Sketch to feasability via HFDM

MARC DURAND: OK. So I guess that it's going to be time to start. First of all, let me thank everybody for coming. I know it's late. It's certainly the last of the day. But I think it's worth, you know, it's an interesting topic. So thank you for coming.

The first part, so who I am. My name is Marc. And I have been working for Atkins, part of the SNC group, for the past two years. I'm based in Dubai, so in Middle East, where we are working on large scale projects and a really fast-paced project. And I'm what they call the digital disruptor director. So, basically, they call me crazy most of the time. And we bring crazy ideas to try to evolve and bring what the future could look like, I think it's a nice way to put it. And then the other difficulty is to convince people to make it happen.

So the main theme of today, for people who have seen the opening keynote on mainstage, we presented a project called caterpillar. And today, [? I thought ?] about it. What can I explain to you what it is exactly and the type of application that we can create with it. OK? And I'm really open to any questions. If you have any specific questions, don't hesitate to interact. I think it's more important to explain and show passion here. And that's it.

Then, Mustafa will join me and take over explaining to you a little bit what's under the hood, because it's important to understand the technology underneath it. And then we'll conclude a bit about some integration that we've done now with some of our colleagues of the Revit IO team. So I think it's going to be interesting. So what it is about. First of all, at least my part of the program was about, how do we translate and do we fix one of the biggest problems we have to date, which is mimicking what we do best as a sketching application?

And no matter where I go, in any offices I have been all over the world, I still see people sketching on paper with a pen. I don't know if it's a natural feeling, or if it's people who feel more comfortable about it. But everything starts with a sketch. And to paraphrase words that I said in the keynote is, we try to bring this napkin sketch concept into a brand new digital way, which potentially will bypass the learning curve of software. Because the idea is, use technology to help you do what you do best. And that was kind of the puzzle that we were trying to solve with some of my colleagues.

So, again-- and I hope that everybody in this room understand the concept of sketching. Now, what do we want to do out of this sketch? It's obviously what everybody else want to do, which is a 3D model. Smart or stupid, we don't really care about it for today. But the idea for us was, OK, how do we take a sketch idea and translate it into a mass for feasibility purposes, and link all of this mass to all type of analytics that we could dream of. So cost, program, anything we want.

But, again, bear with me for a second. The idea is to be able to be as active as possible and be as smart as possible. So when you look at the 3D model like that, what's your first view of it? It's dumb. We can't do, really, much with that. It's kind of a pretty picture, like we would call it. OK? And the idea is, how do we turn this pretty picture into something that your client will want to see? So the third part of the question is, what is the client's requirement?

And that's where we'll jump to the next slide. This is what the people that we are facing on the database wants to see. What's my return on my investment? What's the money that I will make on the project? So when you start to think about it, they are not always interested about what it looks like, but more about how much they will make out of it. OK?

And, again, when you start to think about, how do I connect these three parts together, it's basically, a client comes to me because he has a project that he wants to do. And then we try to design and sketch an idea and translate it. And you know this linear workflow that Andrew was mentioning earlier? This waterfall workflow, like you mentioned it. Which is, I take an idea, I make a sketch, I make a mass, I make a program. I make my Excel sheet to see if everything fits with what my client's requirement is. And then you realize that it's not working.

So then what do you do? You go back to square one. And then, when you meet the requirements, whatever, and you go to your client with this-- I'll go back one slide-- this pretty picture thing saying, tah-dah, that's your master plan. And guess what, the client is like, I don't like it. Go back to square one. So the goal with today's technology was more about, how can I interact with my client without having to do anything? How could I sketch? And all of the other things that I've just mentioned can come for free, basically.

But since we like to make it even more complex is how my client could say, just imagine, bear with me, is going, I don't know I'll take one [? edge, ?] takes this progress bar and goes down saying, I want less commercial. That would be nice if the sketch evolved and the mass evolved by itself. Because you have this bi-directional communication, that you see the data being an entity and you interface it in different ways.

So my sketch is one view of the information. The 3D massing is another view of the information based on the 2-D sketch, obviously. And then all of the analytics comes out together. To solve this problem, we developed, internally, a code name that we call because we haven't come up with an application name for the moment. But we call the code name Caterpillar. And as you can see, it's a full web-based application, which is kind of cool because you don't need any software to run it. You just need a web browser.

