Primarily identified as a way to enable online gaming (including on mobile devices), this innovation is a major opportunity for CAD (at least, we believe it is!). As a W3C standard, WebGL and HTML5 will kick off proprietary or specific plugins currently required to visualize 3D shapes. Just imagine the benefits for CAD users: get free from proprietary plugins and just use the native features of your web browser! Remove all the heavyweight clients from your machine! You, project manager, review the progress of your design team on your iPhone or Android phone, whether you are in the train or in a cab!

Furthermore, we do believe in the latest technologies to enable what we call “CAD in a browser”: we mean a complete online CAD/CAM toolsuite running in Firefox, Chrome or Safari. Lightweights client rather than heavyweight installations on a local machine. Truelly multiplatform. Requiring low machine resources. Poor deployment and maintenance costs.

Open standards and open source software makes this ideal become possible in a short future: FOS browsers, FOS js libraries, FOS distributed and highly scalable”NoSql” databases, and pythonOCC of course

An experimental WebGL renderer for pythonOCC was developed to test these features. The python script takes any TopoDS_Shape and generates a three.js compliant javascript file. The test was ran from the STEP file of a robot. Below is the screenshot of this STEP file displayed in an heavyweight PyQt client:

The resulting html file can be viewed in a WebGL compliant browser. The following screenshot was taken from MacOSX/Firefox-Beta10 and was successfully tested with Chrome 8 and Webkit (latest rev.).

Useful links:

• the online demo: http://webgl.pythonocc.org
• the source code: http://code.google.com/p/pythonocc/source/browse/branches/webgl

• Wrappers for C++ static methods have been implemented in a SWIG compliant way. Static methods like: BRep_Tool::Surface() is now callable with either BRep_Tool().Surface() or BRep_Tool_Surface()
• Improved project quality : 64 unittests currently pass
• Many improvements related to the graphic side (better qtDisplay.py, abstraction of the CSF_GraphicShr env var)
• Multi selection of entities (vertices, edges, faces, shapes)
• Many fixes and improvements to the ‘addons’ or Level2 API (meshing, PAF, new experimental DYN package, topology etc.)
• New samples (Traits based parametric modeling, display customization etc.)
• License move from GNU General Public License v3 (GPLv3) to GNU Lesser General Public License v3 (LGPLv3)

376 changes were committed to the subversion repository. Check http://code.google.com/p/pythonocc/source/browse/tags/0.5/CHANGELOG for a detailed list of changes.

The following results were obtained from a new installation of Ubuntu 10.04 from the original 32 bit CDROM, on a Virtual Box 3.1.6 machine.

Step 1 : set a new package source.

Insert the following line in the source field:

deb http://ppa.launchpad.net/cae-team/ppa/ubuntu lucid main

Step 2 : search/install the python-occ package

From the Synaptic Package Manager utility, search for the python-occ package :

Mark python-occ for installation :

Wait for the download to be completed :

Step 3 : set the CSF_GraphicShr env variable

The CSF_GraphicShr variable must point to /usr/lib/libTKOpenGl-6.3.0.so :

Step 4 : install a GUI manager

You can choose PyQt4 or wxPython. I rather user PyQt4 :

Step 5 : run the HelloWorld example

You’re ready to run the following example (that creates/displays a box) :

Thomas Paviot and Jelle Feringa, managers of the pythonOCC project, will attend the 12th NASA-ESA Workshop on Product Data Exchange, 18-20 May 2010, Norway. Here is the abstract of the lecture they will give (slides of the presentation will be uploaded as soon as the conference has took place).

Implementation of a SaaS based simulation platform using open standards and open source software

Thomas Paviot*, Jelle Feringa**

*tpaviot@gmail.com, *jelleferinga@gmail.com

Abstract: Software as a Service (SaaS) is underlined in the literature as an important area of developement, and is an active field of research for software or organizations integration. Major PLM vendors already provide SaaS/web based solutions (e.g. Product Data Management software coming with a set of web services and lightweight internet browser clients), but many issues still have to be solved in order to cover the whole product lifecycle engineering workflows. In that sense, the presented paper deals with the simulation issue: currently, most of the CAE tools (FEM solvers, pre-processors and post-processors) used by engineers require expensive hardware, costly licensing scheme and might not be interoperable. This gives rise to high operational and maintenance costs.

The arrival of a healthy biotope of open source CAD/CAE tools, such as Code_Aster, pythonOCC or Salome give rise to new engineering workflows that allow to overcome these issues. Developing a SaaS/web based CAE simulation platform would allow one to overcome these barriers. Such an intent however requires a large range of technologies to be closely integrated, e.g. web technologies, remote public services, data visualization, 3D modeling, meshing, simulation, optimisation etc. The goal of this paper is to present an approach based on open source software and open standards that facilitate the achievement of a SaaS simulation workflow, by using a strategy of interoperable components rather than an integrated monolithic solution.

In the first part of the paper, we present the global functional architecture of such a simulation platform. Open standards are underlined as a key-concept for enabling this functional architecture to be technically implemented: open standards for data representation as well as a topology of webservices intended to data exchange. In the following section, the different frameworks that support the solution are detailed: web server, parametric modeling kernel, on-line data visualization, meshing and simulation/optimisation frameworks. The third section introduces two scenarii aiming at modeling common industrial workflows and, at the same time, validate the presented solution. As a conclusion, implementation issues are detailed and a functional demonstration is presented.

Keywords: Software as a Service (Saas), Knowledge Based Engineering (KBE), parametric modeling, product data visualization, Finite Element Method (FEM) based simulation.

Slideshow of the presentation:

Screencast of the demo

STEP file import (6.1Mb, 75895 entities)

Bien qu’il existe des logiciels libres dédiés à la modélisation 3D orientée jeux, la communauté open source souffre encore du manque d’applications CAO 3D industrielles. Les enjeux sont pourtant très importants, notamment du fait de problèmes d’interopérabilité liés à l’utilisation de logiciels commerciaux. pythonOCC est une plateforme d’applications CAO/PLM, basée entre autres sur la bibliothèque C++ OpenCASCADE 3D.

Le diaporama de la présentation est disponible au format PDF (3.6Mb): PYTHONOCC_PYCONFR_2009_PAVIOT

pythonOCC was presented at the last PDE2009 conference (http://step.nasa.gov/pde2009/). Here is the abstract of the presentation: The combination of open-source software and open standards are known to be considered as a key success factor of further technology development. The interest around the STEP standard is growing these last years, since the release of Application Protocols that allow to foresee new software solutions able to ensure a consistent product data description covering the whole lifecycle. This presentation focuses on the features of a free 3D modeling/data exchange library in the context of a research work dealing with the STEP AP239 standard for PDM/ERP interoperability. The goal is to merge these two approaches and describe the product from design to manufacturing, with a granularity that allow to keep the control over product configuration.

The slideshow can be downloaded in PDF format (3.2Mb): PYTHONOCC_PDE2009_PAVIOT_FINAL