ArchiCAD Layer Theory, Part 4: Different Types of Models


This post is part 4 in my series on ArchiCAD Layers. If you’re just joining us, here are the links to the first three posts (make sure to read the comments as well!):

  1. Part 1: LEGO and Layers
  2. Part 2: Do you need that Layer?
  3. Part 3: Layer Combinations

Bonus Posts that I don’t think are coincidental:

  1. Layers and Layer Combinations (from ArchiCAD SADC blog)
  2. Filter Your Layers (from the Cadimage blog)
  3. Deleting Attributes and the One Layer Model (from Shoegnome…more on that at the end)

We all have our preferences as to how many Layers and Layer Combinations we need to work efficiently. Our predilections come with associated views on naming conventions (seriously, if you haven’t checked in recently on the comments here, you need to) and also depend on what the project type is and what your collaboration environment looks like. But there is another major factor that will determine the details. To me, the primary defining characteristic of what determines the quantity and quality of your Layer and Layer Combination system is the goal of your model.

Who Are You Modeling With?

If you remember my blog posts from 2013 about the Primary Benefits of BIM, you might recall I have some pretty strong and clear views on the potentials of BIM. But just because the long term goal of BIM might be to connect our virtual model to the physical built world, in the short term each of our current projects probably has some very specific needs. Depending on if you are using ArchiCAD primarily to enhance production, coordination, design, or integration, you should be looking at your work in a specific manner. While we often tackle a mix of these aspects of BIM on each project, there are times when the model is all about one thing, or about one thing above all else. Let’s look at a few examples.

The Most Basic Layering SystemBIM Collaboration

If you are interested in integrating IFC data deeper into your model, you should consider turning Layers into an Attribute that supports the whole world of IFC data. I’m not going to try to explain it further here. Rob Jackson has written about this (check out point 7). And more recently, Nathan Hildebrandt has also written a post about this. Not everyone is doing IFC exchange right now (though this is quickly changing), so this Layering system isn’t for everyone. BUT…if collaboration and coordination of models between authoring tools is the fundamental reason you are modeling, then a layering system that ties into IFC data is worth checking out. Just as importantly, if OpenBIM collaboration is a direction you want your work to head, then aligning your Layers in a way that supports this is of high value.

CAD Collaboration

It pains me to even write that. But it’s true. There is definitely a segment of users who deal with this everyday: consultants or clients that require CAD. For those ArchiCAD users it might be very legitimate to build a layering system that looks like it came out of AutoCAD release 12. If having layers like A-wall-new and ANNO-dims-Int means data flow between ArchiCAD and ancientCAD is improved, then go for it. But if that’s not your top requirement, don’t show me a file with that sort of terminology. I will yell at you and tell you you are wrong. Because you will be wrong. Unless clarity between your advanced work and those using ancient methods is the hinge upon which your world revolves.

ArchiCAD Collaboration

There are a few ways to look at collaboration between ArchiCAD users— though they can be simplified to serial teams and parallel teams.

  1. Multiple users accessing the model independently (serial teams). This could be a file being handed off between team members within a firm or a file being sent from a firm to an outside designer/freelancer and back. In this case, you want your Layers and Layer Combinations to be something that can be easily adhered to as different people of different skill levels and knowledge control the file at different times. That probably means clear names and minimizing the complexity of the layering system.
  2. Multiple members of the architecture team working in the same file via Teamwork 2 and the BIM Server or BIMcloud (parallel teams). If having lots of ArchiCAD users in the same file at the same time will be the overarching theme of your model, then you should think about reserving elements by Layers and how the model might be parsed via Layers to better accommodate so many people. The details are beyond the scope of this article, but think about how your layering system relates to the number of people in the file. It’s worth noting that some other strategy discussed here (IFC Layering, CAD Layering, etc.) might support this—or not.
  3. Multiple disciplines in the same file via Teamwork 2, etc. (parallel teams). Once the file has more than one discipline (architects, structural engineers, MEP engineers, landscape architects, etc.) actively working in it, you’ll find your layering system shift. Do you want architectural and structural  Layers to be clearly defined (yes)? Do you want architectural and structural Layers to be equal in quantity (probably not)? Do you want to micromanage the other disciplines (maybe)? How will you segregate out the Layers? Will you handle that with extensions? Do architectural Layers start with A or Arch or 0 or ??? Will it be A Anno Text or A.Text or A – Text or Text.Arch? How will you arrange Layers and handle Layer Combinations so that each others’ Layers don’t pollute and distract? Needless to say, as the make up of the collaborators change, so too will the Layers and Layer Combinations. By the way, if you don’t believe non-architects are using ArchiCAD to do their work, think again. Here’s one example. And here’s another.

