Conceptual Structural Design Using Revit Adaptive Components and Dynamo

DESIREE MACKEY: All right, and we're going to go ahead. And we only have an hour. So we're going to go ahead and get started and just let people still trickle in. OK, so my name's Desiree Mackey. This is Brian Mackey. We are married. Just get that out of the way right now.

How many people have been to one of our classes in the past and already knew that? OK. All right, have anybody never been either one of our classes? OK, well, this should be interesting. So we very rarely get to speak together. While we have similar interests, our work just doesn't tend to intersect.

BRIAN MACKEY: Which is ironic, because we actually met at Autodesk University. That's how geeky we are.

DESIREE MACKEY: How many people knew that?

BRIAN MACKEY: At a speaker socially.

DESIREE MACKEY: We met at the speaker social right here at AU. So we very rarely get to do this. So we're excited about it. But I just want to warn you. How many of us have children? And have you noticed that after you feed those children they go from being this nice, calm, little package of cuteness to it something literally runs around the room? And he does the same thing after he eats. So--

BRIAN MACKEY: And well after lunch. So if I start doing laps, you guys will understand why.

DESIREE MACKEY: All right, so when we are putting this session together, these examples came from work that I was working on. And often, when I actually get to work on a project instead of working on everybody else's projects, I do that work at home or in the evening. And inevitably, because Brian is a super geek, he and I end up working on things together. So that's how this happened.

BRIAN MACKEY: I think she just comes home because she wants help. That's what I think.

DESIREE MACKEY: So we also had a problem with doing a class together that we had an argument over whose computer to use. How many people have heard me say the number one thing we argue about at home? Keyboard shortcuts, that's not a lie. So if you notice, we have two computers up here. So we solved this problem. So we can say married by the end of the session, hopefully.

All right, so this election is-- I'll just skip right to that-- about two conceptual structural models. So I'm assuming generally, most everybody in here is structurally related or will not be offended if all the examples are structural, OK? All right, good. So the two examples are pedestrian bridge that is just about to go into construction, so real project, and then arena roof that didn't ever really get off the ground. So we've got both ends of the spectrum.

But we went through this conceptual phase. And with the bridge, we actually went through the CD phase. So I'm going to take you guys through how we did. Well, we'll take you through these things of how we did these models and these families. All right, so I have to remind myself what's up here. All right, so this is the bridge. It's three span at an intersection. And it had a very specific client-desired look to it.

It is kind of sort of a secret, so shh. But they're in construction, so it's not going to be secret much longer. So the look that they wanted was this inner and outer spiral of this bridge. But they didn't know much more about that. They didn't know exactly how they wanted the angles of the members to be, and they didn't know how it would look as it got bigger because each of the three spans are slightly different. So they wanted to be able to stretch, and bend, and change how this looked to decide what it would be.

But that's a complicated piece of structure to stretch, and flex, and bend, and be able to do a quick feasibility study saying, hey, yeah, that's going to work. Because the structure of the bridge is entirely aesthetic. So there's no hidden structural members. We just used what they wanted with the helix, is what we came to call it.

So we wanted a flexible model. They wanted to be able to flex absolutely everything. So we had several iterations of this guy. Then the roof was a little bit more simple in the scheme of things. What we got from the client was a CAD file that he had blocked out the shape that he wanted and the look.

And he goes, well, I started doing this and realized I don't even know if this is buildable. So I want to send it to you guys. And can you do a quick study on this for me?

Which, first, show of hands. How many structural engineers actually get input from their architect that early on? Sorry, I thought it was pretty--

So then we took it one step further. We're like, you know what? He wants to take this into conceptual iterations. So we said, well, how about if we just give you back a tool that you can flex and play with? And then when you've come to all these iterations you want us to check, send it back. And since we built it, we have ways to get our structural information out of that. So that's what we did.

So we're going to start with the bridge. And we're going to-- it's going to probably be fast. I hope we'll get through all of it. But please feel free to interrupt and ask questions. We tend to talk quickly, so this should work out just fine.

