I like to see this in a Wolfram Blog post and I already hear the first paragraph: "In the new Wolfram Language we implemented this new kind of thing or as we called it complexity speed-up... the harder your problems get, the faster Mathematica will solve them!" — halirutan2 mins ago
I do not know how best to formulate this as a question but the new PositionIndex is horribly slow.
Using Szabolcs's clever GatherBy inversion we can implement our own function for comparison:
myPosIdx[x_] :=
<|Thread[x[[ #[[All, 1]] ]] -> #]|> & @ GatherBy[Range @ Length @ x, x[[#]] &]
Chec...
Now this is really strange. I just realized that on Mathematica V10 the Sin plot isn't gray (and neither is the Plot[null,{x,0,1}]). However, a more complicated plot that I created is gray.
If you don't mind let me try to isolate the problem with the more complicated plot and post back in 5 minutes
Okay, I think I found a minimal example in V10. For some reason a simple plot exports fine. However, when I combine a plot with a legend I get a gray backround.
@JeffDror That is weird, because then the gray background must happen during the plot, because no matter what you plot, in the end it is just a Graphics object which is exported. That's why I tried an empty Graphics first. The export command doesn't see whether you plotted something or you typed it by hand.
@JeffDror No difference at all. OK, I guess you have a workaround for now. I don't know whether setting the printing environment to Printout Gray introduces other things. Anywhere in the style system there seems to be this definition that your printed stuff is gray for legended plots.
Finding this without me sitting at your pc could become a bit tedious :-) I guess when I post this as answer, someone with the same problem might be able to find the source of the problem easily.
@acl I have finished the program and the performance is satisfied. However, I noticed while the matrix is copying back at the end of compile, it use 3 matrix memory. Do you know any method to circumvent it or is there a way to use pass by reference?
In version 10.0, when I leave the Mathematica section idle for some time (of order an hour), the kernel quit automatically. The syntax highlighting is gone and the variables that has defined before lost their values. There is no error message (tried to launch from terminal and still no error mess...
Does anyone have a hint how to improve quality of output for BoundaryDiscretizeRegion[RegionDifference[Ball[], Ball[{0, 0, 1}]]]? I've tried everything obvious and not so obvious I can think of, and results stay essentially the same. That makes impossible to achieve results I would want to see on playing with constructive solids geometry.
@rm-rf I don't think that changes the fact that you can't go knee up to the head. What if basketball players believed that was the natural way to jump for a ball?
In any case, what is hardly debatable is that it was not free kick for Germany. That was weird
@MichaelE2 Yeah. There are of course many things to criticize, if one focuses on "how near we were". But overall I ḿ satisfied, it was a great cup, and the winner was clearly the best team overall throughout the cup
even having trouble with algeria, ghana, usa, or whatever
It was entertaining too. The semifinals only had "candidates" by pure chance
I only pity the stupid rivalry between argentina and brazil
Argh. WRI has both Wolfram Programming Cloud and Wolfram Cloud, Wolfram Cloud supposedly (?) being the thing tied to Mathematica. And I see no option to upgrade Wolfram Cloud subscription (whatever that is).
@Pickett well, I just think that maybe Extract works by calling Part (or some C function corresponding to Part) and that it is totally unintentional that Extract can work with Span in the first place
@JacobAkkerboom I thought Span would be supported in V10, I guess I remembered wrong. Now I don't think I'll report it, since it's undocumented. Thanks for your help.
@rm-rf Good find! I was half expecting v10 Import to do that automatically, but from what I've read here they are still smoothing the edges of Association/Dataset.
Data is, by its very nature, an evolving world of consumption and distribution. Conrad Wolfram recently spoke at TEDxHousesofParliament, giving a 15-minute talk about the evolution of data and how current and forthcoming technologies will continue to play a role.
Will be using fewer semicolons in Mathematica 10. Output forms that I always suppressed in previous version of Mathematica now present nifty little pods containing juicy information.
@Szabolcs. There are quite of few functions that have received this treatment. It would be nice to have a list of them. Don't know how to generate one though. Do you any ideas.