Which means, as we mentioned a lot of times, no need for specific hardware. No need for specific software. And even better, it works really well on tablet because we believe in flexibility of designing anywhere we want. But we went beyond that. We even thought of, how can we bring community together? So we started to think on how can I bring 10 users, 20 users, 30 users with the power of the cloud to interact together? So, basically, we started to have this crazy idea saying that would be nice because the company we work for, we have about 50,000 people globally.

So we don't work together more by constraint rather than by desire. And that would be nice to, for one, to sit on his sofa because it's 9:00 in the evening, where the others start to work with the team and start to be able to scribble all together to share experience and share ideas and for everything to update in real time. So, as you can see on the screen right now, we speak about, obviously, desktop capabilities. But, mainly, we will focus, today, on the iPad, or the tablet, actually, I should say, not to give any brand. But we like Apple. So we use the iPad Pro, really well optimized for it, because it's a good feeling.

Now, let me kind of break down a little bit the text below it. And, again, guys, it's really important to understand this part. What we've done is, we've created, basically, an application, let's call it a sketch application, which you can see on the top of this diagram. And then we linked it to the Forge service specific one. So we have been exposed to it in the early the early stage. And that was called, back in time, at least, HFDM, which means high frequency data management.

And that's a little bit a good way to represent what HFDM does. Because, as I explained earlier, it's kind of, you plug your application to microservices. And when you do something, it triggers an event. And the event pushes back the information everywhere ours is connected onto. So imagine, if we want to make it in a simple way, it's a really good technology where you could have a Revit model online. So a Revit IO, for example, is a good example. And that you would connect that, almost in real time, in the Google spreadsheet, for example. So when you start to change and interact with one, everything would update in the others. That's, I think, an easy way to explain it.

Another way that we have exposed that couple of time, as well, it's a little bit of a mix between a Git repository concept and Google Doc kind of information. But, again, Mustafa will explain to you a bit more the technical behind it, because he likes it and that's his part. Then, the last part, and that's super important to me is, everything is web-based. I can not speak, and explain, and repeat myself enough explaining that we have to run away from software. In a sense that, we have to be more creative.

We have to be able to edit on the fly. We have to be able to work together. We have to not worry about the type of OS we are even using, or the type of software. Or even if hardware. Because, obviously, the creative spirit, they are more Apple-driven, based on my experience. And we have conflict of OS quite often. So everything is web-based. And you start to think about it, OK, I have an application that I can interact with from anywhere. And the idea is to plug this application to third parties.

And I'm sure that in the room-- which I haven't asked yet. Let me just ask quickly the type of people we have in the room. So who is more on the design side? We have some designers, nice. And who is more on the development and programming side? And what are the others, then? Do I have any engineers? Ah, nice. I'm an engineer too, by the way, guys. OK. So that's cool. So it resonates with you what I am trying to explain, I hope.

So, again, web-based to have the biggest reach as possible. So when I do something, it's shared in real time with the mass. And, again, speaking about design, our experience is, we start to have smaller teams that are clients are facing. And we start to have back office for productivity in order, what we call a GDC, which is our Global Design Center. So it's really important for us to be able to send, almost in real time, to the back office, for them to be able to produce and be productive really efficiently.

And what we also try to do is get, or remove, suppress, would be the best word, the delay that we have to bundle, send somewhere, wait for it to be done, download back and then present to client. Because this could be days between the point [? we're at. ?] And, trust me, we are [? basing ?] [INAUDIBLE] least, as I mentioned earlier-- so our week is from Sunday to Thursday. So we are already, I think, 12 hours difference from here. Then, we have full day off. Basically, [? if my ?] end of week, I wait three days before you send me update, which is half the week. So we really need to find a way to change that.

Now, let me jump into a little bit of the application. So, first of all, because, as I said, I like interactive talk with people. So do you have any questions so far? Or is it clear? Good? Perfect. So let me jump in a bit about the application to explain to you how we made it. And guess what? We made it extremely simple. Because my goal is to be able to have the, let's say, experienced person, which, by default, is maybe less technical, to be able to utilize this kind of application sitting with his client in real time.

So a good scenario for you to think and forecast how this app could be best used is, you come as a designer with the iPad, sit next to your client, tell him, wait, open your TV and log on to the website that you are seeing on the screen right now. And, basically, everything you do on your tablet will generate a trigger and will update in real time on this TV. Except that you sketch in 2-D and you will see that everything comes in 3D on a screen, depending on the view that you want to work with. OK?