Layers for Concrete PoursWhat Are You Modeling For?

Beyond collaboration and production of traditional documentation, there are a number of other reasons you might be modeling. Your layering system should support those reasons. And in fact unique layering systems can help you see all the interesting ways one can use ArchiCAD. I want to spend the remainder of this post just listing some interesting goals of an ArchiCAD model and how, if these are the main focus, Layers might support those goals. At a later date, I hope to revisit some of these modeling types to dig deeper, as each has further implications that go well beyond just using different Layers.

  1. Energy Modeling. Imagine if your primary purpose is energy modeling. If that’s the case, then your model should be done in such a way that it supports how Energy Evaluation, EcoDesigner Star, and/or 3rd party energy models see the model. Is there some way to structure the Layers so that it makes exporting to gbXML easier? Or is there a way that your Layers can be grouped by “affects energy model” and “doesn’t affect energy model”. What would that do to your modeling and design methodologies?
  2. Rendering Modeling. Of course all our models can be easily rendered, especially with the awesomeness of the built-in CineRender engine. But if the primary purpose of your model is for rendering, what does that mean? What should you model? What shouldn’t you model? How can your Layers support all that? What if you are exporting your model to a 3rd party rendering software? How could you use Layers to enhance that exchange?
  3. Cost Estimating Model. Imagine you are doing an early design and just need basic cost at the end. Could your layering system support segregating the model by some criteria that supports who assigns dollars to what? Yes, there are other ways to handle this. But if you are modeling fast, what if your model and Layers made costing easier within ArchiCAD, or when the data is exported to some other program? Imagine if you assigned elements to Layers according to dollars per square or linear foot, per unit, etc. So you might have a $100 Layer, a $10/ square foot Layer, a  $1/linear foot Layer, etc. Maybe that’s crazy. But maybe it’s not.
  4. Schematic Model. If you are doing a schematic model, one that will get thrown out later, what does that mean for Layers? How many do you really need? Maybe just one. Maybe just a few. If the model is only used for screen captures and quick explorations, how might Layers be simplified to support that?
  5. Iterative Model. Imagine you have a base model. Then you want to test design options to show a client. Imagine if the base model is one Layer and each design option is another single Layer. Then switching between Options is as simple as switching Layers/Layer Combinations. You want to show ten schemes plus the existing conditions? That’s eleven Layers to manage and eleven Layer Combinations. Or do you want to show Scheme A and Scheme G together? Turn on those two Layers. Simple. Done.
  6. Zone Only Model. We don’t always need all our tools. What if you are building an early model with just Zones. How could a simplified layering system add complexity and detail to that model?
  7. Time Based Model. Perhaps you’ve download the Construction Simulation Add-On. Maybe you haven’t. Either way, you want to explore how the model can be used to talk about when different elements need to be built. Modeled and Layered correctly, this could be done quite easily. Perhaps your Layers are Step 1, Step 2, Step 3, etc; or week 1, week 2, week 3; or perhaps January, February, March…
  8. Detailing Model. Exploded axons, intricate 3d details, etc. Could a series of details just be the same view with different Layers turned on or off? Absolutely. If your layering system is designed to handle that. In this case maybe Layers are by CSI number or just something like Flashing, Air Barrier, Brick…
  9. Competition Model. What if you are doing a training exercise or a competition or some other atypical collaborative event. How might you structure the Layers differently?
  10. Atypical Schedule Model. I am working with a client who needs to schedule elements by elevation (it’s a framing model and each elevation’s lumber needs to be listed separately). My solution was to create a Layer for each elevation. In fact, I have a layer for each View, as this allows me to schedule each view separately via Layers. It’s not your average layering system or model, but by thinking about Layers differently I was able to do something that one can’t otherwise do (schedule elements based on their appearance in a particular view).

