Everyday tools, quiet habits


Greetings. preserves every pixel exactly. Quality loss enters through wrong tools: JPEG intermediates, silent downscaling, or flattening transparency against unwanted backgrounds. A PNG-to-PNG merge with matching dimensions is mathematically clean. How do I merge PNGs with different sizes? Decide the policy first: pad smaller images to the shared dimension (preserves all pixels, adds background), align and crop the excess (uniform edges, loses some content), or scale to match (uniform, but resamples pixels). Good tools make the choice explicit; bad ones scale silently. Can I merge PNGs with transparent backgrounds? Yes — transparency merges natively, with each image's alpha channel preserved into the combined file. Where transparent regions from different images overlap, the alpha values compose. For a flattened result instead, choose the background color explicitly during the merge rather than discovering white-on-white invisibility later.

Nuclear


Even with the inclusion of Chernobyl and Fukushima, nuclear is per terawatt the lowest death source of any method of generating power. It also does not contribute to greenhouse, and the old bugaboo of long term million year nuclear waste can be negated by the construction of molten salt breeder reactors or mixed spectrum reactors which can fission the long term actinide waste from existing reactors and produce 20x as much energy from the waste as the initial reactor produced from fuel. These reactors are safe by physics and the fast and mixed spectrum versions have no known radioactive isotope dispersal mechanisms.

What we are calling waste now consists of two components, actinides, those elements that are as heavy or heavier than Uranium, these are the long term waste components that last a million years. The other is fission product, MOST (not all) of these decay away in 300 years, the few that don't can be left in a reactor with a sufficient neutron economy to be transmuted or can be transmuted with a particle accelerator using 5-10% of the reactors output. The waste actinides contain 95% of the fuels original energy after the first pass in a light water reactor. A molten salt fast or mixed spectrum reactor can safely recover the remaining 95%.

The existing method of one pass then bury is the equivalent of filling a cars gasoline tank with 20 gallons of gas, traveling far enough to burn off one gallon and then dumping the remaining 19 gallons on the ground. Is it dangerous? Of course it is, it still has 95% of the energy potential. This is the same thing we are doing with nuclear but do not need to.

I think why nuclear is scary to so many people is because radiation is invisible, you can cure that for $100-$200 by purchasing a Geiger counter or scintillation detector, the latter are somewhat more expensive but they measure gamma ray energies so the source can be identified.

reshared this

7.x Kernel


We have reverted from 7.2.6 kernel back to 6.18.52 kernels for friendica.eskimo.com and our other federated sites.

The reason we have done so is the scheduler in 7.x has disabled the ability to specify a non-preemptive kernel, but on an 18-core 36-thread very busy machine, this results in a lot of trash, wasted CPU and memory cycles and eventually some of the kernel processes don't get enough CPU, the whole machine backs up and does, where it remains sane under 6.18.x kernels.

I did file a bug report with kernel.org to try to make them aware of how the new scheduler logic was NOT working for some hardware configurations and workloads and was basically told to fuck-off.

I think it is sad that Linux which up to this point has been highly configurable and worked for a wide variety of hardware and workloads is now being optimized as a Windows replacement and a game console OS not a server OS.

Roland Häder🇩🇪 reshared this.

This website uses cookies. If you continue browsing this website, you agree to the usage of cookies.