Wolfram Mathematica

2:01 AM

a funny bug that took me 3 hrs to find: I wrote If[x>= distance, doThis,doThat] and it turned out it never executed any of these branches becuase I have not initialized distance to a numerical value ! So it was bypassing the whole thing. The THEN and the ELSE parts. oh well. it was only 3 hrs this time. One time I wasted 8 hrs to find one bug. So I am getting better :0

rcollyer

3:51 AM

@NasserM.Abbasi Right. If has three possibilities: True, False, and who knows. A way around this is to use If[TrueQ[ test ], doThis, doThat] which guarantees a true/false answer. However, it hides the fact that test may not be initialized, and will likely result in other bugs. :P

Nasser M. Abbasi

Mathematica needs a good spring function. To use to draw mechanical springs in 2D or 3D. Currently all what I see is little bit of ad-hoc code here and there, and no clear API to use the code. We need a good robust function with good API, at least for 2D case.

@rcollyer, thanks. the variable was initialized and numerical. I just moved code around and the IF statment moved above, and so it became un-initialized. This is the cost one pays for using Dynamic interpreted code I guess vs. compiled languages. One gets the speed to try things, for the price of the user having to be much more carful.

Matthew7.7

4:18 AM

In[42]:= Integrate[1/s, s] + Integrate[-1/s, s]
Integrate[1/(s - 1), s] + Integrate[-1/(s - 1), s]

Out[42]= 0

Out[43]= -Log[1 - s] + Log[-1 + s]

waht's going on? why is the second not 0?

Stack Exchange

4:45 AM

0

Q: Should [math]/[mathematics] be blacklisted tags?

(I had been meaning to ask this a few months ago, but only got around to it today.) Should math and mathematics be blacklisted? Ostensibly, these tags are as redundant as mathematica, but if any of you guys have dissenting opinions, I'd want to hear them.

CHM

5:28 AM

@Matthew7.7 I don't think this has to do with Mathematica ;)

CHM

5:44 AM

Actually, there might be a way to tell Mathematica which solution you want...

Because this result is not false, it just depends on the value of s

Nasser M. Abbasi

there is a branch cut for ln. ln(z) = ln(abs(z)) + K, where K =Pi*I if z<0, otherwise 0, where I is sqrt(-1).

Simplify using assumption z>1 to see.

CHM

Integrate[1/(s - 1), {s, 0, .99}] + Integrate[-1/(s - 1), {s, 0, .99}] Works.

@NasserM.Abbasi Yes that's why Mathematica gives the full form I guess

Nasser M. Abbasi

or can just help Mathematica a little, and take the minus sign out yourself: Integrate[1/(s - 1), s] - Integrate[1/(s - 1), s] , this gives zero now :)

CHM

Oh, heh.

Matthew7.7

6:03 AM

I'm confused. How does this ahve notthing to do with amtheamtic?

I did integrate f, -f

added them up, and Mathematica gave me a non-zero expression

so now, I'm looking at it, and I can't help but wonder wtf

CHM

@Matthew7.7 Well, what I meant was, it's not a wrong computation.

Matthew7.7

so the error is between the chair and the keyboard?

CHM

No no no

Read what Nasser and I wrote

He provided a fix

And I, an approximation.

Matthew7.7

his fix is to move the - sign out

but I need symbolic expressions that I cna plug into other computations

and it bothers me that they're not canceling out on this simple example

CHM

@Matthew7.7 I found it surprising too

But that's a good question, you could ask it on the site.

Why does it choose one symbolic representation over another.

Nasser M. Abbasi

6:33 AM

Mathematica simplifies -1/(s-1) to 1/(1-s) and this causes the -ln(1-s) to come up. To prevent Mathematica from doing this simplification: Integrate[1/(s - 1), s] + Integrate[Hold[-1]/(s - 1), s]; ReleaseHold[%] now the result is zero.

Basically, the result is zero, ONLY when ln(s-1)=ln(1-s). And this is only true when s=1. Special case. Mathematica gives answers for general cases, not special cases.

Transcript for