I’m sure I could go on and on. Perhaps some of these potential models are foolish. ArchiCAD can probably handle many of these models in other ways using different tools and Attributes. But maybe not. Perhaps a combination of atypical Layers and other functionality within ArchiCAD makes these concepts even more detailed and versatile. More importantly, hopefully you are now thinking of an 11th and a 12th model type—and are inspired to dig deeper into alternative Layer methodologies.

Framing LayersAll the Crazy Layering Systems you need to try.

To make many of these modeling types work, especially if you need to temporarily create one, you need to quickly create a One Layer Model (or a model with a limited number of Layers). I recorded a video on how to do this and wrote an accompanying blog post on Shoegnome with more information about creating a One Layer Model. In the blog post I also talk about the larger concept of creating models with simplified Attributes. As that post ties into the discussion above, it’s definitely worth reading. Here’s a link to the blog post. I’ve included the video below.

Are you following Graphisoft North America on Twitter? Click Here to keep track of all the latest ArchiCAD News in North America (and beyond).


  1. D.


    First, thank you for your article series about layers, even when I confess have the last two on my “read later” list. Anyhow they are a great take on the never ending discussion about layers.

    I just have a comment. You mention using layers to deal with design options (Been there, done that also) and I wonder if you ever met or used the design option feature Graphisoft included in a short lived AC version geared towards USA general contractors. Even when they took it out from AC, it was a fair first attemp and I hope GS still has it shelved somewhere.

    On the same design options topic, have you ever tried using the remodeling capabilities of AC for alternative design management? I guess if we use renovation filters with design option names, the could be up to the task and without messing with layers. Hope I expressed myself clear on this last part.

    Thankyou again and keep the good work flowing,


    • Jared Banks


      I know of the long lost AC version for general contractors, but never had a chance to use it. I think it actually was before my time or when I was so new to ArchiCAD that I didn’t have a good sense of what was going on. I think the latter, as if I recall correctly that was around v9? Maybe. I don’t know!

      As for using renovation status and filters for design options. No I haven’t used it for that, but I know anecdotally of what to do. One of the interesting things you can do is select elements and have them only show on one Renovation Filter. (Show on current Renovation Filter only) This actually allows for a lot more options than just having the base model as existing, option 1 as demo, and option 2 as new. But it also sounds a bit like a nightmare.

      Honestly my working method for versions of a scheme has nothing to do with Layers at the moment. I do the base model, save it. Then do a Save As for Scheme 1, a Save As for scheme 2, etc… There are some tricks to streamline the process that involves making all the schemes, then setting up one file with all the views, layouts, renderings, etc., then copying all the schemes back into that file doing a Save As each time. So that the process goes: existing conditions, schemes, documentation for schemes, copy schemes back into documentation file. I’ll do a video or a proper post or something. Basically what that allows is to be loose for design, explore different paths, but only have to set up all the documentation once (more or less).

  2. D.


    Thank you fir the tip on the renovation filters. I didn’t know it, i will tried as soon as I come back to the AC trenches, on personal leave now due to family issues.

    I have used a method similar to yours (save as Option #) in the past. It a workaround but still had some drawbacks in our case:

    – It works fine for single residential projects, but when dealing with 40+ mixed use buildings… It gets complicated.
    – If after three or four options the client asks for a “hybrid” that uses portion of the previous design options (this happens almost always) things become interesting.

    In any case, that’s the beauty of our profession.

    I follow your blogs, so when you pitch the post or video I will take a look, and comment if I have time.

