Wolfram Mathematica

12:01 AM

Mathematica code will run even if it's written procedurally

When I started using MMA, that's how I coded, because it's more logical (for a non-programmer)

@termrewritica Well, what CHM said. That applies to you since you're new. I learned Mathematica in the old days, so the books where I learned graphical ways from were... different.

CHM

@termrewritica have you just started using MMA?

term-rewritica

@JM and also unforunately, out of print?

J. M.

@termrewritica Just checked; yes.

term-rewritica

@CHM: the state of my matheamtical knoweldge consists of the first 74 pages of Power Programming w/ Matheamtical Kernel

CHM

12:04 AM

Ha! That book was mentioned some days ago I think... as being out of print

I'm currently reading the Cookbook - very interesting for a beginner

term-rewritica

the mathemtica guidebook on symbolics look really intense

like there are pages and pages of equations

J. M.

It demands a good measure of stamina, yes.

term-rewritica

very different experience from reading "JavaSript: The Good Parts"

CHM

@termrewritica I think you'd benefit more by starting with easier material. Not that I'm an experienced user, but I've been through the initial "wtf" phase, and I can assure you that you have plenty of material in the Mathematica Documentation to keep you occupied for some weeks.

I have to say, before commencing Mathematica, I had only dabbled in C and python - with the usual HTML and PHP, but nothing really profound.

term-rewritica

nah, I really like the style with example + exercises every few pages

[away]

CHM

12:08 AM

@termrewritica I was talking about the Guidebook for Symbolics

@JM can you buy Mathematica online at WRI's store?

J. M.

@CHM Not where I live, unless I pay a fair amount of sundry fees. Luckily I know the lone reseller where I live.

CHM

Ok. It was removed from the store for me, but apparently Sjoerd can still buy it. Was checking if I was the only one.

J. M.

...and I have to leave now. See y'all later.

CHM

Bye

term-rewritica

12:26 AM

Intersting, matheamtica is 1-indexed for elements, i.e. [[0]] gets the type of the object

term-rewritica

12:57 AM

as I look over matheamtica's syntax

(in particular, the notion of varName_Type) for function arguments, and how it breaks from convention, it strikes me as absolutely obvious

that Mathematica could only be invented by someone with the courage/arrogane/brilliance/ego to deviate from the norm

Ajasja

7:22 AM

@termrewritica Well, pascal had varName : Type style declarations since 1970's.

Szabolcs

12:52 PM

@Jin If you need us to send you any other mathematical shape, or vector drawing that is easiest to generate programmatically, just let us now!

3

Oleksandr R.

1:21 PM

@termrewritica not the type, but the Head. Heads are part of every expression, but Mathematica is essentially typeless (or has only one first-class type, whatever way you want to look at it).

rcollyer

2:15 PM

@termrewritica as @OleksandrR pointed out, there is only one type, and that is expression. Everything in Mathematica is an expression, and how it behaves is determined by its DownValues, OwnValues, SubValues, and UpValues plus whatever attributes are attached.

Hola, @RM!

R.M

@JM Could you double-@ jin so that he sees the above?

@rcollyer hola señor!

rcollyer

@RM Now that I've exhausted most of my Spanish vocabulary ... Hi.

R.M

lol

rcollyer

Long time no talk to.

R.M

yeah, I thought you were spending time in the thesis bunker

rcollyer

2:18 PM

Sort of. The past couple of weeks have been mostly a bust in terms of work. Today, though, is shell script day. Yay! :P

R.M

is just syntax sufficient here? I thought we had a tag for such questions, but I could be wrong

rcollyer

syntax works for me.

R.M

> Questions about built-in operators and other special characters of Mathematica [...] should use this tag.

yeah, I guess that works. Somehow seemed strange being the only thing there

rcollyer

somethings can only be stuffed into one pigeon hole. Although, I pity the pigeon that is stuffed into multiple holes at once. :P

R.M

or the hole that has to take many pigeons

That sounded dirty

rcollyer

2:22 PM

:)

@RM In many ways.

Szabolcs

Hello everyone!

How's the weather in California? :)

R.M

Fantastic!

It couldn't be any better

Afternoons are better spent on the beach, which is about 100ish meters from my office

Szabolcs

nearing 40 deg Celsius here ... I'm not handling it too well.

R.M

Hello @stevenvh

Mr.Wizard

Welcome @steven

stevenvh

2:24 PM

Hi Wizard, thanks for your time so far

Hi @stevenvh

@RM ,,|,,

no problem.

Looks like we're not alone

:-)

rcollyer

@stevenvh Not always, no. For a while there, we were among the chattiest of SE sites.

stevenvh

2:25 PM

Hi Szabolcs

R.M.

Mr.Wizard

I'm trying to get this answer working for Steven. If any v8 users want to test that for me I'd appreciate it.

CHM

@RM Wow.

R.M

@MrWizard I concur with Steven.

stevenvh

Wizard, I'm quite new to MMA, so it's possible that you let me try things I don't understand yet, like Sequence. So bear with me :-)

Mr.Wizard

@stevenvh Indeed. It's quite busy now. (a good thing)

@RM thanks for the confirmation.

R.M

2:28 PM

That was actually my first thought and my draft written yesterday actually uses Sequence to splice it (and use the fact that it colours both the same)... but I wrote that without mma and when I tested it, it didn't show up which is why I didn't post it :)

Of course, I was trying with ConditionalExpression and not Piecewise

Mr.Wizard

@stevenvh would you please do:

plot = Plot[. . .];
InputForm[plot] /. Line[x__] :> Line["<<>>"]

and paste the result here?

