In particular from the article I was confused by this:
> NumPy doesn’t offer a way to store data outside of the array buffer—there’s no concept of “sidecar storage” in NumPy.
But then it goes on and say to he strings are stored on the heap (which clearly is also possible with dtype=object) with an arena allocator. Reading the NEP now
Well, there was no concept of sidecar storage. Now we have the hack we came up with for StringDType to store data on the DType instance and also make it so StringDType arrays don't share StringDType instances, unless the array is a view.
EDIT: looking back at the NEP, I'm not sure it does a great job explaining exactly how the per-array descriptor works. Ultimately it's powered by a hook in the DType API: https://github.com/numpy/numpy/pull/24988. There is only one spot in NumPy where array buffers are allocated, so we hooked there and made sure any arrays with newly allocated buffers get a new DType instance.
On the theme of alternatives, I've been using https://www.followthatpage.com/ on and off for years. In particular for my partner's job hunt, watching some local trade association billboards or single school's `/jobs.html` pages.
Also has a reasonable free plan for personal use (20 pages daily + 20 pages weekly + 1 page hourly).
Definitely much simpler in terms of diffing (only text), but it has this 2000s vibe.
I also tried a similar setup (with gphoto2 getting the stream of jpegs, ffmpeg, loopback device) but the deal breakers for me we're:
- high latency: in the half-second range, consistently out of sync with the audio
- bulkiness: I had to place the camera on the side of the screen rather than on top, so I consistently looked away from the camera
With that experience I didn't find it worth to upgrade the setup (with mini HDMI cable, capture card, power supply) and ended up buying a crappy webcam instead
Are you referring to "US International with AltGr dead keys"? I use it too for programming, English text and sporadic Italian text -- so I'm not annoyed by the dead keys unless I actively look for them.
FWIW, growing up in Italy I have the same complaint as the OP about not being able to type È in the italian keyboard -- it comes up somewhat often in prose.
My only complaint is that I occasionally write to colleagues named Paweł or Michał which are not typable
That's funny, the swedish keyboard can handle éèÉÈúùóò and so on, with deadkeys but we don't need much of them. Also üîñ on top of the native ÅÄÖ. The keyboard has five deadkeys concentrated on two buttons with shift and altgr alternatives. The only special keys we realy need is åäö and é. Very seldom à or á, forget which one. And then usually it's a borrowed word from french or italian.
Currencies also: €£$
The annoying part as a programmer is the way to get to {[]} with altgr button and right pinkie finger, that takes dexterity... Oh, and <> that are bottom left beside z on the same button with shift.
> FWIW, growing up in Italy I have the same complaint as the OP about not being able to type È in the italian keyboard -- it comes up somewhat often in prose.
I always get a bit confused by those namings - maybe? On the one I'm talking about I'm realizing now it's actually two presses rather than one for É (E -> altgr+').
Other commentor mentioned EURKEY, which has it but is otherwise quite similar
> At least in my experience, I don't see a lot of trouble with number vs string in CSV data.
I definitely had my fair share of trouble with locale-defined number format. Importing a column where a thousand is spelled "1.000", any integer between 1000 and 999999 would be wrongly parsed as a float between 1 and 999, while for any other number (like "0,1" for one tenth, "1.000.000" for a million, "1.000,56" for a thousand euros and change) the parser would give up and keep the string.
I usually have had more luck importing as text, then doing some string replacement of separators before finally converting to number.
You might want to read about Maxwell's daemon [1]. I came across this a few weeks back, two links away from the HN homepage.
I found it intreaguing and read more about it but I can't quite umderstand the proposed solutions, likely due to my lack of background in thermodynamics.
Clearly, Maxwell was much better (and funnier) in articulating what I was asking. Looks Maxwell's demon can work some of the time at a molecular level, but on average, you need to do work to extract useful heat from matter at a macroscopic level. The hypothesis being that you need to do work just to identify hot/energetic molecules from the cold/low-energy molecules.
