Wolfram Mathematica - 2014-10-22

« first day (1008 days earlier) ← previous day next day → last day (3459 days later) »

1:19 AM

I wrote something... but it's going to take a couple of weeks for it to solve the problem. I posed it as an optimization problem, but I can't find a suitable energy measurement. This is too slow: ImageMeasurements[ImageAssemble@Partition[imgParts, 7], "Energy"]

So well, I think I'll leave that problem for now. But it is interesting :)

1:57 AM

@Pickett I'm using as "energy" (gf[ImageAssemble[1,2]] -gf[1] -gf[2]) ...

gf= Gradient filter

works good, too slow too

@belisarius ImageAssemble[1,2] <- what does 1,2 stand for?

@Pickett Any two pieces

that's why it's slow

way too many permutations

@belisarius I see. I assemble the entire image and then compute a measurement on the entire image.

@Pickett Yep. Also too slow. Too many possible results

there must be a way to reduce the viable configurations

@belisarius My attempt is to use a genetic algorithm, i.e. stochastic optimization. But so far it isn't working :)

2:03 AM

@Pickett ping me if you find something :)

I will!

I'm out of ideas

2 hours later…

4:24 AM

@acl Very long story. It's sort of related to study of probabilistic, local layout optimization of data structures. Current state of the data structure layout affects which of those random samples are feasible.

2 hours later…

6:32 AM

Regarding my own comment:

I have been able to narrow the problem to the use of GraphData in the sense that removing those expressions prevents the hang/crash though of course it also breaks the program. This jogged a memory of another question about loading data inside a Dynamic expression causing a hang as the result of curated data being repeatedly (erroneously) loaded. I can't seem to find that now but I think it may be the same problem. I suggest you search for it. — Mr.Wizard ♦ 6 mins ago

Can anyone remember the question to which I refer?

6 hours later…

12:42 PM

@Mr.Wizard That happens a lot to me with love letters

1:42 PM

In case some of you didn't know that this has changed:

7

A: Is there any point forcing a post to Community Wiki after 6 owner edits?

I'm happy to announce that starting today, community wiki is no longer automatically forced on a question for any reason, including this one. This has always been a long standing issue, how this mechanic worked. It was never a comfortable one, even for this team, but it had a job that it was doin...

1:53 PM

@halirutan Is there a box to check somewhere to avoid having your post bumped to the front page when you make an edit? I like to make minor changes and clean up my posts when I have an idea or new info, but I would just as soon not bump the post unless I make a more substantial change.

Live from Champagne 6 minutes to Stephen Wolfram keynote!

I bet it takes quite a while before getting into anything that's actually new.

2:20 PM

@Murta Link?

@belisarius I think the only public info will be short blurbs on the Twitter tag #WolframTechConf.

@belisarius Tw*tter. As usual, I guess Wolfram Technology Conference is mostly under NDA, which annoys me mightily, considering the fact WRI tries to keep it both secret and somehow advertise with its' occurrence.

@dionys No, there isn't, but I would bother about that it is bumped.

@Murta I assume "Champagne" was autocorrect by software keyboard, or such? I pondered several minutes (without any intense thought) the possibility of this actually happening on the Champagne region, and oddity of not mentioning the specific city. :)

This far those tweets seem maybe slightly more open than last year, but completely devoid of new information. Wonder ponder how it'll turn out in the end.

@MichaelHale Thanks! I'm out of the twitter/facebook/whatever thing :)

2:32 PM

@belisarius twitter.com/hashtag/WolframTechConf

@kirma :D

thanks

2:55 PM

> Usually when you think about machine learning it's lots of black magic and a bunch of weird things. -Stephen Wolfram

@rm-rf A NKBM really

... So we decided to wrap it in a black box wrapped in W|A calls wrapped in an enigma wrapped in PR nonsense...

@rm-rf And you can invoke it in infix notation too!

I linked some time ago a Stephen Wolfram talk from some earlier event earlier this year. After 30 minutes, he started to talk about stuff that was sort of new, and not entirely this sort of hubris.

Usually there's something new "hidden" in there...

But it's sort of the same thing as Apple keynotes and whatnot. 80% of the stuff they say is repeated from event to event, with outrageous self-praise.

@kirma Multiple undo announcement

3:03 PM

@belisarius I think it was announced around the time of v10 getting unveiled (?)

@kirma yeah, well, whatever. Another hype-type conference

Session crash recovery (on basis of multiple undo mechanics) was also under consideration. Announcing that would be actually something.

> Everything is symbolic now including times and dates.

@rm-rf What does that mean?

more atoms?

Time to pack bags and look for a different platform/language...

3:09 PM

@belisarius Overhead.

@kirma :D

@rm-rf They will end up forcing you to compute everything in their cloud of petaflops servers

user image

@belisarius What, I see no signs of peta there... only flops.