So, here, what do we have? Simple. Home page. So, obviously, like every application, you have to subscribe to the home page, which has, at the moment, I'll be honest with you guys, we kept that for internal purposes. So you can't really go there. But in the short future, I think that we will open it to the public. You go, you create your own project. And what happens when you create your own project, because we try to do things smart but easy, is project setup.

So what's the project setup? Project setup is, it opened, we connected that with the ESRI ecosystem. As you can imagine, I think a good Intel probability with ESRI, now, and Autodesk, we decided to go with them. So what happened on your big iPad? You have the ESRI map that comes, and you use the search engine to be able to go to the destination you want. OK? So, again, why did we do that? If people are working on master plan, it will, again, resonate with them.

But our [? fund, ?] we start project making a screenshot of a Google map, and then start to sketch on it. And guess what? When we sketch on it, you don't even know the scale. So when you do a sketch, you don't know your area. So how do you manage program? By, again, going in this waterfall effect of waste of time. I'm sorry to say, but it's still true. And then, people are doing smart things, but the old way. So the other things that we wanted to also make sure is, when I do my sketch, wouldn't it be nice that, not only I sketch in real space, but my GFA is calculated in real time.

And also, wouldn't it be nice that, if I do save as Revit, it comes in the real space. So the GPS coordinates are translated and transferred as well. Guess what? That's what we did. So you go there, this-- to, again, echo the presentation we've done on the London Olympic, I put the London screenshot-- you choose the type of map you want to see, which is, in ESRI, out of the book capability.

And then, you press a button that basically prints a screenshot of your mass in real scale, takes and make a ratio to translate and transfer the coordinate, or the spatial coordinate, of your plan. And it goes on the back burner into a database. Again, I will not be speaking too much about the data structure. But if you're interested, we can definitively explain to you how it works.

So that's for the project setup. When the project setup is done, then, on your home page, you will have two buttons that will appear. One is for what we call the sketcher, and the other one is for the viewer. I'm going to start with the sketcher. Again, what do we need to do? To speak to the designer of the room. Where do we sketch generally? If we leave the paper, we're going in Design or Illustrator. Correct? And we do with that is morphographic purposes, or for reporting purposes. Because what we do is generally vectorial.

So, for the more, let's say, geeky person, you will know that we can work with the format colors SVG. That is actually pretty cool. And what we've done is we connected this technology of your sketch is actually converted, more or less, in an SVG. So everything you sketch on the app, you can have access to it at any point of time, and you can exploit it for report. But the good news, and this is, again, technology that is evolving is, we will, in the short future, be able to connect that to third party applications, and look at the guy who developed the HIVN.

And wouldn't it be nice that, the day you open your Illustrator file, and you change your color [INAUDIBLE] updating the sketch, that's the type of logic we have to have. OK? So you go into your sketcher. And what the sketcher does-- so that's where the smart part is. The sketcher allows you to sketch on what we call a layer base. And I hope you agree with me, it's a bit of Photoshop-ish type of logic to keep everybody used to a certain logic.

So what you do is, you create a layer. And on each layer, of course, you can change color, because we have even went down this path of changing color too. But the idea is, you assign recipes. OK? So you will hear me speaking about recipes a lot. And recipes, it's almost like a script that will be launched every time you do an event.

Let me make it clear and simple. If I do a shape on a specific layer, the layer will take the shape and potentially extrude it from a specific height, assign some specific information as well, like the type of asset that you want to do, or that could be a document, that could be anything that can already be connected to a certain feasibility, that could be connected to formulas, that could be connected to a script, that could be connected to Dynamo script, which you will see after, that could be connected to anything you think of.

So you sketch. And as soon as you sketch, an element, literally, is created, and we connect to it. OK? So that's where, for the time being, what we're doing is, we allow the user to select predefined recipes. But we can generate them on the fly. And we have the capacity, as well, to generate some kind of coding experience directly into the app. But we give access to that only for the, let's say, advance user, in a sense.

So, here, what you can see is, again, opacity, blending, stroke, type of color, the stroke of the color, and the important part, for us, is the building type. OK? So bear with me for a second. I have created a couple of layers. Some are residential, some commercial, some mixed use, whatever you want. OK? And these, after you start to select the layer you want, and you sketch on it. OK? So far so good.

