How to leverage the power of computation to preserve design visions

PRESENTER: Hello, everybody. First of all, thanks for attending this. The first thing I wanted to do before even introducing the talk, I wanted to get a sense of who is in the audience. So I wanted to ask you, who in the audience is an architect? Can you please raise your hand? OK.

And engineer? Perfect. And who is on the construction side maybe? Great. And then, is there anybody that doesn't fall in this category? OK. So I'm very curious to know.

AUDIENCE: IT.

PRESENTER: IT, perfect.

AUDIENCE: [INAUDIBLE]

PRESENTER: Perfect. OK, sounds great. So this talk is going to touch architecture, engineering, art, construction, it's going to be a big mix. My name is Alfonso Oliva, and I'm the director of LERA plus. LERA plus, it's a computational and research group. We are based in New York. We are part of LERA, former Leslie Roberts and Associates structural engineering firm.

And a little bit about my background before actually diving into the presentation. The reason why I ask the question who is in the audience is because I'm actually mixed myself. Don't get tricked by my New York accent, I'm originally from Italy. That's where I started my studies. I started my studies as surveying engineering. And I started actually practicing. And then after that, I got my Bachelor and my master's in civil engineering. After that, I got this great opportunity to move to New York where I studied structural engineering and I started working as an engineer.

My first experience in engineering was design of tall buildings. After a year designing tall buildings, I understood that I wanted to do more. I was very much, in a way, bored by the tedious tasks. And I wanted to automate these tasks. So what I did, I basically started scripting a little bit every now and then. And then, I understood I needed more. So I went for a master's in architecture with focusing computational design. So that's where the mix comes. After that, I kept working, I saw the value, and I looked for new opportunities, and that's when I moved to LERA, and I founded the LERA plus.

We had a group of by five people now. And we work in a very big mix. You'll see. We work from everything from skyscrapers all the way down to furniture. The reason for that-- we also work on sculptures-- is because I'm an artist myself. I'm actually now conducting a PhD in computational design with focus in the arts, and specifically in sculpture design.

I hope this gives you a good overview of the reason why LERA plus was founded. One thing that is at the core of the group, it's this. We basically believe in preserving architectural visions. We understand how important it is to preserve visions, both in architecture and both in the arts. And we also understand how sometimes engineering can step in and can modify that vision. So we really preserve visions like in architecture, in the arts. And we do this through engineering algorithm aided optimizations and custom software solutions.

We work, as mentioned before, across any kind of size, everything from extra large to extra small. Today, I'll have five examples, five projects that we worked on, and we start from extra large, and we move all the way down to extra small. Anything that you see in this talk can, of course, be applied to any scale. So I just tried to show you some significant examples.

The first one is the HTT Tower. It's a project that we worked on recently. The location is Luanda, Angola, and it's with the architects [INAUDIBLE] of Zapata. This is a rendering of the building. As you can see, there is a main podium at the base, and there is two towers connected by a bridge where there is a pool. And we'll see also a shot of the pool in virtual reality that is part of what we did for this project.

As you can see, as LERA Plus, we provided structural discretization, and translation across finite element analysis and BIM. We usually use Revit for documentation, and we did our discretization in Rhino and Grasshopper. And then we imported the modeling to Etabs-- that is one of the software that we use for final element analysis.

This is an example, just like a screenshot, of, like, different models across different software. And we understand how we did that. We developed-- as I mentioned before, we are very much custom software development driven. As matter of fact, in November, we ended up opening a software development company. And in order to achieve something like this, we actually developed our own interoperability platform.

This is an example of another platform that we developed. It started with another project, actually. And now, of course, we use it across every single project that we have in the firm. This is an example of how we were furnishing a space, and then moving furniture around. And I'll go a little bit in detail about this platform, because we think that this is part of the future of the AEC.

