WAV is uncompressed. MP3 is highly compressed, but lossy. FLAC is compressed and lossless. If you wanted to store lots of master copies of audio, you could use WAV or you could reduce storage space by using FLAC instead.
So basically, if there's a distribution of possible outcomes (with estimated duration plotted on the x-axis, and probability on the y-axis), then we put bounds on either side of the distribution and take a 'central' point on the x-axis between them. But instead of a simple average, we can use a log scale on the x-axis, so that we represent estimates of different magnitudes well. Now the simple half way point on the x-axis is on a log scale, and (I think) equivalent to the approach you described.
So one question here is: Why reduce the distribution (with long tail or whatever) to a single estimate number? If the distribution represents the range of possible outcomes well, then the single number throws away most of the information in the distribution.
Yes, I don't see the logic of using fertile agricultural land for solar PV. It seems like short term thinking and a planning failure. I've heard at lest one similar story regarding reforesting, i.e., a scheme that was planning to use fertile land in Wales that was ultimately canned due to local objections/resistance.
For the battery storage, perhaps they will ultimately choose https://en.wikipedia.org/wiki/Lithium_iron_phosphate_battery. The cable manufacture seems like more a 'when' rather than an 'if' problem. The scheduled first phase of construction is 2025-2027, which we can reasonably expect to be delayed by 2-3 years if they are trying to build new manufacturing capability. Overall this still looks to be in the realms of possible/viable IMO.
This is anecdata, but I don't think it affects my overnight sleep if my nap is post lunch, although it definitely does if I sleep later in the day, say from 4pm to 6pm. I think there's enough time and mental exertion in the afternoon and evening that I sleep normally overnight. Possibly it's a slightly shorter sleep overnight, but a net increase per day overall, or, at least no shorter overall.
As I understand it REM sleep is at the start and end of each cycle, thus 45 mins would put you squarely in a deep non-REM phase. That said, the cycles get progressively shorter in an overnight sleep, so it's possible a daytime nap sleep cycle would be much shorter than the full 90 mins.
I find if I'm working on a difficult problem, I feel like I can make more rapid progress if I take a 'nap' early afternoon in that post lunch drowsy period. I often start thinking about the problems again immediately on waking, and typically find that is the most productive period in terms of breakthroughs and making mental connections.
I put 'nap' in quotes because I don't set an alarm (unless there is specific need to), and for me that usually means a 2 hour sleep! I figure that is made up of 90 mins of one sleep cycle, and some time either side for falling asleep and waking.
Timsort may give much improved performance if the items in an array are already partially sorted. I had a specific need for this, and couldn't find any good quality implementations in the dotnet world, so ended up putting the time in to port it and update it to the latest fixed/patched version.
As part of the work I looked at the source for Array.Sort and noticed it was doing some redundant null checks when you call it with a secondary sort array. I submitted a pull request to the MS github repo with some supporting performance numbers, and they merged it in.
It looked to me that the code there was originally one class that used null checks to support the secondary sort array, and at some point the code was copied to a second class that does the secondary sort only, and the first class had that logic removed, but some of the null checks apparently got left behind.
Next steps for my timsort will involve moving from dotnet standard to dotnet core, and starting to leverage some of the new functionality, such as Intrisics for x86-64 and Arm. So this article will be a good starting point for that.
In my experience it's often quite obvious when someone doesn't understand something and they say they do. Someone who properly understands will typically offer some tidbit of feedback from their understanding as an acknowledgement. And if not forthcoming I will extend the conversation to try and tease out that bit of feedback. I think this works because the people who genuinely understand something are usually keen to demonstrate that understanding in some small way.
If you cross byte boundaries, i.e. compactly pack the bits such that there are no wasted bits; then you have 4 bits per decimal digit, 50T digits becomes 200T bits == 25TB. Technically there is still some slack in there because each 4 bit block represents 16 values, and we only use 10 of those.
I agree, I think there are a few things in the Windows file explorer that conspire against its good performance (file preview is a big factor, but recycle bin content seems to affect it too), and it does seem to get worse over time. I think there's a market now for a third party 'back to basics' explorer.