« first day (1620 days earlier)      last day (2854 days later) » 

Meta
Meta
1:25 AM
0
Q: Shall we ask questions which are not really questions?

WjxSometimes some questions will suddenly bump in my mind while actually they have no practice use. Sometimes I'll even 'make' questions according to these ideas. They're great ideas, but they cannot solve present problems. Shall I post these questions? e.g.:this and this are all partially made-up...

discussion
0
Q: How can we clearly state in our title that the question is always open to new answers?

WjxEveryone will be influenced by a question's answer number. When there's already an answer to a question, people will hesitate to answer it. But sometimes we need multiple solutions to a questions, the more the better; or sometimes we need multiple ideas and approach to a question. Usually peopl...

discussion
 
Mariano Suárez-Alvarez
Mariano Suárez-Alvarez
1:54 AM
@Szabolcs I eventually found it :-)
 
 
12 hours later…
Shutao TANG
Shutao TANG
1:39 PM
Hi, everyone. I have a question about the operation in LibraryLink.
 
Sascha
Sascha
2:03 PM
I hope this does not turn out to be a stupid question but can anybody explain to me why something like trainingset = {1 -> "A", 2 -> "A", 3.5 -> "B", 4 -> "B"};
c = Classify[trainingset];
Dynamic[c@RandomReal[4], UpdateInterval -> 1] lets my run cpu at 80%? It gets even worse without setting UpdateInterval to 1
 
 
2 hours later…
Pickett
Pickett
3:55 PM
@Sascha FWIW it has not effect on my CPU. Mathematica 10.3, OS X.
 
Michael E2
Michael E2
4:48 PM
Funny bug in V10.4.1: Reduce[ForAll[n, n >= 2 \[Implies] 1/2 (-1 + 1/n) \[NotElement] Integers], n, Integers] returns False, even though it's obviously true.
 
JasonB
JasonB
5:41 PM
@MichaelE2 Would you consider the bug to be gone if you instead got an error message: "This system cannot be solved with the methods available to Reduce."
 
Michael E2
Michael E2
@JasonB Yep :)
@JasonB Oddly, I could get that error message with Solve sometimes, with slightly different input, which I've now forgotten. (Had to be an equation to solve over the integers.)
Still, you wonder about "proving" things with M....
 
JasonB
JasonB
If you ask `Reduce` if the opposite of your statement is true, `Reduce[Exists[n,
n >= 2 \[Implies] 1/2 (-1 + 1/n) \[Element] Integers], n, Integers]` then you get that error message
(is that the opposite of your statement?)
 
Michael E2
Michael E2
5:57 PM
@JasonB Yes, that's the negation. Well, bugs are often quirky. If it were pervasive, a lot of wrong results would have led to its detection.
 
 
2 hours later…
Szabolcs
Szabolcs
7:57 PM
@ShutaoTANG Yes, that fails. The reason is that P is of type double *, i.e. a pointer to doubles ( = mreal). It can only be dereferenced once.
MTensors use a flat storage, row by row. That means that the matrix {{1, 2, 3}, {4, 5, 6}, {7, 8, 9}} and the vector {1, 2, 3, 4, 5, 6, 7, 8, 9} have exactly the same underlying storage. You must know whether you are indexing a matrix or a vector and calculate the index appropriately. In the first case, the element m[[i,j]] will be indexed as m[3*i + j] in C, where 3 is the number of columns.
m[i][j] simply isn't valid if m is of type double * because it tried to reference twice.
 
 
2 hours later…
Sascha
Sascha
9:48 PM
Can anybody else replicate my issues with cpu usage? I run Mathematica 10.4.0.0 on OS X 10.11.3
 
 
2 hours later…
Michael E2
Michael E2
11:19 PM
@Sascha Yes. M 10.4.1, OSX 10.11.5. Seems like a bug worth reporting. (My guess is that some internal variable is being tracked that shouldn't be. Well, at least I had that given to me as a explanation for something similar a few years ago, with respect to a different function obviously.)
 

« first day (1620 days earlier)      last day (2854 days later) »