@Adnan actually... I still can't take X from the docs and paste it into my editor. I have to find it in the source code, paste the "\u..." Somewhere, and use that instead...
@El'endiaStarman I read a fair bit of CS Lewis growing up. He was possibly more averse to separating different mythologies than anyone I've heard of. I liked how The Cosmic Trilogy managed to combine several into a coherent storyline
@trichoplax Lolz. That book was one of the only books I've ever had to start multiple times. It really took some willpower to push through the excruciatingly boring beginning.
@Adnan it's the only choice I have. I tried opening Info.txt in both vimpager and TextEdit, but if I do that, (almost?) any command I paste (non-ASCII) is incorrect. The greator one was normal, but the same just happened for "last" and "a[0:-1]"
Y'know, we've had a lot of "alphabet" challenges recently. (one two three four.) While I love a good challenge, and those challenges were very fun, I think it's time for a change of pace. We need to exclude such challenges in the future. It's time for automation!
You're going to find some alphab...
@Adnan okay, cool. then here's my idea: say you have a string that's terminated by ¤, as in "...¤. Then you could interpret the contents of said string as a series of functions to apply on the TOS, After, concatenate the results together.
@quartata Not really. I pulled it once a couple of weeks ago because somebody complained about a non-working feature, but since nobody tells me to pull, I rarely remember.
@Dennis feature request: add a button to TIO that writes an explanation of your code and any interesting golf tricks you used so that you don't need to write anything at all, just copy and paste.
@FryAmTheEggman No, the type itself. I'm implementing a Number type and using binary right now, with a fixed amount of digits after the comma (1024). That works nice and all, but Number('1000.1').exp() has a huge absolute rounding error. Switching to decimal would fix that.
@Dennis I don't think there's a nice way to get around the cases where one has little error but the other doesn't. I think you could have a "decimal" type in addition and rely on the user knowing what they are doing?
@trichoplax I'm not concerned with how literals should be encoded atm. Ideally, it would be a numeric library that could be used in a variety of languages, and - if I manage to make it fast enough - not necessarily for golfing.
So I've been working on a BF/Sesos solution to codegolf.stackexchange.com/questions/89241/…, and I was wondering if anyone had a good idea on how to set the pointer to the maximum of a bounded set of cells
@trichoplax It isn't. It's just very slow since adding two rational numbers involves GCD and whatnot, and numerator and denominator get huge rather quickly.
@Dennis I see now. That does sound painfully slow. Could you get any speed up by storing the numerator and denominator as prime factorisations? Hope I'm not going to just repeat the previous conversation - I'll go and read it...
@trichoplax Heh, sympy is known for a lot of things, but speed.isn't one of them. I'm not sure how to deal with prime factorizations when it comes to transcendental functions though...
Or plain rational.
I guess I'll just make the base variable and let it default to something sane.
@Dennis I guess for cases like that you might have to recalculate the prime factorisations from scratch, so it might give a speed up for summing the inverses of the first 10,000 natural numbers, but a slow down for other tasks. I don't know how to get around that...
Right now, real and imaginary part are stored independently. Although I guess absolute value and direction would have its perks. A pain for sum/difference though...
What's the distinction between non-symbolic and symbolic numeric types? Google is giving me advanced psychology papers, which I don't imagine are the right answer.
@trichoplax No. Even rational may be too much. I just want a single numeric type that can be used for everything (integers, non-integers, complex numbers, characters), so creating a language that can deal with all of them becomes easier.
@StevenH. In SymPy, exp(1) is just exp(1), not a numeric approximation of the value. Unless you ask it to evaluate, all computed numbers are represented exactly.