Janine George

As I am at a federal institute, this is part of the institution by design. It's not always easy to orchestrate.

We also have some research data management funding in Germany that supports joint work on infrastructure. I really think this is a great effort.

I would love it to be more frequent.

I also burned through 4 non-science books within 2 weeks but I now have to prepare a few talks again. 🫠

I'd rather be involved in a project where I'm one of many than constantly reinventing the wheel.

🫠

(I had my ERC starting grant intwrview shortly after)

(Not the water itself but the metal surface of the boiler while it was still heating. Thus, it was insanely hot. The burn also looked very interesting. )

I think the last time I was this tired, I burned my arm with a water boiler by accident. It was not nice.

Completly sleep-deprived. 🫩

And thanks! 😅

Honestly, depending on the size of the todo list, those items can be indeed a challenging. 😬 Especially, the lunch bit.

Ooooof, my brain. 🙃

Not sure if I have done anything useful today, but i certainly was able to remove some tasks from my todo list. 🫩

We have updated our preprint on ChemRxiv discussing GW-BSE calculations on CdS zincblende 'quantum dots'

#chemsky #compchem

doi.org/10.26434/che...

Error-handling or human-in-the-loop-style workflows would therefore be really important. I do understand that this is more complicated.
While one can restart jobflow-remote jobs with different parameters, I am also not sure this is ideal for reproducibility.

Photo of small wooden sign saying "I have a crystal for that"

Not sure this was intended as a materials discovery sign... but it is now #chemsky #compchemsky

2025 HybriD3 Materials Theory Training Workshop The 2025 HybriD3 Theory Training Workshop will be held at Duke University on November 21 and 22, 2025. The workshop is aimed at graduate students and postdocs entering the field; senior undergraduate ...

For anyone near the Research Triangle (North Carolina): The groups of Yosuke Kanai and ours will be organizing a Theory Training Workshop - primarily intended for incoming graduate students or postdoc but open to anyone (subject to capacity):

hybrid3.duke.edu/workshops/20...

#chemsky #compchemsky

Thanks to everyone that submitted an abstract 🙌🏻.

⚠️Don't forget to complete your payment before October 10th:

➡️ jtms2025.sciencesconf.org

@reseauscf.bsky.social
@sfp-physique.bsky.social
@cecamevents.bsky.social
@icmcb.bsky.social
@icgmmontpellier.bsky.social
@univbordeaux.bsky.social

Income conference logo

That's a first for me: Conference Logo in 3D and with light 🤩.
I love it.

Thanks for sharing! (Very nice talk!)

Other 3.  G. Gryn’ova , T. Bereau, C. Müller, P. Friederich, R. C. Wade, A. Nunes-Alves, T. A. Soares, K. Merz Jr.,  EDITORIAL: Chemical Compound Space Exploration by Multiscale High-Throughput...

Due to popular demand, I am sharing my slide deck on #environmental costs of #ML in #compchem & #materials_science www.grynova-ccc.org/other.html - incl. work by @skepteis.bsky.social @pschwllr.bsky.social and others.
Please send any other relevant resources my way, I’ll be maintaining this deck.

Guidelines for robust and reproducible point defect simulations in crystals Many physical properties of functional materials are governed by their impurities rather than their bulk characteristics. Defects in crystals can activate electronic and ionic conductivity, create act...

Defect calculations have many pitfalls and key considerations for achieving good accuracy 🎯

In this perspective, we discuss these issues, how to avoid and how we can make defect simulations more reproducible – particularly important with more ML developments! 📊

chemrxiv.org/engage/chemr...

Ah, I see! Thanks!

As you can imagine I was mainly asking from the perspective of someone interested in developing workflow tools. I wanted to see if there are any blockers for applying them for this specific purpose.

As far as I understand, your own developments facilitafe submissions to HPCs rather and do not reuse other workflow tools. (Please correct me if I am wrong. As I do not work much on defects, I have never used your codes directly)

Hi, nice work as usual!

How do you see the role of workflow tools (e.g., workflow languages and definitions) in this field? Wouldn't they in principle bebvery suitable to solve some reproducibility challenges? I am asking out of curiosity.

Winter is coming. 6 °C in Dresden this morning.

Heute geht's direkt wieder zur nächsten Dienstreise. 😬 ...

Grad ein bisschen viel.

I won't. No worries. I am too stubborn anyway.

It might be that it is not too well supported in pymatgen as the code mostly focuses on solids

pymatgen/src/pymatgen/symmetry/analyzer.py at 682bfd855dd89264b0ff8d6bee4815f8834cc5ac · materialsproject/pymatgen Python Materials Genomics (pymatgen) is a robust materials analysis code that defines classes for structures and molecules with support for many electronic structure codes. It powers the Materials ...

github.com/materialspro...
I think you should be able to set the tolerance at least. Operators, I would need to take a closer look.

I see.

In my case, both of the comments were not related to my workplace.

I am happy to help with pymatgen. This should be fine.

Conda environment with Python 3.11 first, then "pip install pymatgen". Then, you need to read the file. Molecule.from_file("....xyz") might work.

If not, please let me know.