It's a virtual reality platform, and the name is Immersify. I don't know if any of you in the room is familiar with it, or run into it. It's on different parts on the web, so you might have seen it. We just kind of started marketing it. And the reason why the platform was born was actually this.

So these are all the players that are on a typical project that we work on every day-- whether it's a building, whether it's a museum, whether it's a sculptural piece, or whether it's, like, a piece of software. So in a typical building like this, you might have engineering, architecture, somebody in the construction side, and the owner. And what happens is that every time they need to discuss something, they need to get on a call, or maybe in a room. What you do in that call is that you understand if there is, like, any clash. It could be, like, any P against structures, or structure against architecture. And then you need to document it. Right?

The first step of documentation is to annotate-- like, so take notes notes during the meeting. Once the meeting is done, you put that on a plan, maybe-- like on a plan view-- maybe by hand, or maybe with Bluebeam. And then you transfer that information to whoever is working on the BIM model. What happens after that is that everybody does that-- like, on the architectural side, on the engineering side, on the MEP, and so forth. And then you have to get back on this call, or back on in this meeting, and do exactly the same thing. So this process is very time consuming, and sometimes frustrating.

So what we want to do with Immersify is that we want to break the loop, and we want to make everything easier for you, in terms of collaboration. I'll have examples and short videos to show you some of the functions that Immersify has.

And what we want to do, in brief, is that we want to get everybody in VR, in the real model-- model that has the architectural 3D model, the engineering 3D model, MEP-- whatever you want in there, whichever contents you might have. And we want everything to happen in there. We'll see how you can annotate. We'll see how you can move elements. You can add 3D content, whether it's a chair, whether it's a column, or it's a beam-- like, anything. Or a piece of-- you know, for the MEP.

And then, after that, you can actually either save your session, or you can import all of these into Revit, for instance, or into Rhino, or into SketchUp. We support different platforms. And then again-- these are always not very visible-- but then again, we want to go back into virtual reality for the next coordination call. So we want to streamline a workflow. That is a big part of what we do at LERA Plus.

So these are some of the functions that we cover. I'm going to start from this one-- Multi-user. That is really the reason why this was born. Right? We wanted everybody. We wanted like five people, or 10, or 20, to be in the same space at the same time. So the platform supports that. And think about this as a game. I don't know how many of you has ever played Doom, for instance-- Doom 3D. It's a very old game, but pretty beautiful, I would say. And our VR platform, as any other platform out there, is actually built on a game engine. So it works exactly the same way.

What you can do with it is all of these. So the second thing that goes hand-in-hand with Multi-user is the Remote Connection. So you don't have to be in the same room to do this. There can be one person that is in New York, one that is in Australia, one that is in LA, one that is in France, and so forth. And they can all connect in the same room, at the same time. So you just join the VR, you share a link, and that's it.

The other thing that is pretty interesting, and I mentioned briefly, before, is the live editing function. Once you are in there, you can do a bunch of stuff, and I'll show you some of it now. And then you can either save your edits-- and this means that you save it locally on your computer, and you don't affect any of the 3D model-- or you can actually export that into the 3D models. And Revit is going to notify you. It's going to tell you, hey, this piece of furniture has been moved here. Do you want to accept that?

You can annotate. And one of the most beautiful things is this-- it's the One-click Connection. So we support Revit, Rhino, SketchUp, Navisworks. And any kind of model that you have right now, today, with one click you can jump in virtual reality directly. And if you're thinking about, wait-- to do this, I need one of the VR-ready computers and I need a headset.

Yes, you need that. But Immersify also supports this kind of computer, that is not a VR-ready computer, or any kind of tower that you might have. So that means that you can jump in virtual reality right ahead, and you can control with your arrow, like, where do you want to go, and you have all the functions in there. So you don't need any special equipment. The VR equipment just to be, actually, like, in the space, and have different kind of feelings.

