On Tuesday, April 18, 2023 at 6:27:50 PM UTC-6 William Stein wrote: On Tue, Apr 18, 2023 at 5:15 PM aw <aw.phon...@gmail.com> wrote: > In high-precision environments like RealField(1000), Sage should *definitely* use the math semantics, because physics people, or engineers, or any other applied type folks, have zero use for 1000 bits of precision in anything that they do.
Search for the word "physics" here: https://www.mpfr.org/pub.html Bill, I flipped through several of those papers, and from what I could see, the amount of physics work that uses more than 256 bits is still small. It may get more popular in future, but at the moment it's small. The amount of work being done in 1000 bits is very small, in 10000 bits is probably zero. So what I said was correct to zeroth-order, and probably to first order. Maybe I start to be wrong at second order? When talking about how things are in the world, or what the ideal thing to do is when there are many options to choose from, it's hard to be exactly right. The most important thing is to get the zeroth-order right. I think all of my posts here have gotten the zeroth-order right. Almost all of the replies to my posts, including this one of yours, have nitpicked something at first-order or second-order and ignored the main, zeroth-order point I was making. I'm not trying to insult you here, I'm just making a sociological observation, like a sociologist might make. You and your people have a *very strong* tendency to nitpick details instead of staying focused on the big picture. Here's one part of that big picture: Alpha: enter "e^1.1", press "more digits" 3.0041660239464331120584079535886723932826810260162727621297528605... Sage: RealLazyField()((e^1.1)).numerical_approx(200) 3.0041660239464333947978502692421898245811462402343750000000 What do people think about Sage, when they do some checking and confirm that the Alpha answer is right and the Sage one is wrong? Do they think good things about Sage, or bad things? That's big picture stuff. Your devs have been nitpicking me to death saying it can't be done, performance considerations, not possible in Python, etc etc. Wolfram did it, so why not you guys? It's funny to even say it that way, "Wolfram did it", because it's such a tiny, trivial thing. I'm pretty sure you guys could do it, you just don't want to, because of various first-order and second-order reasons. The zeroth-order should override the first and second-order. Because the zeroth-order matters more. Anyway, I'm just trying to help here. Like I said in my first post in this thread, I like the idea of Sage. I want it to do well. But it's pretty hard to be a booster for Sage when I look at the above two outputs, Alpha vs Sage. Fix it or don't fix it. Up to you. By the way, if you haven't already, you really should read all of my posts. There's a lot of good stuff in them, from various angles. I have to get back to my main work. Probably we will run into each other at some future date. Later alligator, -aw -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/sage-devel/b1f39e56-e61c-49eb-a9d5-dda45a326c8cn%40googlegroups.com.