All right, so starting out, so the general approach of the bridge was it screams repeater, right? It repeats through this inner and outer helix. But it also needed to be adaptive because it needed to stretch and bend. And those angles potentially needed to stretch and bend. So we started with a two-point adaptive component that we brought into a host family and repeated it. And it iterated from there.

Then structurally, what we wanted was to be able to get those coordinates out of the model. And we found out recently that a lot of our analytical software talks to Excel really well when you just have coordinates. And so we can dump from Dynamo into Excel, and dump those coordinates right in, and do a real quick conceptual feasibility study. So that was the goal, the goal here.

So we started out with this two-point adapted component, which I was like, all I need is a square. I just need a representative element. I don't even care if it's structural, it's a generic model, whatever. And Brian, unfortunately, has developed a far more exciting two-point adaptive component that he's going to show you in a second that ended up been really useful for this bridge, specifically the way the outer and the inner members needed to intersect.

Think about insertion points when you're thinking about structural framing. You have the top, bottom, left and right. It became evident real quick that we needed that sort of functionality. And so we essentially made a structural beam out of this two-point component. And then you want to switch?

BRIAN MACKEY: Sure. So, of course, Desi just said she needed a rectangle family. Yeah, I don't play that game when it comes adaptive components. So, for me, I'm going to cover a little bit what it was. I've been dealing with adaptive components since they came out. I've taught a few sessions on it. And anybody here besides myself find adaptive components extremely quirky, right?

So some of the tools I have I used to eliminate that. So what originally started with this family I had is I just have this rectangular profile. And I don't know if I was deprived as a child from playing connect the dots. I don't know if my parents deprived me. But I really, when I'm doing adaptive components, like to host points on points on points on points and connect those points with lines.

It's the most stable way I've found for dealing with adaptive components. So in this particular example-- I'm not going to really go into it-- but I have my two host-adaptive points. And then these points are hosted to that. And they've just setup with an offset. So I just have an-- inside here, you'll see in the offset parameter right here I just have a parameter for half width, et cetera. So I like to host points on points to get things to accomplish. Yeah?

AUDIENCE: When you say you host [INAUDIBLE] points, do you mean you're hosting a [INAUDIBLE]?

BRIAN MACKEY: Yeah. So when you're hosting a point on a point, you can always come up here and hit the Set button. And then using the Tab key, you can get to any one of the three work planes that the point has. I am also slightly a type A personality. She might say more than a slightly. But usually what I will also do is name every single one of my points, because then it will tell me what the host point is.

So if I start going through, who here names the reference planes? You all should, because in 2018, you can actually see the name of the reference plane in the project environment. I blogged about this last week. So it's huge tip for me. But I do the same thing with adaptive component families.

So normally, if I'm going to host a point off of another point, they will all be labeled. So I'll say it like point one left, point one right, et cetera. So when I'm looking at an additional point, that will come through.

So what I ended up doing is I take that rectangular family, and I host it into then another adaptive component family. And what I've found with adaptive component families that's very buggy, is if you host adaptive component into an adaptive point that's nested, more than likely, it will screw up. Anybody put a rectangle family in, and it starts rotating and twisting, and you're like, what? Well, that's the reason for that.

I cleaned that up. We're being recorded. I usually say something much harsher than that. So what I have found is, between my two adaptive points, I will draw a reference line. And then I will host a point-- again, I may connect the dot, deprive the ability from my experience, I host a point onto that reference line. And when I do that, that eliminates the twist factor happening in your adaptive component family.

So, yeah, sure. It would have made much more sense to take this rectangle and just host it into the middle of, into the adaptive point family. But more than likely, that will blow up. So I take a point, and I literally host a point on this reference line. And what's great about a point hosted on a reference line, is that it also allows you to then move that point from either the beginning or the end of that reference line to deal with it.

So since I usually want that point to be where the adaptive point is, I always just parametrize this to allow me to move that point in and out. And I just call it control point. So I can just go ahead and move that point in and out relative to the length of the line-- I don't really care where it is-- just for the ease of editing this object. So I don't want it on the end because I've got a point on top of the point when I'm creating and editing the family. So that the only reason why I have that.

