Open source and the future of nuclear physics | Hacker News Reader

Open source and the future of nuclear physics

github.com

191 pointsBerislavLopac16 comments

One project that I didn't see in the article was the rtos https://marte.unican.es/ which a number of prominent experiments use.

It's weird to see an OS developed from a place pretty close to home...

Not when you live in Palo Alto! So you get an idea of what it's like here. If where you live isn't particularly fancy, it'll be an even more accurate experience.

You live in a firewall?

I've always found academic research to be one of the most natural homes for open source software. It's in the paper formatting software (latex) and much of the lab software.

Didn't nuclear physics also help fuel the future of open source and the web in general, e.g. CERN.

Yes! Turns out open, transparent, low-barrier scientific and technological development snowballs itself. Who would have thunk :)

Where is KiCad? CERN pumps quite a lot of money into KiCad.

https://www.kicad.org/

There is a broader statement: open source is fueling the future. In fact, holding off projects close sourced only makes it profitable for one-party for a while, until an open version spins off. Open source copies are bound to happen.

The title could be adequately rephrased: Intellectual Property Has Been Preventing the Future of Nuclear Physics

Alternate theory: open source is destroying the future. In fact, significant progress is occurring from concentrated capital investments made by public entities or corporations, and development slows down or regresses once open source alternatives become available. The likelihood of open source piracy occurring decreases the overall desire for coordinated, large scale technology development.

What are the problems which require

> coordinated, large scale technology development

and are going unsolved because of "piracy" by the available open source projects that address them?

I'm struggling to come up with an example.

The article mmentions Moose, MFEM, OpenMC, but the really good stuff relevant for us nuclear physicist I would expect being talked about aren't mentioned: Bison, Transuranus, Marmot, Serpent, Vasp, Rattlesnake, Fispact.

On the open source open access side I'm missing Lammps or Geant4.

LAMMPS: https://lammps.org/ - MD Simulator (Sandia Labs)

GEANT4: https://geant4.web.cern.ch/ Monte Carlo code for particles through matter (CERN)

Moose & MFEM are great projects and important in itself, but not relevant in the nuclear physics domain. Software which is remains export-restricted for good reasons. Here a list for the curious mind:

FISPACT-II: https://fispact.ukaea.uk/ Inventory & Source term (UKAEA)

MCNP: https://mcnp.lanl.gov/ - Monte Carlo code (LANL)

SERPENT: https://serpent.vtt.fi/serpent/ - Neutronics (VTT)

VASP: https://www.vasp.at/ - Ab Initio DFT (U Vienna)

TRANSURANUS: https://data.jrc.ec.europa.eu/collection/transuranus - Nuclear Fuel Performance Code (EU/JRC)

BISON: https://bison.inl.gov - Next-Gen Nuclear Fuel Performance Code (INL)

(edit: formatting)

I won't be surprised if one day I saw "Open source is fueling the future of nuclear bomb" on HN.

That would be old news!

My first job as a programmer, in the early 1990s, was working on a tool that was used in one of the last US underground nuclear tests. It was a finite-element electromagnetic effects simulator, designed to help get off of actual bombs blowing up and on to simulation to help design for radiation and EM effects on military equipment.

My part was a 2.5d viewer that showed the input model and let you rotate it around and zoom in and out and such. We used X386 (the predecessor of XFree86), the Athena widgets, and GNU tools on a commercial SysV Unix for PC hardware. This was the era of the actual 386; maybe Linux existed but it was well pre 1.0 (I know because I started installing Linux in the 0.99 kernel days, and that was 3 or so years later)

Open source has been fueling the future of every single technical endeavor for 30+ (40+? more?) years.

Reply on news.ycombinator.com