@m_goldberg The only specially formatted object I found that didn't receive this treatment was FittedModel. All the rest that didn't display their full contents look like this now.
@m_goldberg From the top of my head, SparseArray, Graph (with no layout), Dispatch, CompiledFunction, LibraryFunction, LinkObject, OutputStream, InputStream, all colours (RGBColor, etc.), TimeSeries, TemporalData, ...
@m_goldberg Probably not officially ... but somebody is going to spelunk it anyway :-) It has always been possible to create things like this, with Format, MakeBoxes, etc. One has to be very careful though not to let the formatting function evaluate anything in the expression that shouldn't be evaluated. DateObject's formatting had some bugs which did this in the betas. (It's fixed in 10.0 final.)
More off the top of my head: ClassifierFunction, PredictorFunction, NearestFunction, RegionNearestFunction, RegionDistanceFunction, LinearSolveFunction, LiftingFilterData, ContinuousWaveletData, DiscreteWaveletData, ParametricFunction, LinkObject
Hmmh... how to take uniform samples from a region, v10-style? Specifically, one with a higher-dimensional embedding.
This is not uniform:
Point[{x, y, z}] /. FindInstance[ RegionMember[ RegionIntersection[Sphere[3], ImplicitRegion[x + y + z == 0, {x, y, z}]], {x, y, z}], {x, y, z}, Reals, 1000] // RegionPlot3D
Hmmh, that was awkwardly expressed. But the question remains...
@Szabolcs Econometrics. I wanted to do a PhD in econ, but I worked in econ consulting for a couple of years and really liked it. I specifically studied industrial organization / competition policy, and decided to go to law school with the idea of being an antitrust consultant.
@acl Well the interest was in response to my experience in law school. Legal information management is incredibly archaic, and particularly in the US a very real consequence of that is shutting many people out because they can't access basic information about the law. Which in theory should be public.
I guess you already know the first Hippocratic oath statement "
To consider dear to me, as my parents, him who taught me this art; to live in common with him and, if necessary, to share my goods with him; To look upon his children as my own brothers, to teach them this art; and that by my teaching, I will impart a knowledge of this art to my own sons, and to my teacher's sons, and to disciples bound by an indenture and oath according to the medical laws, and no others."
@mfvonh Sorry the prev msg was answering your " but because most of that material is created by judges, it is buried in prose" comment
I actually gave a talk at school about how the law has a lot to learn from medical information management
I basically got booed off stage
It's just arrogance at this point
The armies of (sociological) rationalization are at the gates :)
@acl So anyway I ended up learning Mathematica because it is pretty much unparalleled when it comes to representing and manipulating arbitrary information symbolically
@mfvonh Medical info management is sponsored by the labs. Lots and lots of money there. The law works for corporations but it is (still) not a part of them
If anyone has experience with the new TransformedRegion function would you please look at this? It's a toy solution to this question. If you evaluate it at t=0 it should give a disk. Instead it gives half a disk.
@belisarius I would say that legal information management is poor on purpose. Lawyers' jobs are not actually that complicated; the most complex part is searching for information to make sure someone doesn't sideswipe you with something by surprise. This sounds bad, but it's actually good for lawyers (and judges) because it gives them way more room to hedge and argue about the meaning of this comma or that footnote. In the end legal services are more expensive and lawyers get more prestige.
At this point it would be considerably less expensive to manage legal information using standardized forms, etc. as is prevalent in medicine. Much of that is driven by regulation, or by major market players like insurance companies. The equivalent of that in the law is the court system. They can literally make arbitrary rules about whatever they want -- including what color paper you print things on.
@belisarius Oh yes I meant specifically in the US. Europe in particular is way ahead of us in this respect.
And yes MDs oppose standardization for the same reason lawyers do. They think of their profession as an art and to diminish it through standard processes and forms and computers and all that threatens their prestige and income. Of course there are also legitimate complaints about hampering needed flexibility in medical choices, but at least here the conversation has not even made it to the point where we are actually discussing those types of issues on the merits.
@mfvonh He came from a rich and rightist family. After being elected bishop he "awakened" and turned (somewhat) left, defending the rights of the poor in El Salvador. The poor in El Salvador are 99% of the people