Hmmh... how to efficiently perform least squares fitting constrained to a semialgebraic set?

Welll... I suppose it reduces into a case Minimize should handle reasonably well - I hope.

@PatrickStevens For low level functions like Nest it is unclear what purpose high level code would serve. Nest[...] := .... what would be on the r.h.s.? The function is not written in Mathematica code.

@Pickett I was just wondering how it was done - is it automatic for non-Mathematica code, then, but all other functions can be spelunked with PrintDefinitions?

@PatrickStevens Have you read the spelunking question here?

It is done basically the same way. I would go so far to claim that the PrintDefinitions was implemented after WRI saw how nice it can be done :-)

@PatrickStevens The second answer in the question linked to by halirutan is by the developer of PrintDefinitions. It says that it only show upvalues, downvalues, subvalues, ownvalues and attributes; so with functions that don't have high level definitions it is automatic in the sense that there are no definitions to show. So I think the answer to your question is that all functions can be spelunked and there is no way to prevent that.

Hi everyone, I have the following confusion:

Mircosoft Edge

However, when I open the Internet Explore, this time the wolfram demo could be ran automatically.

Is there any solution to revise this pheomenon?

@PlaysDice It does what it's name suggests: prints definitions of a symbol - though I forgot to add that it is in the "GeneralUtilities`" package; see here.

Ahh, I see there was already an explanation by others.

@IstvánZachar thanks, @Pickett too. GeneralUtilities` requires v10.1. When I upgrade I look forward to exploring!

I'm trying to pass a ! Exists[x, ! ...] as a constraint to Minimize. How do I prevent Mma rewriting Exists into ForAll in this process?

@ShutaoTANG Maybe Chrome supported old browser plugins (if it worked for you before, I don't know how else to explain it), officially the don't support plugins like the CDF player. Other browsers are phasing it out, and this is probably what you are seeing with Microsoft Edge.

@Pickett How is "most popular" calculated on package data?

@TheToad Each time someone clicks on a link his IP is registered as having clicked on that link, and the popularity ranking is the number of visitors (in this sense) who have clicked on a certain link.

One IP cannot be counted twice.

@Pickett Ah nice, thanks

@ShutaoTANG Neither Chrome, nor Edge support NPAPI plugins anymore. Chrome might have a switch to turn it back on ... otherwise just use Firefox while they still support it.

@Pickett Good, so I don't have to worry about clicking too much :)

:)

Did you know Mathematica could do this?

Graph[ExampleData[{"NetworkGraph", "JazzMusicians"}],
 VertexSize -> Scaled[0.01],
 EdgeStyle -> Directive[RGBColor[0.8, 0.2, 0.3], Opacity[0.2]],
 VertexStyle -> Black,
 GraphLayout -> {"EdgeLayout" -> "HierarchicalEdgeBundling"}]

Useless method of approximating pi: Table[With[{e = 2.^-n}, Length@NestWhileList[#^2 + 1/4 + e &, 0, # < 2 &] Sqrt@e], {n, 0, 35}]

Stolen from this video:

I'm trying to build a tool to construct a color distance metric from own visual judgement.

It repeatedly gives two large random color patches, and ten small ones, randomized interpolations between those, overlapping with both. The goal is to pick one that is "at equal distance" from both large patches.

A very simple test case, measuring linearity of greyscale values, seems to reconstruct the perceptual response curve surprisingly efficiently on basis of this data.