Has someone experience with mod for floating point numbers? Assume the simple example for integers, where calculate mod 2:

Table[{x, Mod[x, 2]}, {x, 0, 20, 1}] // ListLinePlot

Looks as expected. On real data with finite precision, this doesn't work. Let us scale everything with 0.1

Table[{x, Mod[x, .2]}, {x, 0, 5, .1}] // ListLinePlot

@halirutan: Compare NumberForm[BaseForm[0.2, 2], 53] with NumberForm[BaseForm[0.2, 2], {60, 60}]

the issue has to do with the fact that 0.2 has no finite representation in binary

What confuses me a bit is that Mod[1,0.2]<0.2 is False which is not supposed to happen according to the documentation

Just noticed that I posted the wrong expression above - what I meant to write for the second expression is NumberForm[BaseForm[Mod[1,0.2], 2], 53]

@Mathe172 Sorry, it was already very early morning and I had to sleep. Why this happens is clear to me. The question (that I forgot) is, why is this not accounted for? If you look at

Table[{x, Last[QuotientRemainder[x, .2]]}, {x, 0, 5, .1}] // ListLinePlot

you see that this gives the right answer. This function must do some more involved computation.

Btw, the same behavior of Mod can be found when you write a real C program with fmod.

@halirutan: C is more correct: see here - at least here fmod(1.0,0.2)<0.2==true

@Mathe172 Yeah, I had a similar C snippet..

@b3m2a1 I would like to update the install instructions for my packages to use your paclet server, one everything settles down. Do let me know when you consider it stable enough. One thing to change is to update the URLs after the name change (which was a good idea—a short name like "palclets" is better)

@b3m2a1 Could you tell me more about how to add extra metadata without putting non-standard things in the PacletInfo file?

@Mathe172 In JLink this happens as well. The method-names are looked up.

@halirutan it just seems strange that it is so slow, especially compared to 11.2 and to something like ?context`*, which finishes a lot faster on the same context

@Mathe172 @halirutan Try Block[{Internal`$EqualTolerance = 0}, Mod[1, 0.2] < 0.2]

Related Q&A: (48810), (87821), (132890)