And I did have to make a tweak to this family, because originally, when we put this into the bridge, we got to the corners, and we had the big void in the corners. And graphically, we just wanted to overlap them for very quick ease. So I did redo this. And basically, this point then has this point here nested onto it. So I could do a positive or negative offset to get those to move in or out.

So if I wanted to go through here and make this move to the other direction-- so like a structural framing member with your start extension, or cut-- they changed it. It used to be start extension. You could do positive or negative. But anyways, so I basically built that into this. I nested that family in there.

Then what I found over the years is I'm constantly having one that's got an insertion point in the middle, an insertion point on the left. And I had 15 different families. So I decided, of course, to add a little parameters to this, where if I just simply hit a checkbox, I can position this stuff all around with offsets. So, therefore, I've got one family that I can have an insertion point in the middle center, I could have it in the top center, top left, top right.

So basically, I can add nine insertion points to this simple family. Just starts really coming into play, especially was really beneficial in this project here. But also, with just simply the adaptive component families where it becomes a pain-- has anybody ever put one in? And on this adaptive point, it hosts this direction.

On the next one, it hosts that direction. And you're like, ugh. So rather than loading three different families, I can just simply change the insertion point by checking that box. And it's just basically a simple conditional statement saying, hey, if that's checked, it would be half of the width, or depth, or height, or whatever it's going to be. So it's not that confusing.

And then since I nested in that adaptive component family-- so I've got the two here-- what's nice is it only goes back one step here-- in that rectangle family, these are model lines. So they can't be reference lines. They are model lines. So since you can't use a profile in an adaptive component family, you have to use an adaptive component family with model lines in it.

So those are model lines. So what's great about it, is once I got the rig all set up, I can then just select those two nested families. They create form, and I got my rectangle, OK? So it's a little bit roundabout way to get into this and start going through there. But that's really what this family allows us to do.

All right, so then once we had that done, we built this. This is the host family. And it's got-- yeah, OK. So it's got-- I ended up with eight or nine lines, reference lines that are divided. And then this adaptive component is hosted on it and repeated.

DESIREE MACKEY: And so we parametrized absolutely everything about this, the number of divisions, and the length of it. So we can have the three different segment lengths. And then the architect wanted to know-- which you'll see when Brian opens up the family-- the architect wanted to understand what were the angle of all of the members, how was it going to look. It was very about how this is going to look from the ground, essentially, and as you're driving under it and as you're walking through it. So we have a lot of-- everything is flexible. It makes for a complicated family.

And then so what we did and what I did so many times or so many iterations of this-- let's see if this works-- so this bottom member, they wanted several different iterations, one where it slants, and one where it's horizontal, perpendicular to the bridge. So there's this spiral look. But what they've found out, or what the architect found out, is that if you continue the spiral so each element has an angle, from the ground, if doesn't quite look like that. It's not the aesthetic that they wanted.

So they wanted several different ones. This one, I think, is the final one with the bottom that's perpendicular. So that's the outer helix. And then there's these inner pieces that run the length of the beam or the bridge. So then I think that's it. So here's the family.

BRIAN MACKEY: OK, so, I, as I said, using adaptive component families went along the same approach to this, as I did going along with that two-point repeater. So I still have-- in this case, they're not adaptive. We ended up at the very end just adding a length parameter between one and two, or no. They still might be adaptive. I don't remember. We went through so many iterations of this.

DESIREE MACKEY: I think this one, this is from the earlier iteration. Towards the end, it got really complicated for the architect to be-- or even the project team. I was building this and then placing it for them by the end because it was just so many things were parametrized that it was hard to flex. It was hard to understand, which I kept saying that from the beginning. But they wanted it that way.

So what we ended up doing once we had the length locked down, is I started pulling that functionality out of it. So at the end, this adaptive component was one point. And it was just set length. Yeah.