So we use this a lot, and it's actually out there. It's a licensed software. And these are some of the functions. This is a tool that shows how you can hide things. The interesting thing is that when you import a Revit model in here, every single parameter attached to the element-- it's in there. And then you have an Inspector. So with your left hand-- you have a little screen on your left hand in VR-- you can point to things, and it's going to tell you, OK, that column, is it 30-by-30 in concrete, and every single property that it has. A piece of MEP-- it will tell you exactly what it is, and so forth.

The hiding tool, it's by categories-- same hierarchy as Revit. We just followed that, because we believe it's a very great way to do it. And then this is a function to strike dimensions. They stay there for a bit, and then they disappear. Very useful.

Then, the next tool I wanted to show you-- it's the Tag tool. The Tag tool is this little diamond that you see appearing, here, in a second. And that diamond is very powerful. Because once you attach a tag to something, then that tag can contain almost anything. It can contain measurements. It can contain a video that you can take, and you can talk over. It can have an audio note. It can have a sketch. It can have anything.

And then you can import that tag tool inside Revit, or Rhino, or Navisworks, so whoever is working on the BIM model can actually go in there, click on a tag, and in your Revit panel, you'll have your voice note, you'll have your video. So you can say, hey, there is a clash here. Can you please fix it? So all of the sudden, you don't need to call somebody up and then double check that it's done. They can mark it as "done" within Revit. You get the notifications, and everything is done.

The next function that I wanted to show you is the Move tool. You can actually move elements. All of this is based on MEP. And the graphic that you see here, it's in Coordination Mode. This means that it's very raw, and that's what people prefer when they're working on MEP or engineering. We have also another mode, that is called Presentation Mode. That is, basically, a rendered view. I think you've seen a lot of that in VR, so today I wanted to show you the other side-- more on the coordination side. But please keep in mind that if you want, you can have a beautiful rendered view, as well, with Immersify.

This other function is super-powerful. It's called the Library. That Library is nothing else than a folder on your desktop, where you can upload any kind of 3D model that you have. It's going to appear in there, and then, in VR, you can actually place it. You can move it, you can rotate. So all of the sudden, you can be in VR, and you can design. You can furnish a room. You can, if you want, design a whole building. You can add a column, beam, and start designing the whole building, and then transfer it to Revit or Rhino or anything else that you might be using.

This is another function. This is kind of the beginning of what we are exploring now-- that is analysis. This is the first step towards environmental analysis. Here, we basically embedded function to study the sun path across the whole year, at every single hour, and every single date towards the year. So this means that you can understand how your building might be affected in terms of shades by the surrounding buildings, and so forth.

OK. Let's scale down to large. I don't know how many of you are familiar with the Lucas Museum of Narrative Art. It's a project that was born with the idea of making it happen in Chicago. So that was the first location, and then it moved to Los Angeles. Now, it's actually under construction. We've been working on this for quite a while, with MAD Architects and Stantec. This is the LA option. Today, I'm going to focus mostly on the Chicago option, because we can't show much of the LA option. But what we've done for this, we've provided computational support, ongoing computational support. Because, as you can understand, on a project like this, complex geometry outside and inside the building, as well.

So we've done a lot for it. And I'm showing you, today, some of the parts. Again, very much focused on coordination. We want to streamline workflow. We want to make our life easier, but also life of people that work with us easier. So what we noticed when we started this project is that there were a lot of people involved, of course, and a lot of software involved. Because in our industry, sometimes-- actually most of the time-- each single person is probably using three softwares, or four, that are different from your softwares.

So this was the first question. And it was like, how are we eventually getting all of this together, to build this? From that, that's where Cuttlefish was born. Cuttlefish is an interoperability platform for the specific workflow for Lucas. That is what I'm showing you now. That's what we did. We were receiving-- every week, we were receiving a Rhino model from the architect. Sometimes, even twice per week. What we needed to do-- when needed to discretize the model in order to build our structural model. And then we needed to document that, and send it in by the end of the week, and we needed to, of course, perform structural analysis.

