are lacking fully buildable details and they are not extracted unmodified from a common model but are “fixed” before saving, so they are “stand-alones” or orphans, with no useful relationships to surrounding objects.

The value of these drawings is only slightly higher than a scan.

3)      A vector-based 3D drawing or 3D file derived from a “visualization” platform, or an inexpertly used parametric modeling platform

REVIT, Bentley, Rhino etc. produce 3D models  which can be sent whole, or which can be the source for views which are arranged on sheets and sent as vector-based drawings. The problem with 99% of the output I have reviewed is that a lack of constructability input, combined with the use of tools for creating features such as stairs, windows, etc. produces drawings with impractical or missing details, so the drawings are not directly useful for construction.

The value of these drawings is only slightly higher than #2, above, though the ability to rotate a model can make some features easier to understand.

4)      A vector-based 3D drawing comprised of cut-file-level data derived from a parametric manufacturing platform or a 3D scan

I routinely receive drawings like this from another subcontractor, such as a structural steel fabricator, who is working in TEKLA, SolidWorks, etc. or from a scan service, which has scanned a site for us. Models like these can be used as a framework in to which our work is drawn, and the level of trust is so high and verifiable, that we can proceed to production without measuring a site.

The value of these drawings is very high, a quantum leap above category #3, but the cost of producing them is also high. That is unless you take in to account that the detailers who generate these cut-file-level drawings are doing work that was formerly done on a shop floor by hand, work such as layout and cutting. Seen in this light, the cost is actually low.

5)      Part drawings and cut files taken from a coordinated 3D model

These drawing are actually the goal of the entire drawing process. The value level cannot be increased, except by actually feeding the data in to a machine and cutting the part, or by clamping the parts together on a bench. The production cost of these files is low as the modeling software makes cut file output simple.

6)      Erection plans based on a 3D model

These drawing are the goal of the coordination process. The value level cannot be increased, except by actually putting a strap around a piece of steel and hoisting it in to place. The production cost of these files is not as low as cut file production because quite a bit of expert input goes in to a smooth erection plan which is well coordinated with surrounding trades and adjusted as required to deal with weather.

Final thought

A lot of time and money is wasted on construction jobsites due to avoidable failures of coordination and communication. Cut-file-BIM, or Real BIM, if I can call it that, is already happening, but so far it is a bottom-up process driven by innovative fabricators. The time has come for architects and builders to make the required organizational commitment and process changes, so as to avoid becoming irrelevant in the eyes of their clients.

Julian Bowron October 2011


Comments

August 24, 2012
Richard Blair says:

Julian, I hear you loud and clear. In 1991 I was the Project Manager responsible for 300,000 sq ft of pick modules in a 1.1 million sq ft distribution centre. 3D was not affordable in those days, so we were using 2D CAD. While working on the project, I had the good fortune of being introduced to a new hardware technology called a “Notebook” – Eureka!!! Within a week I never again took another drawing or piece of paper to the job site. My drawings, project schedule, variance reports for budget, etc., were all contained electronically on the minimal 5MB credit-card-technology hard drive that was available on the first Notebook ever built. It went with me floor by floor.

At the job site, we were able to review drawings and even generate drawings of missing details for the construction crew. As-builts were completed within 2 days of the completion of the project as we were able to update the drawings during the construction process. This was all possible because I had digital data on-site. I had 6 jobs offers within a 3 month period from 3 different companies that were also working on the project.

This was 21 years ago. Back then, I thought that everyone would be doing this within 3 to 5 years. I was wrong.

Many manufacturing companies jumped on this band wagon years ago and some have gone paperless. The AEC industry is years behind. The project I described was much smaller in scale and detailed content than a building. However, the scale and speed of hardware is many multiples larger today. The software is also much more intelligent. There are many creative ways in which what you describe can be accomplished. You know because you have done it!