BRIAN MACKEY: Yeah, so here or there are both adaptive components, adaptive points-- excuse me. So, again, between those adaptive points, I drew a line and I hosted a point on the line, just like what you saw I do. And again, for all my adaptive component line-based type families, I do that.

And, again, here we come into-- I've now got that point there. These points are hosted to the vertical surface here with an offset. These points are hosted to these points with an offset. Then once I did that on each end, I connected those points with a reference line, OK? And then once you have a reference line in there, you can always select that reference line.

And for those of you who don't know, you can just come up here. We obviously didn't do it with this control point. But you can just come up here and say I want to divide the path. So they also-- I don't remember when we did this. Maybe this was in '16, or-- in the older versions of Revit, if you had two lines together and divide it to the path, it would divide them independently. But that's been fixed for the last few releases where you could have six lines together, divide the path, and it ignores the fact that there's a join.

So really nice feature when they added that in. It's like the old AutoCAD. Yes, I said AutoCAD, the old AutoCAD Divide command. But so I can come in here and say, OK, I'm going to go divide this path. And then you can also then parametrize it. So do you want to do a fixed number, or a maximum spacing? And we didn't go back and forth with the theory on here. Are we going to go with a fixed number parameter or are we going to start going through with a quantity parameter? And we ended up going with just fixed number because--

DESIREE MACKEY: Which was really, actually kind of difficult because the current client wanted the-- we'll called him the panel points of the bridge-- so one revolution of this helix to be the same in each of the three segments that were all different lengths. So there was a lot of coordination between that and a lot of formulas, which made me happy, about figuring out how many divisions it was, and then how many, what the length was. So it was actually a little bit complicated.

BRIAN MACKEY: Well, and it also even made it more confusing because they wanted six divisions. But we were technically, because of the angle, skipping five bays to get it to happen. So they're like, oh, they need six. But we've got to skip four. So if they want six, we got to type in 40 or something weird. So it was a little bit confusing.

But then once that was done, I'd just take that rectangular two-point family, and I'd just go ahead and place it on here. So for those of you who haven't played with repeaters or divisions, if you just do one and you say repeat, it repeats it along every single one. But like I said-- and I forget the number-- we had to skip a few.

DESIREE MACKEY: Yeah, it's four.

BRIAN MACKEY: I don't know.

DESIREE MACKEY: It's four with the flat bottom.

BRIAN MACKEY: I'm going to pull an Emeril and pull it out of the oven in a minute and go, bam! So we're right near one of his restaurants. I figure that's appropriate. But the other thing that you'll start to find-- and actually, I want to do this one slightly different-- is when you have that on there, it also matters the way you pick when you're going with adaptive component families.

So this one, it didn't really matter. You can start seeing on here. So sure, I want one that's to the middle right. But if I take both of these, why I have the family that allows me to change it, if I go from the middle right to the bottom corner-- because this one I went up, and this one I went down, you can see they're on the opposite sides of that line.

So understanding with adaptive component families, directionality, the way you draw things becomes very critical. And when you have to delete things and redo it in three weeks, you're like crap. Did I draw them going this way or that way? Inevitably, you'll get it wrong.

DESIREE MACKEY: Well, it puts in your face what we all preach to the people we work with, is always model in the same direction. Do it consistently. But when it comes up to you-- like, here, we'll model this two different ways, so it's going to behave in opposite directions-- you suddenly have to really adhere to those rules. And you're immediately seeing why.

So that was challenging with this, to get them all. Especially once you got all the members in there, and which point am I grabbing, it became complicated.

BRIAN MACKEY: So with that, what we started doing is we then just selected this and selected this. And then you can go up and use the repeat command or the array-looking icon. And then since we had three spaces between it, it will continue that going across. So we had to do this going up this side, across the top, down the other side, across the bottom, and make that spiral effect going around. And that was what we did just to start, just to get the concept to show them, hey, here's what the spiral pattern will start to look like.

DESIREE MACKEY: So, you want to pull it out of the oven?

BRIAN MACKEY: Yeah.