And these two are very linked to each other. Right? Because while you're doing this, and once you're done, then you need to communicate all the member sizing, and everything needs to be documented. As you can understand, on a project like this-- and on many other projects-- this is a very stressful process, sometimes. So the reason why we built Cuttlefish is because we wanted to streamline that. And we did it. We had the Grasshopper definition that, as an input, just the new geometry from the architects-- we basically built a model that will dev every single structural element in the whole entire building.

After that, we had a custom component that we developed, that will basically read all this information, store it in a database, and then, through our custom application, Cuttlefish, we could automatically rebuild that in Revit, and in our software for structural design-- in this case, SAP2000 and Etabs.

The platform, today, is more versatile, and it supports other software. The reason for that is because some of our clients have asked us to extend it to other software. And the next step that we wanted to take, here, was, OK-- but what happens when, next week, this is changing? Like, what's happening to this element? How do I know that this element changed here, or changed here? And if you have three people working here, or five, how do you know who is updating what? And how do you know that that's actually updated in the Revit model? That's a lot of QAQC that you need to do. Right?

And that's what we did. We brought this to the next level. And what we did, we built a system within Cuttlefish that's called-- we call it Compare Tool-- that works across every single platform. And what it does, it asks you for two very easy things. Two models. So you can upload a Revit model, and a model from Etabs, for instance. What it does, it compares them, and then, in output, it gives you this. It's a centerline model, where the green elements are elements that are there, and were there, in the previous model. The red elements are elements that, as you can see here, have been deleted. And then orange elements are elements that have been moved. So all of the sudden, you can explore and you can understand what's happening in your BIM model in no time. You can say, OK-- 10 elements were moved. 20 were deleted. What's going on here? Let's fix it. So you basically keep time for other tasks.

This one is one of the workflows that we followed. You'll see a lot of Excel to handle the database. This was five years ago. Now, everything is on the cloud. We updated everything. This is an optimization that we did on the outside facade of the building. It was to understand how much concrete we needed to use outside-- shotcrete, in this case. And we were able to optimize that and, of course, save a lot of quantities and money, on a project like this. And this is part of the workflow on how we use Dynamo to actually import all the [INAUDIBLE] in Revit. So the amount of drafting in Revit was basically close to zero for the whole structural system.

We were able to import-- that's a shot of the whole model. We were able to import every single element. And then this one is the second part of the workflow, where we are importing the modeling to, in this case, Etabs-- our software for structural design. This is just a quick teaser about a perimeter truss that is basically something that was changing a lot. It was changing, maybe, two times per week, or three times per week. So what we did-- we basically found, with Dynamo, a way to update that automatically into Revit. It's not shown on this video, but later on, we also automated the documentation of it.

I have another example on how we automated documentation on this project that I'm going to show you right now. And it's connections. Right? On a project like this, you have 200, 300 connections that you need to document. They are all over the place. You need to create all these views. Something changes, everything needs to change. Right? Revit takes ownership of part of that, because it's very good at updating visuals. But you also need to be in control of that. So what we did for this-- we developed our own workflow. This is part of the building, just to make it clear, a little bit.

So what we did, we brought these Dynamo definitions, that what it will do, it will generate rectangles at connection locations, and at these. Each single rectangle was automatically generated. The second part will create the callouts. The third part of the definition will name the views according to the location and specific group that we input. And the fourth part, the last one, will generate sheets, and automatically place the views on the sheets.

So you can imagine how this can save you-- across the span of the project, it can save you months of work. Imagine updating these all the time, and keeping track of it at the same time. So we wanted to take that to another level. So we were like, wait a second-- we've been doing all these translations across these softwares that, they don't talk to each other, but there is a way to make them talk. What if we actually push it to another level, and we get a PDF, and we try to embed that information in Revit?