@rm-rf They are modelling it on RPis, then they will use ReplaceNode[ graph, {node...}] to switch them to relativistic natural-language quantum computers

Did a popularity/interest check with Wikipedia traffic for programming language designers. Wolfram ranks higher than Alan Kay.

1	Ada Lovelace	73015
2	Grace Hopper	38045
3	Dennis Ritchie	32913
4	Richard Stallman	22707
5	Donald Knuth	14087
6	James Gosling	12809
7	Paul Graham (computer programmer)	11777
8	Ken Thompson	11742
9	Bjarne Stroustrup	10161
10	Guido van Rossum	9834
11	Edsger W. Dijkstra	8276
12	Stephen Wolfram	7677
13	Alan Kay	7223
14	Konrad Zuse	7204
15	John McCarthy (computer scientist)	5996
16	Larry Wall	4626
17	Rasmus Lerdorf	4338
18	Anders Hejlsberg	4322
19	Brian Kernighan	4229
20	Niklaus Wirth	3575

what is your metric?

3:23 PM

That's just Wikipedia article page views in the past month.

Lovelace outdid him on hype factor.

@MichaelHale Most of it must be self-refreshes ;)

Lol, maybe

My personal, not so pretty opinion is that Lovelace and Hopper are high on the list only because people explicitly want to treat women by different standards. For "enlightened reasons", that is.

Lovelace actually has a pretty significant spike this month due to Ada Lovelace Day in mid-October commemorating the first computer program. Hopper is pretty close to her normal numbers.

3:32 PM

@MichaelHale Well, that might explain a bit.

@kirma You don't like COBOL, I see

@belisarius I doubt if those visits result just from numerous COBOL coders... :)

I actually saw several programmer friends passing around a link on Facebook blaming male-targeted ads in the 1980s for the decline of women in computer science. The numbers were much higher in the 70s.

I recall an over-the-top sexist Forth ad/book cover from the eighties or so...

@kirma Ha! I used cobol a lot back then :)

3:35 PM

@kirma Here's the chart they showed: engadget.com/2014/10/21/npr-women-in-computer-science-drop

This was it: globalnerdy.com/wordpress/wp-content/uploads/2007/09/…

13 minutes ago "SemanticImport takes unwashed data and converts it to canonical natural language." - @stephen_wolfram #WolframTechConf

what the heck is " canonical natural language." -

?

Life of Forth coders must be... interesting. Also, it involves kinky outfits.

@belisarius Heh

4:08 PM

"Coming: Example Data and high performance computing options that people can use." - @stephen_wolfram #WolframTechConf

Tweeted by WolframResearch on October 22, 2014 at 4:02 PM

It has the words "high performance" in it.

@Pickett it still lacks "relativistic", "quantum", "natural", "new" and "wolfram" ... but they are doing their best

:D

4:23 PM

Wolfram Blog: Wolfram Data Summit 2014

posted on October 22, 2014 by Wolfram Blog

This year’s Wolfram Data Summit brought together innovators in data science, creators of connected devices, and leaders of major data repositories for two days of high-level discussion about challenges and opportunities facing the worldwide data community. This annual Summit offers an exclusive group of thought leaders an opportunity to meet and share insights into new [...]

4:56 PM

63

Q: Computer Algebra Errors

In the course of doing mathematics, I make extensive use of computer-based calculations. There's one CAS that I use mostly, even though I occasionally come across out-and-out wrong answers. After googling around a bit, I am unable to find a list of such bugs. Having such a list would help us rem...

computer-algebra big-list

Some old but interesting stories

@Pickett at which point I usually turn to C or Fortran... :-D

5:23 PM

@blochwave I guess that works... if you can live without free-form linguistic input.

5:45 PM

Is the syntax coloring bug here new in 10.0.1? I hadn't noticed it before.

StringCases["abcdefg", Shortest[x__] ~~ __ -> x]

Shortest is red because it parses as an unknown option.

5:59 PM

@MichaelHale I can't see it in 10.0.0.

They probably already know, but I went ahead and submitted it.

6:36 PM

@MichaelHale I read "engadget" as some sort of "endgadget" so often... :)

A gadget to end it all!

@kirma Hehe, I think humans have made enough of those.

7:07 PM

Has anyone seen the game Scribblenauts? It'd be fun if the System Modeler library was that big. The Scribblenauts models aren't very detailed though, I think. Just a 2D segmented graphic with some basic animation.

I've thought about just taking a list of the most common words and just starting to manually program basic versions of them all. Like Human[] makes a simple stick figure, etc. Then after I have a few hundred trying to write simple stories to see what it makes.

2 hours later…

8:41 PM

@belisarius You always find something to make us smile. :D

9:03 PM

nope

misunderstood

« first day (1008 days earlier) ← previous day next day → last day (3459 days later) »

Transcript for

Oct '1422

Wolfram Mathematica

Welcome! This is the main Mathematica chat room for mathematic...

join this room
about this room

00:00

06:00

12:00

18:00

all times are UTC

site design / logo © 2024 Stack Exchange Inc; legal

mobile