Then, let's be quick here about the tools. So the tools, since we wanted to be as mobile as possible, we designed that for an iPad, as I explained to you earlier, or a tablet, sorry. So what we did is, we did the radial kind of menu. So you sketch with your hand and you press with your thumb. The menu pops up, and then you have access to different features. So guess what type of features we've put? So all the Boolean extraction that you can imagine, union subtraction, any type of Boolean ways.

Then, obviously, you have all of your scale, you have all of the basic. I'm not going to list them all, that's not the purpose. But you can move, you can rotate, you can scale, you can copy, you can cut, you can delete. It you can do whatever you need. And this comes under three actions that were grouped differently. Because we started to look at, by working with people, that I think a big menu is not really practical. But having three menus-- one to kind of edit, one to design, and one to interact with-- was well done.

And just to let you know, at the moment, we support pretty much any type of shapes. So we go from square to circle to ellipse, but also to free end and to the pen tool as well, to be able to change the [? handle ?] of your sketch. So, on the sketch part, as we can see, when you start to scribble a couple of sketches, all of them, they were done onto the application. What happened next is, you go into the 3D Viewer. So I'll play some video after on, guys, don't worry.

This is the 3D Viewer. Again, same interface. It's better in landscape mode, so you just have to turn the device. It's not very complicated. And then you have the under layer that comes. And all of the mass that we've created earlier, they just extrude, basically. OK? And what becomes interesting as well is, you have the same mimic of all of your layer base strategy. So if you create a layer in the 3D viewer, or in the 2D catcher, the layer of data is automatically in the 3D Viewer. And you can stop to interact.

So, again, what becomes super interesting is the logic and interaction you have with your team. So imagine that I'm looking at sketching a little footprint of a building, but I don't know yet what type of building I want to do. What is cool is, at any point of time for my calculation, in the 3D Viewer, I can start to select this mass, change its layer and all of the [? feasibility ?] recalculates. Your sketch [? re-change, ?] update, so the color will change, basically. And that's the same entity you play back and forth.

So what I love about this concept is, you know when we look at teams, you have the thinker, you have more as the doers, and you have the one who's running after the others. The idea for us is really to be able to put everybody at the center to contribute property by removing the [? heavy ?] [? lift ?] of, I have to update my PowerPoint, I have to update my [? end ?] [? design, ?] I have to update my report. You change the color of your mass, now you have to update my Excel spreadsheet for my client. It's just going nowhere.

So, now, let me show you a little bit how it works, which hopefully the video will trigger. I don't know, sometimes I'm having difficulties with that, and today is one of these days. So let me just do one thing, guys. And I'm sorry for the format of it. But just to avoid wasting time. So, here, before I play it. This, bear with me, one screen another screen. It's in the real world to make it more interesting. One guy in the US and one guy in Australia working on a different device, and that's how it looks.

I will explain to you a bit, because we didn't do any voiceover. Obviously, for, how can I say that, confidentiality purposes, we took a fake project. Because we don't own the patent. Now, for people we know don't know this project, in Dubai, we love to create island in the middle of the sea. So we did a fake example for the EU. So as you can see here, I am opening the layer base on the right hand side. And again, I will put some post to really make sure everybody understands.

On the right hand side, which is one, could be many, and believe me well a lot of dashboards, so I can show a lot of things. But for the purpose of today, Mustafa did a quick interface to facilitate. So here, we could look at the mass with the GFA, which is a formula that takes the footprint of the sketch multiplied by how many flow, which is what dictates the height of the building.

So here we are. Now you will see how complicated it is. You first select your layer. So, obviously, you can zoom into it and do whatever you want. Select your option on this radial menu, move on your sketch. And here, we are going to simply take a block. So we are blocking. I don't know if people are familiar, again, with the master plan logic. But you first block, and then you subdivide, and then you go in detail. And as you can see in the 3D Viewer, it updates, literally, in real time. And that's on any application.

And I don't know if you've paid attention, but even if the pie chart re-updated, so this pie chart could be a total third screen. So you could literally have this big dashboard [? feasibility ?] coming in real time. You could have 20 screens, if you want, with different views. Because don't think, what is important for you to understand? And I know, sometimes, it's kind of not obvious. But if a guy in another computer turns on the 3D view, it's not going to turn on this one. It's going to turn on this one. OK?

But if you move a building, then it will move everywhere. So what becomes interesting is the smartness you put in the system and what you connect together. Let me reiterate that. When I turn my screen, you all agree that, actually, I turn the point of view, which could be a camera. So if I don't connect the camera of the user to the HFDM system, then it doesn't update elsewhere. So you have to think about, OK, how do I make an app? What do I want interactive? And what do I want the other to do?

