I will be away a most of the rest of this year, and leave you with the following exciting mix of very good and very bad news:
DevCon Europe is coming up, and the Call for Papers is open, including customer submissions.
DevCon is classified into three tracks:
The core topics include:
Please submit your papers by November 14th and join us for an unforgettable experience filled with learning, networking, and inspiration. This is your chance to share your expertise, innovative ideas, and groundbreaking projects with a vibrant community of developers, engineers, and tech enthusiasts.
We are thrilled that the Autodesk DevCon Europe 2025 will be held in the beautiful city of Amsterdam! Whether you're working on cutting-edge software, pioneering new technologies, or have unique insights into the future of development, we want to hear from you!
Don't miss out on the opportunity to be a part of this exciting event!
My retirement is coming up, currently scheduled before the middle of next year.
Since I have heaps of unused accrued vacation that I need to consume beforehand, and not enough remaining working hours to do so, I am taking leave in November and December 2024.
So, I will be less active both here in The Building Coder blog and in the Revit API discussion forum for the rest of 2024.
Happily, during that time, new growing capabilities offered by the AI LLMs may enable them to step in and help:
Back in 2017, I pondered a question answering system for Revit API.
I called it Q4R4, short for QA 4 RA, aiming for a short yet globally unique term.
Here are some notes on that project:
Since the advent of ChatGPT, I thought I might create a collection of Revit API questions and answers to feed a RAG-enhanced LLM system. Currently, RAG in evolving further into GraphRAG, KnowledgeGraph, etc.
That would make sense if the Revit API knowledge that I wish to use was proprietary and not public.
Luckily, however, I have now spent over 15 years pushing hard to make all Revit API questions and answers publicly available.
Those efforts are now paying off, and there may be less need for my research and efforts to create a closed-loop RAG-base system; instead, ChatGPT now searches the live internet, including the Revit API discussion forum and the blog:
I tested it on a real live forum question on rebar cover geometry extraction and visualisation using C#. The result is promising: ChatGPT was successful in answering the customer's question, cf. the first LLM forum solution below.
So, possibly no need for RAG or any dedicated chat machine if you have the RAG sources on the public Internet. Just use a generic AI that has up-to-date Internet access and tell it where to look for the answers.
This is the first AI LLM-generated Revit API discussion forum solution Q & A:
Original Question: Do we have any way to extract/ visualize the rebar cover geometry given to the host element. I need to create curves with respect to the cover geometry. Could anyone help me on this topic?
Answer: ChatGPT now supports improved real-time access to the Internet.
Therefore, it can be used to answer questions that require looking for answers in the Revit API discussion forum, The Building Coder blog, and elsewhere on the web. Therefore, I simply rephrased your question as follows and received the attached answer. Please check it out and see whether it helps you move forward. Please let us know the results and how you and up solving this. Thank you!
Edited Question for ChatGPT:
How can I use the Revit API to Extract and Visualize the rebar cover geometry given the host element? I need to create curves with respect to the cover geometry. This has been discussed in the Revit API discussion forum and by The Building Coder blog.
Answer by ChatGPT:
To extract and visualize rebar cover geometry using the Revit API, you can follow these steps:
Example Code: here is a simplified code outline in C# to extract cover geometry and create curves around it:
using Autodesk.Revit.DB;
using Autodesk.Revit.UI;
public Result Execute(ExternalCommandData commandData, ref string message, ElementSet elements)
{
Document doc = commandData.Application.ActiveUIDocument.Document;
// Assume you already have the host element (like a structural element)
Element hostElement = ... // retrieve your host element here
// Get geometry options
Options geomOptions = new Options();
GeometryElement geomElement = hostElement.get_Geometry(geomOptions);
double coverDistance = GetCoverDistance(hostElement); // retrieve cover distance from rebar cover type
// Iterate through geometry to find the faces
foreach (GeometryObject geomObj in geomElement)
{
Solid solid = geomObj as Solid;
if (solid != null)
{
foreach (Face face in solid.Faces)
{
// Compute offset face or create curves around the face based on coverDistance
List<Curve> coverCurves = CreateCoverCurves(face, coverDistance, doc);
// Display curves in model (e.g., as model lines)
using (Transaction tx = new Transaction(doc, "Create Rebar Cover Curves"))
{
tx.Start();
foreach (Curve curve in coverCurves)
{
doc.Create.NewModelCurve(curve, sketchPlane);
}
tx.Commit();
}
}
}
}
return Result.Succeeded;
}
// Helper methods to retrieve cover distance and create curves around face would go hereAdditional References:
This approach provides a general workflow and can be tailored based on the specific cover types and face orientations in your project.