And that's what we did. I think most of you are probably familiar with Bluebeam-- if not, probably, everybody. And you know how much information is in a Bluebeam drawing-- how much do you add, how much somebody else add. And then eventually, that needs to get-- somehow-- into Revit. So what we wanted to do-- in our specific workflow, the way it worked is that we had two people marking up Bluebeam drawings. And in this case, it's for rebars. Right? And as you can understand, these tags are probably-- across the whole process, there are probably, like, thousands of tags. And at the beginning, what we were doing is that, we had engineers marking that up, and then we have the BIM team literally redrawing that in Revit. So redrawing the tag, and copying the tag.

So what we did, we basically accessed Bluebeam, and we got all this information-- so tag, tag content, and location-- and then we automatically were able, through Dynamo, and through a little custom component that we developed, to tag automatically every single rebar on the whole project. Across the whole project. So this was very interesting-- up until one engineer came over, and he was like, this is great, but I have no way to double-check if next time we do this process, something changed, and if that's actually a good thing or a bad thing.

And with that, we introduce revision clouds, that nobody wants to do, but they are there. And that's what we did. We basically added to it. We added a function that will basically preserve the old database. It would input the new database, double-check them, and automatically draw revision clouds when something changed. So as you can understand, all of this-- everything that I'm showing you, and that I will show you-- it's kind of a snowball. If you start very small, and then you keep building on it, one, you keep control, and second of all, it becomes something super-powerful at the end of the day, that you can actually reuse on other projects.

So let's go down to medium. This is a confidential application. It's a project in New York. And it was a very fun study. So we are very well known for optimization, at LERA Plus, for structural optimization, for reducing quantities. We brought tons of projects back to budget. We do custom optimizations-- meaning we develop our own software and our own algorithms to solve structural challenges. And this was one of the examples.

So complex geometry-- you'll see a 3D model. Part of it is a 3D model after, to understand how complex this is. The problem here was to understand, OK-- how do we place rebars here? And what's happening to the rebar? You know, they had no control over it. And of course, they wanted to put a price tag to it. How much is this going to cost?

To show you this, we are going to focus on a three-story part of the building, just to keep focus. And this is part of the study. This is the 3D model of that portion. What we wanted to do, at the end of the day, we wanted to understand which bars were straight, which bars were hand bend, and which bars were machine bend. And of course, the red is where the money is, because when you need to machine bend something, it means that it's going to cost more.

This is within the context. So now, let's dive in. OK. So we have two parts to this. We have an interior part, that we're going to call gallery side, and then an exterior, that we'll call canyon. And just to keep focus even more, we'll focus only on the canyon side. We focus only on the exterior part. And you'll see that at the end, it's the same thing.

Let me clarify something-- I promised you blue, green, and red as straight, hand bend, and machine bend. And I see that, here, you can see many different colors. The reason for that is because there are horizontal and vertical bars, and that's just, like, what happens when you overlap the colors. So don't get tricked by that.

So here, overall shots-- vertical bars, horizontal bars. This is the first step. Right? So as you can understand, what we did here-- we put together a definition that would allow us to build all these rebars, and to understand, also, which one would be straight, and which one would be machine bend, or in the middle-- so hand bend.

Now, imagine to get these three-story building rebars, and laying them down on a huge table. Needs to be very big. Right? What you can see here-- and this is just to show you how we did this-- you can see that there are some bending points. Right? And then lots of points which are straight. So what we did, here, was very simple. We just had a limit that was the angle between you can actually hand-bend a bar, and when it actually needs to be machine bend.

But this is very powerful on its own. Because you can do that for both the vertical bars, but you can also do that for the horizontal bars. And all of the sudden, what you have out of this-- it's not just a pretty drawing that you can print out, but it's actually data. That is what I'm talking about, here. And that can be translated into-- later on, it's this. That it's basically a drawing that can tell you exactly the length of a bar, the angle-- sorry. This shouldn't be zero. This should be infinite, because it's straight. That's the radius.

