Retrospective on 10 Years of colour-science
Today is the 10th anniversary of Colour and colour-science!
(colour-science-py3.12) Eris:colour kelsolaar$ git log --reverse commit 90bc42b9fedfc7291c7023247eab14b41d5c29af Author: Thomas Mansencal <thomas.mansencal@gmail.com> Date: Sat Apr 5 14:07:48 2014 +0200 Initial commit. Signed-off-by: Thomas Mansencal <thomas.mansencal@gmail.com>
It is time for a decade of work retrospective.
We would like to start by thanking all the contributors whether they write code, improve the documentation, create issues or start discussions.
Over the past 10 years, we built a strong community of people loving colour science. There have been many passionate discussions about this complex scientific domain. To this day, we still do not have full understanding of how human vision works, so we are excited about the future.
Colour has become one of the most comprehensive open source colour science library, irrespective of the programming language.
colour-science now has almost 50 repositories, the main ones are as follows:
Repository | LoC (Python) | Commits | Stars | Forks |
---|---|---|---|---|
colour | 207000 | 4970 | 1936 | 85 |
colour-checker-detection | 3900 | 362 | 202 | 28 |
colour-datasets | 12000 | 358 | 52 | 11 |
colour-demosaicing | 1500 | 457 |
260 | 57 |
colour-hdri | 8500 | 661 | 126 | 18 |
colour-specio | 1450 | 74 | 3 | 1 |
colour-visuals | 4600 | 40 | 29 | 1 |
colour-science.org | N/A | 523 | 10 | 3 |
awesome-colour | N/A | 114 | 251 | 19 |
To update what we wrote 4 years ago, here are the highlights at this date:
-
Seven actively maintained Python packages:
-
Two actively maintained websites:
An active Twitter account
Affiliated project of NumFOCUS
A collective to track our expenses
-
Active participation to other Open Source projects:
Growing number of contributors
Cited by a growing number of publications
Trended for a few days in 2018 on Github
-
Used by an ever growing number of companies, studios and universities:
Google
Hey Thomas, Michael, Thanks for your work on Colour. It's fantastic, and has already helped us improve color correctness in our processing pipeline at YouTube (with more improvements coming soon, and hopefully fixes for open-source video pipelines too). If you're ever in SF, I'll buy you a round! Thanks, S***
Thank you for sharing my PhD. After 15 years of MATLAB only I will be teaching my first course based on Python and using @colour_science this spring. Thank you for your great work.
— Jan Fröhlich (@Jan_Froehlich) February 5, 2020
What is the future of Colour? We thought we could merge the GPU backend four years ago and ultimately decided against. The main issue is that it was very much vendor locked because of Cupy. In a hindsight, this was a good decision as there has been some exciting new development with the Python Array API Standard. We haven't started work yet to support it, but we have certainly discussed about it.
We also thought that 1.0.0 would be released, but this did not happened! We instead continued to improve the API and a stable release will only be possible after adoption of the Python Array API standard.
With all the written, we are hoping to continue for at least another decade. Feel free to join us on Github Discussions, contact us on Gitter or by email.
The Colour Developers
Comments
Comments powered by Disqus