Goofy

@Nasser I find it dissatisfactory that the last coefficients are unreliable, esp. when expanding about a singularity. For your example which I posted above, I need to expand one degree higher than the degree I want to get the correct term for that degree. Shouldn't that be considered a bug?

deg = 5;
Coefficient[AsymptoticDSolveValue[ode, y[x], {x, 0, deg + 1}], x, deg] -
  Coefficient[AsymptoticDSolveValue[ode, y[x], {x, 0, deg}], x, deg] // Simplify
Coefficient[AsymptoticDSolveValue[ode, y[x], {x, 0, deg + 2}], x, deg] -
  Coefficient[AsymptoticDSolveValue[ode, y[x], {x, 0, deg + 1}], x, deg] // Simplify
(*
(C[1] (39 - 20 Log[x]))/9600
0
*)

@Nasser Umm, your image is missing the backtick for the context. Is that what you really did?

I guess Windows is different?

Charting`$InteractiveHighlighting = False;
Plot[Sin[x], {x, 0, 10}]
Charting`$InteractiveHighlighting = True;

This works for me in V14.1/Mac:

@ThuyNguyen More precise: N@Round[lis, 1/10].

Seems odd that control-equal (W|A) does not know what to do with "Wolfram Language code for a*(b + c)" but does know what to do with the simpler "a*(b+c)".

@Semiclassical Yep, it's a funny bug.

It's kinda like this:

eqns = y1'[t] == 4  y2[t] + 5  a^t && y2'[t] == -y1[t] - 20  a^(-t) &&
    y1[0] == 1 && y2[0] == 0;
sol = DSolve[eqns, {y1, y2}, t]
eqns /. sol /. a -> E // Simplify

@Semiclassical Try this:

Hmm: MaxBend::deprec: "MaxBend->20" setting has been deprecated. Use "Method->{MaxBend->20}" setting instead. I know MaxBend has been deprecated for many years now, but it was suggested by the FE for the autocomplete of Method. (Note that I used Method->{MaxBend->20} and it generates this message.)

@JasonB. Thanks! I always forget to pore through the whats-new updates. :)

@Kuba A thousand syntactic-sugar functions but no DoWhile....

If the integral is very large in magnitude, then convergence is determined by the PrecisionGoal, and changing AccuracyGoal will have only a small effect.

@Brownian_Motion Reducing AccuracyGoal relaxes the convergence criteria (greater numerical error will be tolerated). In particular, AccuracyGoal roughly controls the absolute error (and PrecisionGoal the relative error). If the value of the integral is close to zero relative to 10^-ag, where ag = AccuracyGoal, then NIntegrate will accept a large relative error in the integral.

chronicle.com/article/…

"The University of California Changed Its Math Standards. Some Faculty Aren’t Happy." Chronicle. (Paywall, sorry):