And then which one is a hand bend, and the length, and which one is machine bend. And as you can understand, you can document this on drawings. Yes. But you can also send this directly to the machine that's bending the bars. So all of the sudden, you not only solve the problem, but you actually streamline the workflow also for other practices. And that's a big part of what we do at LERA Plus.

The last step was this. We started from a 75%-- for only this part-- of red rebars. So that's a lot of money. So what we did, we put together a small definition that was tied with [? Part. ?] [? Part ?] was a custom plug-in that we wrote. And it was connected to an evolutionary software. What this evolutionary software was doing was basically changing the shape very minimally-- meaning it was pushing the boundaries one inch in and out. And of course, we talked about this with the architects, and we agreed that with them.

And as you can understand, on this three-story portion, one inch in and out-- it's something that you will never be able to see. So we are still preserving visions, because, as I mentioned before, we're very big on that. And with this small effort-- that, of course, was just for the initial study-- we were able to go down from 75% to 45%.

Another example is a sculpture in New York, that is built-- it went up probably two months ago, something like that. I don't know how many of you are from New York, or if there is anybody from New York, but it's on 31st Street and First Avenue at the new NYU building. Go check it out, if you want. It's a beautiful sculpture by Alyson Shotz. The client was NYU. We were not the engineer on the project. We were the engineer for the building, but not for the sculpture.

So we were called in for a coordination call on the sculpture. And we needed to coordinate, basically, where the cable were hitting the structure above. So we were responsible for the beams above. And what happened during this meeting was that the engineer that was designing the sculpture announced that the sculpture would have deflect 93 inches. So the artist, as you can understand, was no-- this is unacceptable. We can't do this.

So we advised to conduct one of our optimization. And that's what we did for this project. Part of it was geometry discretization, part was structural optimization. And then you'll see how we went all the way down to the step right before production-- fabrication. Fabrication was by C2, in New York-- C2 Fabrication.

So the first thing that we did, we wanted to understand where these 96 inches were coming from. And what happened is that, the engineers that were on the project, they were using four cables to hang the sculpture. To give you a sense, by the way, the sculpture is four stories high, and it's hanging from the ceiling. So we first created exactly the same conditions. We used four cables, hang the sculpture, put it in our FEA software, found the 93 inches of deflection. And we understood that part of it was due by the self weight, and part of it was actually due by the location of the cables, because there was a swing of the sculpture.

So that's what we did. We basically broke that down into two contribution, and we moved into optimizing the first contribution. We promised the artist that we would have preserved her vision, 100%. That was, at that time, maybe, like, an over-commitment, without working on the project yet. But we went for it. So we worked on the first contribution-- that is the location of the cables.

The first thing that we found is that we could have got away with three cables. That is actually what the artist wanted. And that's what we did. We built a model with three cables, and we connected this model to an evolutionary software that was running a genetic algorithm optimization. And it was connected to finite element analysis at the same time. So it was measuring the deflection of each single option. We ran through thousands of options, and at the end, we were able to reduce the deflection from 93 inches to 45 inches.

So part of it-- actually, sorry. To 25 inches. So we were happy about that. But what we really wanted to do, we wanted to get to zero, because that's what we had promised to the artist. And that's what we did in the next step.

Before I moved there, some of you might have seen that the sculpture sometimes gets out of the screen. The reason for that is because of the specific algorithm that we are using. I don't know how many of you are familiar with genetic algorithms. Maybe somebody? Oh, perfect.

So as you might know already, some of you, what the genetic algorithm does, it basically goes around and it explores, like, every single option that you might have. It is affected by the initial state of the optimization, when you're setting it up. So you need to experiment a little bit. And then what happens is that you start finding points of relative minimum, in this case, because we wanted to minimize the deflection, and you start converging to it. So what you get to at the end, it's a relative minimum. It's not an absolute minimum. And of course, there are other algorithms that can get you to something-- maybe more refined, if we want to call it like that-- or to a different solution. But we explored different algorithms, and this is the one that, in this case, gave us the best result.