DESIREE MACKEY: All right, so what happened, what was interesting with this, is that we built up all of these. So there's all the angles that-- those are just reporting parameters so they can understand what the angles were.

BRIAN MACKEY: She's talking about these dimensions here.

DESIREE MACKEY: Yeah.

BRIAN MACKEY: So then the client said, well, we have a logo that has a whole bunch of angles. What angle are you so we can compare that to the logo?

DESIREE MACKEY: Yeah, we're not joking.

BRIAN MACKEY: We're not joking. That was real.

DESIREE MACKEY: In fact, if you oriented it just right, you might be able to guess who the client is, which I think I've removed from everything. So if you accidentally see it, pretend you don't know it.

BRIAN MACKEY: And you guys from that client are listening to this recording in a year, we had your permission to show this.

DESIREE MACKEY: I did. I did ask for permission. So what was interesting about this, we built in all that parametric behavior so we could go from-- we ended up calling it Bays, 13, 16, 17, 18, 19 bays for the different lengths of the bridge. But we found out-- and it's a really strange behavior. And I started asking other people that have done a lot of repeaters if they've noticed this behavior-- and we don't know if it's something that's repeatable, but it was for us-- is that if you change the number of divisions so that you want to add-- so you go from 13 to 16 bays-- the division is centered along the repeater path.

So I want to add one complete revolution to the end. So instead what it was doing, if I added one complete revolution, was adding a half to the beginning and a half to the end, which was annoying. So what this ended up being was three separate families, that, of course, they couldn't decide. So I think I have 10. I have everything from 13 to 20. So there's several. So that was complicated. So this is the final family here.

BRIAN MACKEY: No, this isn't final.

DESIREE MACKEY: No, it's not?

BRIAN MACKEY: So the other thing about this too that we had, that I ended up actually changing and building into these families-- and back to the family and go to it-- is when we put it in there, we realize, oh, we've got to do the inner spiral and the outer spiral. When I had this set up, they also were then going to want to start having some additional offsets.

So that's not that one I wanted. So back in the original family, I went and tweaked it even more so because on that rectangular two-point family, I showed you guys how I have basically the left, right, center checkboxes. But we had to go in there and then also start adding in an offset. Because in some of those cases we wanted to-- we'll make it bigger than that.

DESIREE MACKEY: So now it looks exactly like structural framing in the family, right? You have the z offset and the y offset.

BRIAN MACKEY: Yeah, you get your z offset. So I also started assisting in that as well. So when I told you guys literally, I had this point and then I hosted this point onyo it, I then had to redo it all, rehost another point so it would get shifted off to the side. So I could shift that with a positive or negative number because we really needed to come in there and start playing with the original family that we had.

We were sitting there thinking that, oh, OK. So we'll just have this one set of repeaters. And everything will be on it. The final family ended up having additional lines, to host the inside members, two of the outside members, two of the floor members, two of the roof members two of the et cetera. So it wasn't so simple, but we had this set up just to be able to start coming through--

DESIREE MACKEY: So in the bookkeeping part of it, that was difficult. Because you'll see they have the exact same number of divisions along each of the paths because they had to line up. And so that's what became a bookkeeping nightmare is that this first member went from 0.1 to 0.2. And then this one is 0.5 to 0.6 or whatever it is. But so do you want to switch back and I'll show them the profile?

BRIAN MACKEY: Yep.

DESIREE MACKEY: So this is the finished section of the bridge. So this, you can see the gaps in here. So this is what eventually caused us to have eight divided paths instead of four, and then those offsets to kind of-- all of these members there are hosted off that inside point so that the insertion points-- and then this guy, I think this point we have for the inner helix, we call it. So then there's floors, and the metal deck, and such. Now it looks like a structural element.

And it wasn't green. It's just that so you can see the outside and the inside. OK, now the castings. So the way they wanted to fabricate this was to have each of those-- I'm going to go back there-- each of these corners. So as-- oh shoot-- each of these corners as it comes around this corner. But think about the way that looks with the bridge. So the way the corner comes across, and it's at an angle at the top, and then the side is at an angle there. So you have a compounding goal. Is that the word I'm looking for?