    Thank you,

    • Brian Lighthart

      Whether using layers or renovation filters, the problem arises: We have an existing building to be remodeled. Element 10 (a piece of new wall, let’s say) is included in option A and C only. Element 15 (an existing slab, let’s say) remains only in options A and B. If my typical template model has layers (or reno stati – that’s plural of status) of Existing, To Remain @ Option A, To remain @ Option B, and To remain @ Option C. I can only assign an element to one layer. So, to what layer do I assign the wall? To what layer do I assign the slab?
      So, do I model multiple slabs and walls, assigning each to a different layer or reno status? Or, do I create layers for each possible combination of stati, and layer sets off each option?
      It gets really hairy either way, and hairier with each added option and each added new or demo’d element. The defunct Options Manager (if that’s what it was called) had the same problem. Before that even, Graphisoft proposed creating multiple “sign-ins”, each showing a different option, sending changes from the adopted option, and signing out of the other options without sending changes. The examples were trivial, and all work essentially needed to stop while the options were being evaluated. That’s not the real world.
      To make matters worse, what do I do with an element that is existing, but is being relocated? Or, a composite wall that is being stripped to the studs and re-sheathed to turn it into a shear wall, and then re-faced with new drywall? Only the framing is “existing to remain” ? (Oh, by the way, the doors in this wall are being salvaged, but need the frames modified to fit the new wall thickness).
      The tool that cracks this nut elegantly for commercial as well as residential work will rule the remodeling world.
      What we usually wind up with is a tree of models with each representing a single option in a solo model. These are basic prelim design models, so not real heavy with detailed framing elements or complex profiles, let alone moldings, etc. It entails some duplication of effort or cut/paste between all the versions, to keep the models moving forward in parallel while a client is evaluating options, which sometimes takes a week or two. But in the end, it’s preferable to trying to accommodate all of the options and hybrids in a single model. It also leaves a trail of “bread crumbs” allowing us to return to a previously discarded option that was more appealing to someone who was not involved in the choice until after it was made, but who holds more sway over the organization than anyone who was involved.
      The chosen option is then carried forward until we need to produce another set of options.

      • Jared Banks

        Brian, the solution you describe at the end is what I do too. I like the bread crumb trail because it also means at any point in the future, it’s easy to add something in——when the client inevitably says “remember 6 weeks ago, you showed me X and I hated it? Well actually…let’s add that!”. The robustness and anti-fragileness of the Save As solution is really hard to beat. And if you do all the versions with the idea that anyone could be a deadend, then you can increase speed and playfulness to the models (the main model stays pure and gets developed, while these option models are rarely meant to be developed to CDs——until farther in the process when that would be the case).

        I too long for renovation status by skin, multiple options/phases, and the ability to give an element a unique location for each of its’ renovation statuses. The combination of those things would be fantastic. It all amounts to giving elements a time dimension, which would be incredible. It’d allow construction phasing, etc. So, so wonderful. I’ll have to write something up that codifies that concept and spread it around. 🙂

        Great comment and thanks for pointing out the complexity of options and renovations. I agree, solving that is a money spinner!

  3. busra

    hi, how can we model shear wall like right side of half circle left side is finished as rectagular corners?
    help me… really urgent

    • Jared Banks

      It sounds like the construction method: polygonal for the Wall Tool would be what you are looking for. That construction method for the Wall Tool allows you to create a wall like you would a slab. Or if that isn’t right, try creating the element out of a curved wall and a straight column. Or the reverse. Or you could do it with a Complex Profile Column or Beam. Or a Morph. Or perhaps a Shell. A Slab should also work.

      It all comes down to what the final effect/look/data you are looking for. Check out my posts on ARCHICAD Tools which go much deeper into picking the right Tool for the job.



  1. gbXML export add-on (for free from Cadimage) - Shoegnome - […] to explain an add-on and also teach someone about modeling. Recently on BIM Engine I wrote about different kinds…

Submit a Comment

Your email address will not be published. Required fields are marked *