So the second thing that we did-- I don't know if you remember, but I've shown you the blue sculpture at the beginning. This is just an overlapping of three models. The blue sculpture is what the artist wanted to begin with. It was her model, her 3D model. And as mentioned, now we are down to 25 inches, and we need to deal with that. So what we did, we basically took the result from the structural analysis from the deflection, and we cambered the piece.

And it's this red shape that you can see here. So by cambering the piece by-- at the very bottom, of course-- 25 inches, and then as you go up, it decreases. What happens is that then, we put the sculpture back into [INAUDIBLE], and we let it relax under self weight. And the lavender color is basically the relaxed shape. And as you can see, it was basically off by half an inch, or something like that.

So there you go. Then you are back to the original design. And all of this was to serve more as-- the reason why we show all of this is because we want to serve as inspiration. And we hope you understand the purpose of the whole presentation.

So once we did that, it was great. We sent the model to the fabricators, and they were like, wait a second. We can't build this anymore. They were using a CNC machine. The reason for that is because all of these ribs-- the horizontal and the vertical-- were not perpendicular to each other anymore. They were not 90 degrees. So what we did, we wrote a custom script that will go through each single of these frame, and basically straighten them up, within a certain limit, to make all of them 90 degrees from each other. We sent the model back to the fabricator, and they were like, OK-- good to go.

And after that, they basically fabricated the sculpture, and now it's built, and it's in the space. It's a sandwich of Plexiglas, aluminum in the middle. Outside, there is a film that, with the sun, as you can see, it basically gives these beautiful reflections.

OK. Let's, now, move to extra small. That's perfect. I'm going to leave some time for questions, if you guys have any. Let's move to extra small. And this is an art installation. The location is Shanghai, and the artist is James Clar. He works with-- James works with a lot of lights. The collaboration that we had with him, it's this piece. It's called "First Landing." It was first shown in Shanghai, and it was recently shown in Chelsea-- in a gallery in Chelsea, in New York. What we did for him was the following.

James was building all these beautiful sculptures that are composed by LEDs, and sometimes, also by tubes. So it's a mix of LED and tubes. And of course, you need to have a way to connect them. Right? So all these connections that you see here-- not in this project, but in previous projects-- were modeled by him. So he was modeling all of this by hand, in Maya, actually. And so he will model each of them. Here, they're all similar for this piece, but we'll see another example in which every single connection is completely different. And it was a very time-consuming task.

So what we did for him, we wrote a custom plug-in that would allow him to just draw lines, and then, to each line, assign whether it was an LED, specifically, whether it was a tube, or whether it was something else. And then, what the software will do, it will automatically create a part-- so a 3D model of a connection that was ready for 3D printing.

OK. So this is a shot of the installation that was actually in Chelsea. So this is another piece from him. This is James. This is me, and I'm pointing to him my favorite connection. I don't know why, but I loved that connection. As matter of fact, we are reprinting it. Putting it on my desk. It was just a beautiful thing. Because, I don't know how many of you have worked at that scale, but there is a lot of work into it. And all of these, of course, they need to have holes to host all the wiring, and then they need to be Boolean together, and they need to be ready to 3D print. And all of this is automated.

So what happens is that, from these lines-- just lines-- you get this, ready to go. And then you can attach your LEDs, or your tubes. So this was, like, our first kind of collaboration. We work with a lot of artists on sculptures, on interactive installations-- like, a lot of that. I'm actually an artist, myself. I didn't mention it at the beginning, but that's also part of the reason we do this.

And this is a final shot of one of the connections. So pretty beautiful. Now we are working with James to bring it to a whole other level. So we are conducting another study together, with him. And with this, I end my presentation, and I thank all of you for being here. And if you guys have any questions, feel free to ask. Thanks.

[APPLAUSE]

Yes?

AUDIENCE: What languages do you use in software?