Response: Thanks @jeremy_tammik, I extracted the host geometry, using each face and the known Curve Distance I'm able to extract the geometry. I faced an issue finding the curves of Cover from the face edgeLoops (x and y axis will return error). In that case, first I used the edge curve to make an offset in Z direction and made the offset curve translation in Z direction (to eliminate X and Y axis error). Below is the code i used to find the curve of an edgeLoop curve:
Curve hostFaceCurveOffset = curve.CreateOffset(
coverDist, XYZ.BasisZ); //to move horizontally
Curve hostFaceCurveTranslation = offset.CreateTransformed(
Transform.CreateTranslation(
new XYZ(0, 0, coverDist))); // To move vertically
Thank you for the confirmation, Thomas Tinu, thanks to ChatGPT for the research and answer, and thanks to the entire community for years of discussing questions and answers in public to feed the LLM!
We are happy to present new modeless WPF add-in samples and tutorials:
First, the Revit addin with modeless WPF window with XAML tutorial by Sergei Nefedov came up in the question on Calling IExternalCommand from WPF button.
Secondly, Moustafa Khalil of SharpBIM (GitHub) shared a Revit_WPF_Example in his answer to Revit crashes when exiting add-in WPF application.
Back to AI-related news, a succinct summary of one problematic aspect was provided several years ago by Rich Sutton in The Bitter Lesson:
One thing that should be learned from the bitter lesson is the great power of general purpose methods, of methods that continue to scale with increased computation even as the available computation becomes very great. The two methods that seem to scale arbitrarily in this way are search and learning.
The second general point to be learned from the bitter lesson is that the actual contents of minds are tremendously, irredeemably complex; we should stop trying to find simple ways to think about the contents of minds, such as simple ways to think about space, objects, multiple agents, or symmetries. All these are part of the arbitrary, intrinsically-complex, outside world. They are not what should be built in, as their complexity is endless; instead we should build in only the meta-methods that can find and capture this arbitrary complexity. Essential to these methods is that they can find good approximations, but the search for them should be by our methods, not by us. We want AI agents that can discover like we can, not which contain what we have discovered. Building in our discoveries only makes it harder to see how the discovering process can be done.
I ran into The Bitter Lesson reading about Brain Drain: David vs Goliath:
There are worries that GenAI systems may run out of fresh data as they scale. Synthetic data is an option, but using AI-generated data to train AI can degrade the model's performance. There may be a better solution. Can data quality overcome a loss of data quantity?
New research presents interesting results on human creativity in the age of LLMs: randomized experiments on divergent and convergent thinking:
Large language models are transforming the creative process by offering unprecedented capabilities to algorithmically generate ideas. While these tools can enhance human creativity when people co-create with them, it's unclear how this will impact unassisted human creativity. We conducted two large pre-registered parallel experiments involving 1,100 participants attempting tasks targeting the two core components of creativity, divergent and convergent thinking. We compare the effects of two forms of large language model (LLM) assistance -- a standard LLM providing direct answers and a coach-like LLM offering guidance -- with a control group receiving no AI assistance, and focus particularly on how all groups perform in a final, unassisted stage. Our findings reveal that while LLM assistance can provide short-term boosts in creativity during assisted tasks, it may inadvertently hinder independent creative performance when users work without assistance, raising concerns about the long-term impact on human creativity and cognition.
Moving away from both AI and the Revit API, here is a neat introduction to HTML for people, showing how to create and publish a web site from scratch, right now, for anyone.
Finally, unfortunately, the world is way off target in tackling climate change.