Open source products include permission to use the source code, design documents, or content of the product. It most commonly refers to the open-source model, in which open-source software or other products are released under an open-source license as part of the open-source-software movement. Use of the term originated with software, but has expanded beyond the software sector to cover other open content and forms of open collaboration.From Wikipedia, the free encyclopedia
In the last decade open source projects have changed the way we make and use software – in a staggering shift from earlier days of software development, when the code was a heavily guarded secret, many of the modern frameworks and tools that power our modern lives are put on GitHub or other collaboration coding websites for other people to inspect, modify and use. The examples include Tensorflow, Blender, Linux, Apache, Chromium and many others.
We at TinkerGen also believe in the open-source spirit, which allows exchange, collaborative participation, transparency, meritocracy, and community-oriented development. This is why, for our AI Education Robot, MARK we plan to open-source 98% of the software and certain hardware modules as well. We hope by doing that we can foster the interaction with community and better understanding of what AI and machine learning actually are. What does that mean in practice?
First of all, software for MARK consists of three components:
- MARK Micropython API that can be used both independently in MaixPy IDE or for integration with graphical programming environments
- The Micropython firmware, which is a Micropython wrapper around C code
- C code for MARKduino, MARK carrier board to which the main board and other components are connected
You can browse the source code for Micropython firmware at https://github.com/TinkerGen/MaixPy. It is a fork of MaixPy project by Sipeed, which we maintain to ensure the compatibility with our own Micropython API. 98% of the code is open for browsing and modification, with the exception of KPU C libraries, which contain face recognition processing code, that is a property of Sipeed.
With release of next Micropython API version for MARK, 5011, which will feature a major overhaul of internal code structure we will upload the Micropython API code and C code for MARKduino to our GitHub as well. Right now, if you already purchased MARK you can browse the Micropython API files with uPyLoader utility when connected to MARK with USB cable. We do not encrypt or in any other way hinder the access to code, models or any other auxiliary files that are present in file system of the robot.
Further, to enable makers and hobbyists to customize their MARKs we will upload the top cover and other plastic parts SLT files to Thingverse. Currently we have designers working on modifying the files to make them more suitable for printing on FDM printers, the most common type of 3D printers available in schools and makerspaces.
We hope that by open-sourcing the software and designs for MARK we can involve the community to build even better Make a Robot Kit and introduce more people to AI and machine learning.
Stay tuned for more articles from us and updates on MARK Kickstarter campaign.
For more information on Grove Zero series, Codecraft and other hardware for makers and STEM educators, visit our website, https://tinkergen.com/.