Because one of the things that was really funny, but not funny, actually, is, one delete what the other created because it doesn't like it. It's a little bit like the early stage of Revit. I don't know if you've seen that. The girl makes a [INAUDIBLE] and the other guy is deleting because he didn't do it, that's the same concept. So we have to be careful about that. And that walks in a more collaborative way, which means, as well, people need to understand what it means.

So just to show you that it really works, we did a couple of quick mass in how we can edit. So, here, we wanted to kind of shape quickly a little bit of the head of the island and pay attention on the other screen. Because I think that's where it's interesting to see that, when you do something somewhere, it really re-appears elsewhere. So, now, depending on the tools you have, and depending on the things we have, we can interconnect them together. And, as I said earlier, we have access.

So it's really, really mimicking the walk through that we do in Illustrator. Because the feedback we had with people is, give me a tool that I know how to use. So we started to understand, what do you know to use? And most of the people I know in design and architecture, they like Photoshop, they like Illustrator, they like this kind of logic to make things pretty. And, again, as you can see here, as I said earlier, you have access to your handle. You have access to your notes. You can add point, you can remove point, you can create whatever you want. It's a pretty straightforward walkthrough actually. And as I say, it works pretty well on your mobile device.

So imagine that, at the minute, we see only one screen. But exactly when I'm playing around with this island, someone else could be walking on his own device on this part of the island. So when we start to think about teams saying, hey, we have two, three, four members, and we want to sub-devise by zone, we can do it. But, even better, as a client, because I like to do that, I could start saying, hey, I have multiple contractors, or multiple designers. So imagine. And they come in Atkins and wherever you want. They could all work on a piece of the project and then you get the feedback in real time.

So everybody was linked to his own part with [? these ?] analytics. But the client will have a global analytics, or an aggregated version of the presentation. Just careful of the time, Mustafa. OK, so I will give it to you. OK, guys. So this is kind of a quick snippet of what it looks like. And, again, we have to keep it condensed in a way that we can present it under less than an hour.

Now let me now jump into how it works, and we will explain. So, as I said earlier, how it works. This is a slide that I got from the folks in Autodesk a long time ago. And I like it, from the beginning. Really, if you are used to the Google Docs kind of ecosystem, where we can have multiple users working together at the same time. And if you understand what git is about, which is, this kind of a folk concept, then you understand how HFDM works.

And, by the way, guys, are you familiar with what git is about? Because I'm not going to give a lecture about that. But just bear with me. You do one action, which is this blue dot. Then you do another action. So imagine I create the mass, then I rotate it. And then, I move it, or I scale it up. So I'm at point number three. But then you say, no, no, I don't want to rotate it. So you go backward one time and you say, I will do something else with it.

What is beautiful is all of this gets saved somehow, and you can have access to it back at any point of time. So it's almost like you can have a slider showing what you've done from the beginning of time until where you finish, basically. And this is a quick diagram expanding what the caterpillar app is about. Which, again, we have a 3D viewer, which I hope you would understand how it works now. We have a 2D sketcher. We have this map service that we call the cropper, internally. And then all of them are connected to these data HFDM with a microservice that we consume.

So, again, because we are in [? use ?] so it's important for people to understand how this works, that's where you consume your Forge token. That's exactly here. When I do something, something happened, and that's where, basically, you consume the service. So that is the high view. Now I'm going to ask my colleague Mustafa to come on stage and explain to you really how it works, or the technicality behind it. Thank you, Mustafa.

MUSTAFA SALLAHELDIN: Welcome, everyone. My name is Mustafa Sallaheldin. I am the data science manager in Atkins, Middle East. And I am a multidisciplinary subject matter expert in the AEC industry and software development. And my speciality is building automation systems for BIM by developing Autodesk products.

And we will start talking about the technical part of HFDM. So let's start with the difference between the low frequency data management and the high frequency data management. And for the LFDM, or the low frequency data management, when we make a change in the design, we need to re-upload the whole file again every time we make any change. But for the high frequency data management, we only send the incremental changes every time we make any updates.

So what is HFDM? HFDM has several features. Some of them allow secure and efficient Cloud datastore. So as it can communicate and store high frequency changes, as well as it creates full and fine grained history. In addition, it is compatible with a GDPR and SOC. The second feature is, it allows the collaboratively data changes. So it is like git for allowing the branching and merging. And it is like Google Docs, where it allows the real time updates. Also, you can work offline, make all your changes, and then you can synchronize when you are online.