It's angled in more than one direction. So this is actually a quite a complicated geometry. And so they were going to make this out of castings. And symmetry was an important thing. We wanted to have as few templates as possible. So we spent a lot of time. It's an amazing amount of engineers standing at a table trying to visualize exactly what this was going to look like and how many unique cases we had. And we got it down to two. So at the bottom where the bottom angled element was perpendicular, and then the top, where the top was angled.

Because if you rotated around, like kids toys where the spatial awareness, if you rotated around, the angles end up being symmetric. So then we thought we would model these so we could see it in the project itself. But then we got a little bit-- I'll use the geeky term again. And we thought we could 3D print this. So we went above and beyond to create this.

And part of the impetus to wanting to do that was that we were having a really hard time communicating how difficult this geometry was to our client. So they didn't quite understand that, as you come from an angle to an angle with two elements that are exactly the same size, this little part here is going to get bigger. You can't just turn that corner and keep the same size and everything. Is not a tube that you just bend.

So you have these tapered, bigger edges. And they had all these sorts of requirements on how they-- if you're looking on it, they didn't want to see the element get fatter. So it was difficult. So I spent all weekend one weekend. And it was pretty much brute force, just massing these things out and getting the angles, and going back.

And I'm like, OK, this is what they're going to want, this to [INAUDIBLE] it wasn't what they wanted. I think we rebuilt these three or four times. And that's when I'm going to 3D print it and I'm going to bring it to AU. And so then when we decided to do that, we needed to make it hollow because we didn't want a solid piece because currently, those pieces are solid. And then we had these little filleted corners.

And I was having a horrible time getting the final sort of aesthetic with the rounded corners and everything. And so Brian-- brilliant idea. He's like--

BRIAN MACKEY: Don't judge me. I'm just saying before she makes the statement, do not judge him.

DESIREE MACKEY: He did it in AutoCAD. So--

BRIAN MACKEY: Well, to clarify, I used the Inventor ad-ins for AutoCAD because I do not have Inventor. But--

DESIREE MACKEY: Yeah, so he finished this in AutoCAD. And then we ended up not 3D printing it, so I don't have one to show you. I'm sorry, that would've been awesome. It came in way over budget. Surprise, surprise. So they brought a fabricator on really early. And they actually took all of our stuff and all of our families and used it. So I felt good about that. But they took over this situation from us. So here's the Revit part.

BRIAN MACKEY: So here is that Revit joint. We actually ended up having 30 different variations of this because they kept saying, oh, no. We don't like that. So that was the other thing that the offset parameter started adding to that angle, is even though we were all on the same nodes, they wanted to play with the way this would mold together. So we ended up having to tweak where those point locations were, which is why we ended up having to have an additional repeater on the inside of the outside repeater for the way the bottoms and stuff were coming in and the offset values.

But this one really isn't intelligent. There's not any parameters to this. What really was difficult on this is trying to get, like we said, everybody to understand this. So we just ended up basically coming in here with a whole bunch of rectangles and saying, create forms, and this rectangle, create form, et cetera, to get this. We really, really, really did try to fill out these corners. Revit would not do it.

We couldn't do it with voids. We couldn't build it into the curve's Revit. It just simply would not do this. Trust me, I spent probably way too many free time for her company trying to get this to happen to get it to work.

DESIREE MACKEY: Well, we could do it pretty easily at these places, where the profile is constant, and back here. But in this tapered-- which that was huge amounts of conversation, how far should it taper from this kind of thick knucklely looking [INAUDIBLE] to back here-- we had a really hard time here. And then don't even talk about that corner. How do you-- I don't even know.

So we did actually take this and export this out to DWG. And if anybody's never actually been an AutoCAD, there's the great 3D modeling tool that they Stole from Inventor. So when I was back in AutoCAD, I didn't open that because it was pretty brute force because I don't remember AutoCAD. But when we did do that, we went through and basically hit the Fillet command in AutoCAD. It did it seamlessly. It did it really well.

