With successful digital transformation comes the need for accurate and reliable content retrieval. Of course, disparate formats make unifying search across file types problematic.
The vdR Group is the publisher of CADnection and has been spearheading digital transformation for their clients for over three decades. Over 220,000 global users benefit from the product lifecycle management and enterprise content management technology they have created.
As a sneak-peak of his talk at ACTIVATE, we spoke with Martin van der Roest, Business Development at The vdR Group, about how vdR solved the conundrum of simplifying information access through a universal translator.
Why is search important at The vdR Group?
vdR is in the business of providing data management solutions specifically for engineering, manufacturing and architecture, engineering and construction (AEC) organizations. We have developed significant expertise in the areas of enterprise content management (ECM), product data/product lifecycle management (PDM/PLM), and various enterprise application integrations. And in almost all of our work, there is a need to treat computer aided design (CAD) files. These CAD files, also commonly referred to as CAD models, possess a richness of data reflecting intellectual property, design intent, spatial modeling, quality parameters, analysis, and other product details. Most anything that is manufactured and/or built today is characterized via CAD models.
But finding a CAD model based on its content is challenging, which is why organizations use ECM, PDM, and/or PLM solutions. These solutions effectively provide a “profile card” feature whereby metadata or properties are assigned to CAD models and then these properties can be searched on. But, all the data within the CAD models is still locked-down and inaccessible. You can’t get to the data unless you’re in the original CAD authoring application, which is simply impractical when you’re trying to find a model amongst thousands of models. So we developed CADnection and integrated it with Lucidworks Fusion.
What does CADnection do?
CADnection fuses Lucidworks Fusion with CAD model data. CADnection liberates all the data in CAD models and provides access to not only the metadata/properties, but also the structure and bill-of-material (BOMs), “where-used” and “composed-of” relationships, manufacturing/quality details, geometrics dimensions, and units of measure normalization. And, once you’ve found the CAD model, CADnection provides full 2D and 3D visualization. As a result, we are now able to offer our audience a platform that augments ECM, PDM, and PLM solutions. It’s a very organic approach that combines unstructured and structured data.
Tell us about the search-related innovations in CADnection.
For CADnection, it’s about making sense of all the data in a CAD model. There are about 15 different formats offered by the leading CAD vendors, such as Autodesk, Dassault, PTC, Siemens, and SolidWorks, and each have their own “secret sauce” for structuring and organizing data. So we’ve adopted a normalizing convention. Think of it as a universal translator that gives us a data structure that we can operate against regardless of the CAD model’s origins. We then process the data further to derive many other insights … like the “where-used” and “composed-of” relationships.
What new technologies are you most excited about?
Machine learning. This is frankly a new frontier for us. As I just mentioned, we have developed mechanisms to derive various insights, but these are created via explicit instructions. But as I understand it, machine learning will give us the ability to perform tasks without explicit instructions and instead rely on patterns and inference.
I can see an immediate opportunity related to scanned drawings. There are a quizillion drawings floating around that were manually produced and still need to be used (think oil and gas, all the different utilities, infrastructure and geography-based applications). If we can learn to detect the title block region of a drawing, or the area for notes and instructions, etc., then theoretically we can retrieve text that can imply drawing name, description, date, author, etc. This could be a game changer.
What are you excited to learn about at ACTIVATE this year?
It always amazes me what folks come up with. Search technology, coupled with the recent convergence of machine learning, will no doubt offer exciting new value and insights.
It reminds me a bit of Uber. It took the sequence of the internet, cloud services, location services, payment services … than poof … Uber. Likewise, we have Lucene/SOLR, Lucidworks Fusion, machine learning … and poof … what will we see at ACTIVATE? I’m excited to find out.
Don’t miss Martin van der Roest’s talk at ACTIVATE “Liberating Computer Aided Design (CAD) Data to Drive Product and Process Insights and Visibility” Wednesday, Sept 11 at 11:35AM.