Reputation
30,108
Next tag badge:
480/400 score
65/80 answers
Badges
3 66 147
Impact
~298k people reached

Jan
25
comment Is the divergence between Round and setting an explicit precision intentional or a bug?
As to the extension, yes, you should round at the end with the min of the input precisions (or use SetPrecision perhaps) instead of using the automatic precision tracking of Mathematica. Because the Mathematica precision tracking does not work as you intend (but in a smarter way that makes sense in general)
Jan
25
comment Is the divergence between Round and setting an explicit precision intentional or a bug?
Or, for that matter, 100``0 which doesn't require rounding up shows as 1.0*10^2
Jan
25
comment Is the divergence between Round and setting an explicit precision intentional or a bug?
So, 54``0 with an error of 1 rounds to tens just like an error of 3.16 (54``-0.5). Both give 5.*10^1. But as you said, misteriously, when you get an extra digit by rounding up, it seems weird (even buggy?): 96``0 is 1.*10^2, 96``0.001 is 96. and 94``0 is 9.*10^1
Jan
25
comment Is the divergence between Round and setting an explicit precision intentional or a bug?
In general, it seems the absolute error should be strictly smaller than the least significant display digit unit, so an accuracy of 0 (or error of 1) demands that the last digit displayed is "tens" (just as if the error was 9, 8 or 4). So the 0 accuracy is the borderline value. Try it with any number that doesn't get a new digit when rounding up, or that doesn't lose all significant digits getting 0 (both weird cases)
Jan
25
comment Is the divergence between Round and setting an explicit precision intentional or a bug?
As to the display digits, there are some cases that confuse me, particularly when it has no digits to display and chooses to display 0.. Others really seem buggy, such as 96``0 displaying as 1.*10^2 instead of 1.0*10^2
Jan
25
comment Is the divergence between Round and setting an explicit precision intentional or a bug?
As to 2.5`1*2.1`1 displaying 0 instead of a red output, also seems weird, but it is still a display problem only. It should give about 5.25 with a 40% uncertainty aprox (less precision than a significant digit)
Jan
20
revised OutputResponse and RecurrenceFilter failure in WM10 with increased precision datasets
edited title
Jan
18
awarded  Yearling
Jan
16
comment “Meta-matching” patterns
How would you owant to deal with expressions with things like x[Verbatim[2/;True]] or x[HoldPattern[2]]?
Jan
16
revised What is the purpose of Internal`LocalizedBlock?
added 2 characters in body
Jan
16
revised What is the purpose of Internal`LocalizedBlock?
added 874 characters in body
Jan
16
revised What is the purpose of Internal`LocalizedBlock?
added 75 characters in body
Jan
16
comment What is the purpose of Internal`LocalizedBlock?
Ok, I now can test this. Let's see
Jan
16
revised What is the purpose of Internal`LocalizedBlock?
added 75 characters in body
Jan
16
comment What is the purpose of Internal`LocalizedBlock?
@Mr.Wizard how I think it works. Does it fail easy tests?
Jan
16
revised What is the purpose of Internal`LocalizedBlock?
added 12 characters in body
Jan
16
answered What is the purpose of Internal`LocalizedBlock?
Jan
12
comment A simple experiment to understand CUDAFunctionLoad
Thanks for the clarifications. I used to more or less get CudaFunctionLoad, but I haven't used it in a while, and for a few more days I am far from my dear home PC with a CUDA GPU to try. Hope someone answers soon :)
Jan
8
comment A simple experiment to understand CUDAFunctionLoad
You should post a more concrete question. What result where you expecting and why and where does it differ from what you get?
Jan
8
comment What are all the named Mathematica styles?
Great find!! :)