And AutoCAD also has a Hollow command, which this one is not hollowed. I didn't need it in Revit, but we did that for the 3D printer. So AutoCAD, we used the Inventor Fillet tools. We used the Inventor Hollow it out, so the whole thing was hollow. And we did send that to the fabricator for them to be looking at for 3D printing.

But I did import the DWG back in and exploded it. Yes, I exploded it. But I did up with basically just this direct shape type object in there. And, again, this was more for just to put into Revit to give them the graphics of this is what you are going to need. Because the square, uncompound, uncut things was freaking everybody out.

DESIREE MACKEY: Yes. So we weren't getting those nice, smooth corners. So we actually put these pieces inside the project, and we just popped them right on top of the bridge.

BRIAN MACKEY: In no fancy way.

DESIREE MACKEY: Yeah.

BRIAN MACKEY: Plot a line, a line, a line, a line. They weren't adaptive. They were just, hey, they came from AutoCAD. There you go.

DESIREE MACKEY: OK. So then now we have our pieces. And then came the structural part. So I was thinking that all we really needed was coordinates because we can set up the analytical model to know which members go from which nodes to which nodes.

So I just needed the coordinates. So I thought, well, I'll just grab all the generic modeled elements and get their endpoints, because I'm thinking it's a structural framing element, right? So what I found out very quickly is that when in Dynamo-- no, I don't pretend to be a Dynamo expert. I know just enough to do a few things like this. So I'm sure you can learn a lot more from Carl.

But so what I found is I was having a really hard time dissolving that generic model that had all these profiles, especially the bells and whistles that we had built into our profiles, into one simple curve with endpoints. I was getting multiple curves. So think about every edge on that all the way around the profile on both sides and all the way along centerline. So it was not working for me.

So what I ended up doing is I made a little, tiny-- I called it a node. And I put it on the repeater at each node. And then I was in control over which coordinates I actually cared about, and I didn't have any duplicates. So that's what I ended up doing, is created this little-- it looks like a little ball-- put it in the family, and I was done with it. And then I had a relatively simple Dynamo script-- this was the entire thing-- to pull out the coordinates of those nodes, pull it into Excel, and dump it into our analytical, which was a very simple process once we got there.

So I think, Brian, are you going to run it? And while he's running this, get your QR scanners ready. But we found out today--

BRIAN MACKEY: Literally, at lunch.

DESIREE MACKEY: Yeah, literally at lunch, if you have an iPhone, your camera is a QR sanner. We didn't know that. Did you guys know that?

BRIAN MACKEY: In iOS, your camera is now a QR code reader. I did not know that. That is a new feature of iOS. I went with the new iOS. So yeah. So this is Desi's highly sophisticated adaptive component family.

DESIREE MACKEY: It's a circle, not a square.

BRIAN MACKEY: It is adaptive. So if I go in X-ray, you can see inside of here. There is an adaptive point in there. So it's an adaptive point. She then drew a line and said, hey, create a solid, so very highly complicated thing. But what ended up happening is--

DESIREE MACKEY: No judgement though.

BRIAN MACKEY: I don't even think you parametrize it. It's just a circle. See, I couldn't not parametrize that. But so then she just put in there and used the Repeater command, just like normal, and basically repeated it where she needed to. But, again, it was the same scenario. Put it where they're intersecting. So one was a little bit of a pain in the booty. Just put it in, and make sure you're on the right one, and set that up.

And then with that said, once she had those inside of there, there was a Dynamo script that was written that says, hey, you're looking for that family and we're just going to go ahead and run this, and export this out to Excel.

DESIREE MACKEY: Yeah, super simple.

BRIAN MACKEY: So then she got an XYZ coordinate of every single one of those nodes.

DESIREE MACKEY: So and this is not as sexy of a spreadsheet as the other one that we'll see later because this is all the analytical software I wanted. We just wanted something we can copy and paste in, and doesn't need to be pretty. But so that's how we did it. And I've done this. We created this workflow to get these simple points out.