PRESENTER: As to what? Sorry.

AUDIENCE: What languages, [INAUDIBLE] languages, do you use to build your software?

PRESENTER: Yeah. OK. The question is, what languages do we use to develop our software. So it depends. We use, like, a mix of different languages. It really depends on where you are-- meaning in which software are you operating, in case you are writing a plugin. And if you are developing something from scratch, you can go, you know, with whatever is needed.

So to give you some example, to develop Revit plugins, most of the time we go with C#. There could be some Python within Dynamo. On the Rhino and Grasshopper side, a lot of C# and Python. If we develop something from scratch, we could either go with C# if we need to interact with any of this software. If it's something from scratch, it could be C++, it could be JavaScript.

We also built this interactive pavilion, that we are going to release soon, that is basically-- it's not really, you know, Dynamo or Grasshopper on the web, but it's something similar. So it's parametric structures on the web. That one was a mix of JavaScript, HTML, CSS, and so forth. We also do a lot of web development, in terms of, like, websites, and stuff-- but just on the interactive part. So we've used, for that-- yeah, mostly JavaScript. And some interesting libraries, like 3JS-- I don't know if you guys have ever used it-- just to enable some interaction. Yeah. Thanks.

AUDIENCE: [INAUDIBLE]

PRESENTER: The curved surfaces in Revit? Which one are you referring to? Do you have a specific one? For the Lucas Museum, or--

AUDIENCE: Yeah. The Lucas Museum [INAUDIBLE]

PRESENTER: The curved-- which one? Sorry.

AUDIENCE: Curved beams.

PRESENTER: Yeah. The beams in Lucas-- the beam, itself-- they're actually straight. The structural system is a straight beam on an angle that goes up, and then a horizontal beam-- a straight column. Sorry. And then an horizontal beam, and so forth. That's the whole structural system. The curved surface outside-- what we usually do, we usually import it from other softwares. In this case, we built it in Rhino. And then we have ways to, with our interoperability platforms, to basically transfer it into Revit. So it's a custom process. Yeah.

AUDIENCE: [INAUDIBLE]

PRESENTER: No. The way it works-- thanks for the question, because that clarifies something. The way it works is that, if you are the owner of the session-- so let's say that you start the session, and you load all your 3D models in there-- then, when you share the session, everybody is able to populate the model, and so forth. The other interesting thing, and the reason why we started this software developer company, is that Immersify, as any other software we developed, it's owned by us. So we can actually make changes on the fly, when people need. And we've done that with some of our clients, of course, as well as integrated new functions.

AUDIENCE: Do you have a good example of where you've seen the most success [INAUDIBLE]

PRESENTER: Yeah. So [? JBB ?] is using it a lot on MEP. One of the most successful-- the reason is just because we are in it-- is the Lucas Museum of Narrative Art. The reason for that is because it has helped, like, tremendously on streamlining all these coordination calls that otherwise would have taken, probably, five to six times the time. The other thing is that we have people in New York and people in LA, of course, working on the project. So daily, everybody is in the same model-- you know, in the same VR model-- and it's basically marking things up, and then sending everything to the BIM department for documentation.

But MEP, as well-- with [? JBB, ?] as I mentioned before-- we also had, in terms of the sculpture-- what the owner wanted to do, they wanted to decide the location of the sculpture. So we set up a meeting, and it ended up being a 50 minutes meeting. There was this thread of email saying, no, I want it more on that side, more right, more left, and so forth. So we sent the invite for the VR, people from their laptop, as I mentioned, even if they didn't have the headset. Everybody logged in, and we were literally moving the sculpture. OK, where do you want it? One foot that way. Three feet that way. One foot down. Done. Import into Revit, and everybody had it.

So it's just something that can-- it's up to you, how you use it, because there is so much to it. And again, it's so versatile, and we can add any kind of function that you might need, that I think that's the biggest advantage of any of it. I hope it answers your question. Great.