Downloaded 23 times
More Related Content
Similar to Collaborations in the Extreme: The rise of open code development in the scientific community
Collaborations in the Extreme: The rise of open code development in the scientific community
- 1. Collaborations in theExtreme: The rise of open code development in the scientific community Kelle Cruz CUNY Hunter College @kellecruz
- 2. I am anobservational astronomer Sun Brown Dwarfs Low-Mass Star Brown Dwarfs Jupiter Image credit: R. Hurt, Caltech/IPAC
- 3. I am anobservational astronomer
- 4.
- 5.
- 6.
- 7.
- 8.
- 9.
- 10. I care aboutmaximizing efficiency of science The goal of this blog, AstroBetter, is to provide information and tips about streamlining all the ways we use our Macs everyday.
- 11. There is nowastrolib (https://trac6.assembla.com/astrolib ), also pyAstroLib (which appears to be something entirely different than astrolib? http://sourceforge.net/projects/ pyastrolib/develop), astropysics (http://packages.python.org/Astropysics/), now pyAstronomy, and probably more. It looks like PyAstronomy contains wavelength conversions (already implemented in pyastrolib and astropysics) and coordinate conversions and precession (already in coords, astropysics, pySLALIB, & more). I'm going to be provocative here: As a community, we are doing something wrong if everyone wants to start their own new module rather than contributing to a common, shared, open- source core. We are clearly doing something wrong if people repeatedly implement the same basic functions rather than building on what's already there. What do we need to do differently? How can we make it easier to use a shared repository and shared namespace for all this? https://mail.scipy.org/pipermail/astropy/2011-June/thread.html
- 12. https://mail.scipy.org/pipermail/astropy/2011-June/thread.html I think asit-down is desperately needed to resolve these issues, figure out the mgmt structure (aka, pecking order), for the BDFL to emerge, and for progress to occur. I'd be happy to participate as a non-python/programming expert and maybe provide the voice of the “users”. …if there are > 3 ppl on this list already planning on going, then you guys should definitely get together to have something a bit more formal than just a chat. You could plan the next meeting, evaluate the existing packages, start to sketch out a roadmap, maybe even assign some tasks, or create some leadership roles (e.g., Lead of Phot packages, Lead of Spectra packages, Lead of Xray), etc.
- 13. Describing code ishard To calculate distances and areas on the solar-disc, firstly helioprojective coordinates are converted to heliographic coordinates using the world coordinate system available in the standard solar data analysis software… doi: 10.1007/s11207-015-0808-7 Courtesy Danny Ryan
- 14. Science should be reproducible/replicable samedata + same methods = same results Modified from https://xkcd.com/242/ https://arxiv.org/abs/1802.03311 - Barba 2018.
- 15.
- 16.
- 17. Open Source Definition Softwarewhose source code is made available under an open source license. An open source license grants anyone the rights to inspect, modify and distribute the code (with some possible caveats). https://opensource.org/osd
- 18. Open Source Misconceptions Puttingcode on my webpage makes it open. True False https://choosealicense.com/no-permission/
- 19. Open Source Misconceptions Puttingcode on my webpage makes it open. No license = copyright, all rights reserved. Nobody else can use, copy, distribute, or modify your work without being at risk of take-downs or litigation. True False https://choosealicense.com/no-permission/
- 20. Open Source Misconceptions IfI make my code open, I must support and maintain it. True False
- 21. Open Source Misconceptions IfI make my code open, I must support and maintain it. Making code open means other people can use and modify it, even if you abandon it. True False
- 22. Open Source Misconceptions Anyonecan participate in the development of open code True False
- 23. Open Source Misconceptions Anyonecan participate in the development of open code Code can be distributed with an open source license but developed behind closed doors True False
- 24. GitHub has madeopen development feasible • github is the thing which makes open development possible. • github can be used to work on closed software. 32nd most popular internet site in the US https://www.alexa.com/topsites/countries/US A repository to store, exchange, trade, and talk about code.
- 25. • Version Control:clean, simple-to-track coding history • Accessible: projects are discoverable and widely accessible • Transparent: work in progress is visible • Collaborative: Revisions are debated publicly • Educational: critical coding knowledge is amassed via all of the above https://resources.whitesourcesoftware.com/blog-whitesource/the-octocat-that-changed-the-world GitHub has made open development feasible
- 26. GitHub has madeopen development feasible
- 27. GitHub has madeopen development feasible
- 28. GitHub makes collaborationpossible
- 29. Woo-hoo! Que felicidad! GitHub has madeopen development feasible
- 30. • Python playswell with others highly interoperable with other languages • Python is powerful large ecosystem of 3rd party modules • Python is easy to learn low barrier to entry • Python is open helps with reproducibility problem • Python is zero cost Nobody likes to pay for things Scientists are converging on Python https://www.python.org/about/ https://www.datacamp.com/community/blog/python-scientific-computing-case
- 31. Wide-adoption of Pythonincreases collaboration https://octoverse.github.com/
- 32. Scientists are convergingon Python
- 33. Scientific Python packages are getting lotsof use https://stackoverflow.blog/2017/09/14/python-growing-quickly/
- 34. Scientists are convergingon Python Modified from https://xkcd.com/353/
- 35. 7 §Total downloads viaPyPi, Jan. 2016 — Feb. 2018 *Estimated cost based on $100k average salary per year **defined as more than 40 commits since Jan. 1, 2017 ***defined as more than 100 commits since Jan. 1, 2017 openhub.net Contributor stats via Libraries.io Maintainers stats via github.com 6 Current Maintainers ** Current Maintainers Downloads 27.7 Million§ Estimated Cost $7 Million* Codebase 267,775 lines Contributors 703 70 person-years 4 latest release: v0.22.0 (development began in 2009) *** latest release: v1.14.0 (development began in 2001) 76 person-years Downloads 49 Million§ Estimated Cost $7.57 Million* Codebase 285,480 lines Contributors 564 *** Current Maintainers Downloads 13.8 Million§ Estimated Cost $6.36 Million* Codebase 242,431 lines Contributors 580 64 person-years 5 latest release: v2.2.0 (development began in 2003) Vital code is maintained by a small number of people
- 40. Implications for thefuture of academic science • Software development skills will be considered fundamental & essential. • Software will be considered a research product, similar to a journal article. • Code review will part of peer review process. • All code will be expected to be open and accessible. • Contributing to open, collaborative projects will be the norm.
- 41.
- 42. Conclusions • Science issoftware • GitHub facilitates open development and has changed the way we interact with software. • Python is rapidly becoming the common language of science. • NumFOCUS needs to be recognized as a a vital part of science infrastructure. • Software development needs to be taught and integrated into the formal STEM curriculum.
- 43. Collaborations in theExtreme: The rise of open code development in the scientific community Kelle Cruz CUNY Hunter College @kellecruz