And I've now done this on a lot of projects now. I'll show you a picture of another one at the end, and a lot of different iterations of it with points, or endpoints, or columns, or whatever. It's turning out to be pretty handy. So I'll take the time to make a little Dynamo script to pull some coordinates out. And then the QR? So I'll make sure I'm out of the way.

So if you scan this, you can look at what the bridge looks like. One is just the one family, and one's the whole project. And we sent this to the client. Well, I just thought it was going to be-- I thought I was going to be cool. I'm a structural engineer sending them Enscape stuff. They're going to think this is so awesome. They're going to love us. They were not impressed.

I was like, I emailed you a QR code, and you could see the bridge. Isn't that awesome? And they're like, yeah, we didn't really look at it.

DESIREE MACKEY: Yeah, so the one to the left is directly taken from the project, the family. And the one on the right was taken from after the project was done quite a few months later. And they put the whole bridge together. That one actually came from the bridge assembly.

DESIREE MACKEY: Yeah, so try not to hit your neighbors as you're looking up at the bridge.

BRIAN MACKEY: Yeah, especially the first one, because I just did it. I wasn't even thinking about it. Under the bridge, it's like, they wanted to see it from a person's perspective. So--

DESIREE MACKEY: Well, and that's what was really helpful with this. I don't do a ton of this structurally because it's not generally something that's super helpful. But this was because they were really concerned about how the structure looked from specific viewpoints, so underneath being one. I'll wait until I just don't see cameras in the air anymore so you can scan this and look.

DESIREE MACKEY: So, yeah, and this was done using Enscape. I don't know if anybody knows. I've been preaching Enscape I think since they were in beta. I've been using it for a few years now. I like Enscape. I like the ease of Enscape. And this was shortly after they came out. The family one was shortly after they came out with the save the panoramas.

But we were walking around it, through it, et cetera with the file. So and then I think most of the camera's down now. So and then this is what their final bridge project looked like. So you can see these are the adaptive component families, which really aren't adaptive anymore. They're basically just, hey, we've got--

DESIREE MACKEY: They're pretty stagnant.

BRIAN MACKEY: --inside of there. And then they just use standard Revit structural modeling tools for the stairs, the structural columns, and everything else, et cetera. But that is the crazy custom bridge for what the client really wanted to look like. And so that QR code on the right, I threw a floor in here. And that was this model, just up in Enscape.

So I think the grass comes out gray or something. We literally didn't spend that much time on it. But it was just to let people know, hey, here you go. We believe we also gave them an executable so they could walk around it too.

DESIREE MACKEY: Oh, yeah. That's right. We did. That's probably why they weren't impressed. I was like, I don't really care what color the ground is. I'm looking at the structure. There's no trees. There's no people.

BRIAN MACKEY: There's no road.

DESIREE MACKEY: Yeah, OK. [INAUDIBLE]. Everybody good? Got it? OK. All right, let's talk about the roof, which is a little bit shorter because it's a lessons learned from the bridge onto the roof, and it's a little bit less complicated, which is surprising since it looks like that. That looks complicated. But it does repeat. So are you looking for that blog?

BRIAN MACKEY: No, I have it.

DESIREE MACKEY: OK, so back when-- gosh, I don't know when it was-- 2012, 2013, somewhere in there. I was a while ago. I don't remember exactly what came out, but we were playing with curtain panel by pattern families applied to structure to create space frame, really fun stuff. And we wrote a blog on it a few years back.

And then when this came through-- I think this came from my model. But what the client gave me was this DWG that had a surface look to it, and it had some lines in there like this, like, oh, could the structure look like this? Because, again, the structure had an aesthetic requirement. So when that's the case, you have to be really concerned about how, if it's actually going to be feasible. So thank goodness they asked.

But he started to do this, and he wanted to know if it was feasible. And it made me think of-- probably because it's so repetitive of the space frame thing that we messed around with a few years ago-- so I came home and I was like, do you think this would work? And so it did. So this is what we did.