The third feature is, it allows the granular access control, where each user owns certain parts of the data, and clients can subscribe [? to ?] parts of the data, and owners can restrict the access to the data. So why choose HFDM? Because it allows me to quickly update my data by sending only the granular changes, instead of sending the whole files over and over again. Also, it allows the collaborative design so many users can work on the same design at the same time.

And HFDM gives us the ability to log their changes. So at any point of time, we can go back and forth to change a certain point of the design time. Also, HFDM gives us more freedom to customize our workflows to link multiple design application or get application together. So how does HFDM work? As we can see from the graph, HFDM does branching and merging in two steps. The first step is design exploration by creating multiple branches. And the second step is a synchronous collaboration by merging changes coming from two or more branches.

From this diagram, also, we can see how HFDM delivers the real time collaborative by multiple collaborators manipulating the branching simultaneously and continuous merging for the global and local branches. In this graph, we see how HFDM gives the granular access control, where a client can choose which data to fetch. And the user owns certain parts of the data and the owner can control user access.

So I will hand over again to Mark to talk about the integration between Revit IO and HFDM.

MARC DURAND: So any questions on that, guys? Because I know it's, sometimes, heavy to digest. But that's clear for everybody. And I think it's really important to sometimes understand how all of this tech works together. Because, as I said, when we speak about the technology itself, people don't always forecast how the future looks like/ but what we're trying to tell you is, we're starting to venture into an era where we don't speak about software anymore. And we speak about, when I do one thing, these things are representing and ripple everywhere else. So you start to really be able to work together.

And one of the examples that we wanted to explain to you really quickly, and we ask if we could share this presentation, it's a couple of slides of real cases that we've been doing. So as you can imagine, we still have to have a technical output. When I speak about technical output, I speak obviously about 2D technical drawings. OK? So let's call the technical drawings a PDF, for instance.

Now, we do, and I hope all of you in this room does, generate our PDF through. And what we wanted to understand is while that's super cool, you do this mass, it's cool, it's on the website, it looks cool, the client can see it. But then, that would be silly of us to say I'll go back from scratch and rebuild everything from scratch. We would agree on that.

So what we started to say is, and I hope we all understand, that the shape is actually a series of coordinated points. As I told you, we've connected it to the GIS system so we know where the point sits in space. And I hope some of you will recognize this interface, which is called Dynamo. And Dynamo allows you to import data sets. Imagine, again, I go back in time on CSV format-- imagine just for a minute. But there is some integration that are coming to be able to connect that. So let's say that, coming next, we don't need CSV anymore.

Anyway, so you see that's why sometimes we speak about almost automated, but this part of the presentation is more what could come next. OK? Now, the beautiful part is, when we were starting to look into that, this Revit integration, we were like, OK, so how do we import it? So the client does. Then he makes an export, bring it through Dynamo and generate and [INAUDIBLE]. But what we came across, and that's where Mustafa started to work with the Revit IO team, is, actually, rather than even using Dynamo directly, we can import through an add-on on the Revit IO ecosystem, and basically generate on the fly our data and map it directly.

So the file that you're looking at, actually, can come directly and can be downloaded directly from the application. But that's not it. Because the beautiful part of this concept of Revit IO connecting to HFDM is, you can do everything you wanted to do in Revit, but in the cloud, without having to open Revit, basically. But for you to understand, we do some screenshots to mimic what we do.

So what is happening? I'm bringing on Dynamo, or, in our case, an add-in, the pad of my application. They are reading the recipes. Remember the recipes that we connected to the layers? They are actually recipes that can communicate to the application. So what is happening here is the pad is scaled down for the offset of the boundaries. The recipe gives you the type of assets so you can extrude it. And from the representation [? we did here, ?] of a simple web GL extrusion, we can go now into a plot and a mass in Revit.

So the more you start to be smart, the more you say, hey, my recipe could connect [? vitality ?] and asset. And that's what we started to do. That's the type of recipes that we are working with. So recipes can be a chain of parameter in your Revit element, to be able to run on your [? feasibility. ?] But your recipe could also be your mass itself. So, on purpose, the blue container was my sketch. And then inside our libraries of blocks, as I told you earlier, we go from the zoom out to zoom in, basically, concept. And the script, then, can generate your technical drawings. So bear with me for a second. That's where two [? blocks ?] [? are ?] for client purposes.

