-
Due to recent changes in Microsoft Github's policy regarding the use of Microsoft Github's Co-pilot, all development for this project has been migrated from Github to sourcehut.
-
The new project repository can be found here: https://sr.ht/~tpapastylianou/trac-compatible-gantt-chart-template
-
However, if you appreciate this project and would still like to express your appreciation and/or promote it here, then by all means please feel free to star it on Github.
In 2021, Microsoft announced Github Co-Pilot, an AI-powered tool that purports to "help developers write better code" by suggesting snippets (e.g. in an autocomplete fashion) given a programmer's existing codebase as an input.
Like most AI tools, it does this on the basis of code that it has been trained on. However, even putting aside the potential issue of how the quality of such ground-truth code is evaluated, Github being the traditionally de-facto repository of open source code that it was prior to its acquisition by Microsoft means that the vast majority of code on which Co-Pilot has been trained on is available for training precisely because it was open, and typically licensed via open source licenses, this being GPLv3 more often than not. Meanwhile, private, proprietary, closed-source repositories are effectively left out of the training pool.
This poses an ethical problem for conscientious users of Github. Any open-source licensed code on Github can be sequestered for training and reproduced, verbatim if necessary, completely bypassing the accompanying license. This strips the community of any and all rights and obligations which open source licences are designed to protect, with potentially far-reaching implications.
Therefore, while I'm not conceited enough to think that my code represents the crème-de-la-crème of open-source offerings in the wild, I have nonetheless decided to remove all of my open-source licensed projects from Microsoft Github's platform out of principle, as many others in the open-source community have done already.
For a more in-depth discussion of this issue on HackerNews, see https://news.ycombinator.com/item?id=31878290