Timeline for Is FindRoot wrong about its WorkingPrecision?
Current License: CC BY-SA 3.0
5 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Feb 1, 2018 at 0:59 | comment | added | Max | Fair point about my lazy handling of y[x]. If there's a way you think I can clarify that in the question I could edit it. You could also keep your answer and add a note to help others avoid the confusion in the future. | |
| Jan 31, 2018 at 11:31 | comment | added | Michael E2 |
@Max Ah, I think I didn't realize it was the same y[x] throughout the post. I never do things like y[x_] = Evaluate[y[x] /. sol[[1]]] because it destroys y (you can't use y[x] or the original ODE as symbolic expressions). I overlooked its importance to your question. -- Yes InterpolatingFunction seems to be mishandled. Will delete.
|
|
| Jan 31, 2018 at 5:53 | comment | added | Max | But FindRoot doesn't issue a warning when the function y[x] is insufficient precision. If I understand your answer, you're saying "that's just how FindRoot works", which doesn't seem to answer my problem. My problem is that FindRoot deceives me where it should warn me. I accidentally feed it a function y[x] of insufficient precision, and it glazes over that issue and tells me it found me a very precise answer. That's just mathematically incorrect. Is there a reason I'm not seeing for why the precision of a function y[x] shouldn't be inherited by its argument x during FindRoot? | |
| Jan 27, 2018 at 15:24 | history | edited | Michael E2 | CC BY-SA 3.0 |
Added explanation
|
| Jan 27, 2018 at 3:56 | history | answered | Michael E2 | CC BY-SA 3.0 |