But the script selects each and every plot, make us scrub books around it, make a sheet, drop it into their views, make your section, make your accessibility. You don't see it in the [INAUDIBLE] area here, but basically, there are all of your GPS coordinates for your plot. Which, by default, Revit doesn't do that by itself, as you may be aware. So we have to manually edit them. The data that are blurred there are coming from an external data source. So, long story short, we generate on the fly. This, it was 1,500 plots out of nowhere.

So let me rephrase that one more time, just in case. Because it's late, and I know we're all tired. I go to PDF or 1,500 drawings out of my sketch app. I never open Revit. Isn't that cool? Revit IO is excited. [LAUGHS]. So, again, imagine the floor where we are going nowadays. So going back to what I said earlier. And believe me-- my dad is an architect, which I never could convert into the beam era. But when I started to show him that you can sketch now and technical drawing comes out, he was like, oh, that's pretty cool. I may even start using it.

So it's almost like we will re-empower our knowledge into our industry. So the question I got is, what happened to the younger that knows technology, but they don't have the experience. And I think that that's where it becomes really interesting is, for this, lets say, younger generation that are more technical or technology-savvy to facilitate the creation of these recipes. Because, basically, all the blocks that you are seeing here, that's where you can express your kind of parametric design. That's where you can experiment.

So you don't look at the mass of the project, but you go ahead and zoom into the plot itself and you can render off your feasibility. You could run fractal only there to make [? feasibility. ?] The master plan wouldn't care. The feasibility could come as many iterations as you want. You just connect multi services all together at the same time. And you run iteration, iteration, iteration, iteration. That make sense to you? Yeah? I don't know exactly on the time. We do? We still do, OK.

So, first of all, I would like to know if anybody has any questions. Because I see some faces with big eyes open. And some of them I see--

AUDIENCE: I'd like to see it. I'd like to see how it happens [INAUDIBLE]. It sounds perfect [INAUDIBLE].

MARC DURAND: We have witnesses behind you. [LAUGHS]. As I said to you, the system is-- so, again, because we are segmenting a little bit of step by step at the moment because of all of the surveys that are getting developed and we are interconnecting them one by one. But what you need to understand-- are you familiar, first, with the Revit IO concept? Yeah. So you know that it's, basically, you launch an add-on. Do you agree with me?

So when you launch your query, you can give it a list of barometers. Which, in this case, they are your points. So if you know, also, your Dynamo, you can bring all of your points to make your sketch. So it's exactly the same logic. You just break down the thought process. You pseudocode the delivery and then you look at where you can inject this information to generate the output.

There is one thing that I do have to you know come clear on this presentation. The only thing that we haven't, at the moment, figured out was more of the circulation within the plot. Where you access the entry point and the exit point. So these are things that need to come slowly to enhance. But, as I said, the potential is really really, really interesting. Really, definitively.

And going from an idea of, we do a sketch and we do a mass. And if I go back to the first slide that I show you earlier, because I think that this one, to me, is extremely important. When you go back to this, but on the digital plane tablet that generates this for a quick understanding of what it is you want. But this part is intelligent and you can alter and generate all of your output, which I don't think any of us in this room would disagree saying that that's the boring part, the documentation. I think that that's definitively the workflow that we're putting in place in our company. That's what we aim to do.

That would allow you to do a lot of things. One of it is between us. Imagine this workflow. We don't even have to sanitize the information because it's generated on the fly. You don't need to have your standard because it's generated on the fly. If you change location, your recipe could change. So the building footprint could be regulated by your recipe. So the same master plan I have done in UK, I could have the same version in the US. I could have the same version elsewhere, just by bringing, on top of my recipe, the building code of the country I'm into.

So, as I said, possibilities are literally endless. And today's a sneak peak of what we've been working on with the technology we were given access to. I think there was another question as well.

AUDIENCE: I think you answered it. It was pretty much, re-playing your [INAUDIBLE] data management to generate the 3D bodies and adding on in Revit IO. So I guess you could have just used Revit API directly to create the bodies as opposed to [INAUDIBLE].

MARC DURAND: So, again, we did that, for explanation, we did not use Dynamo. Actually, I'll be shooting myself, but Dynamo doesn't work on Revit IO. Because that would not make any sense, because you can't really see it. But, no, we use API, pure API. Yeah.