I'd like to see what is being generated.

@RM If you've already given this a try it is probably not workable. :-/

stevenvh

Which version do I put in the Plot?

Graphics[{{{}, {}, {Hue[0.67, 0.6, 0.6], AbsoluteThickness[3], Line["<<>>"]}}},
{AspectRatio -> GoldenRatio^(-1), Axes -> True, AxesOrigin -> {8., 0},
Epilog -> {Thickness[Large], Dashing[{Small, Small}], GrayLevel[0], Line["<<>>"]},
GridLines -> Automatic, PlotRange -> {{8, 18}, {0, 1.5}},
PlotRangeClipping -> True, PlotRangePadding -> {Scaled[0.02], Automatic}}]

Mr.Wizard

pardon me, the one with Sequence[ cond, cond ]

stevenvh

That's the one

rcollyer

@RM Just tried with using Which instead, but it doesn't want to cooperate with Sequence.

Szabolcs

2:31 PM

@rcollyer Unevaluated@Sequence[] ?

rcollyer

@Szabolcs that doesn't work either, the second line (the important one) doesn't show up in the plot.

Unevaluated@Sequence[
Which[V < 12, V^2/344`],
Which[V > 12, V^2/344`]
]

Mr.Wizard

@stevenvh It truly seems to not recognize it. Okay, different approach. I remember seeing a form of Piecewise that did not plot the ugly lines as shown in my answer at present; I thought it was Piecewise[{. . .}, Indeterminate] but that doesn't seem to be working this morning. Can anyone help?

rcollyer

@Szabolcs is what I used.

Which alone works, but it doesn't keep the same style.

Mr.Wizard

Hm... I think I see the problem with Indeterminate but I don't know a fix.

R.M

Vanishing function works

rcollyer

2:36 PM

@RM Not sure I understand.

R.M

@rcollyer Piecewise[{. . .}, ##&[]]

Szabolcs

Who is High Performance Mark? Does he have an account here under a different name?

stevenvh

Guys, I don't mind using my own method if there's no real alternative. Just a pity that I have to patch the color manually

rcollyer

@Szabolcs no he doesn't. He just has never been active over here, but was somewhat active on SE in the mma tag.

R.M

@Szabolcs he has an account here, but I was surprised that he deleted his answer here once it was migrated over

Szabolcs

2:38 PM

He did suggest migrating questions here before.

I think it's better to suppress my curiosity

Mr.Wizard

@stevenvh Ah ha! I think I'm on to something now. Please try this in v8:

Plot[
  {V^2/144`,
   Sequence[
    Piecewise[{{V^2/344`, V < 12}}, I],
    Piecewise[{{V^2/344`, V > 12}}, I]
    ]
   },
  {V, 8, 18},
  PlotRange -> {0, 1.5},
  PlotStyle -> AbsoluteThickness[3],
  GridLines -> Automatic,
  Epilog -> {Thick, Dashed, Black,
    Line[{{{12, 0}, {12, 2}}, {{8, 1}, {18, 1}}}]}
  ] /. {sty__, x_Line, y_Line} :> {sty, y, Dashed, x}

stevenvh

wait

rcollyer

@MrWizard works on v8.0.4

Mr.Wizard

:-D

rcollyer

Clever.

R.M

2:40 PM

heh, nice

stevenvh

Perfect!

Mr.Wizard

OK, answer update time. Thanks all.

rcollyer

Welcome.

Oleksandr R.

@rcollyer how long do you have left? My submission date is the 28th of September. Exciting times!

rcollyer

@OleksandrR I have to get everything wrapped up by the end of August. I have a job offer that I'm in the process of considering, so it puts pressure on things. That and my adviser is getting pissy.

@OleksandrR Good luck! And, I hope you're prepared for the snake wrestling portion of the defense!

3

Mr.Wizard

2:46 PM

@steven you should make use of this chat room; it's great for getting help working out simple problems, or formulating more complex questions.

R.M

@rcollyer oh good! Congrats on the job offer! Where is it at?

Mr.Wizard

And just having fun. It's a great crowd here, IMHO. :-)

rcollyer

@RM that is confidential until I accept. I'll let you know one way or the other later this week.

R.M

@rcollyer sure

Oleksandr R.

@rcollyer haha, I had never heard that one before! Good luck to you as well. The viva voce I hadn't thought too much about so far... need to get my thoughts down on paper first!

Mr.Wizard

2:48 PM

@steven I just realized that you're #1 on EE. Impressive.

rcollyer

@OleksandrR :)

@RM sent you an email.

R.M

@MrWizard Something with Sequence doesn't sit well with Indeterminate, for Piecewise[{...}, Indeterminate] by itself plots fine

Also, the same reason why ConditionalExpression fails with Sequence. When the condition evaluates to False, it returns Undefined

Mr.Wizard

@RM Yeah I figured that out after asking. The presence of Indeterminate kills the whole group: Plot[{1 x, Sequence[2 x, Indeterminate], 4 x}, {x, 1, 10}]`

(versus 3 x in place of Indeterminate)

stevenvh

@Wizard thanks

Oleksandr R.

@rcollyer I particularly like: "Q: Could the snake kill me? A: That almost never happens." Very reassuring!

rcollyer

2:56 PM

@OleksandrR as long as you never admit to faking any data, you should be fine. Heard about a guy who did that on literally the last question that was going to be asked. Needless to say, he did not pass.

Mr.Wizard

@OleksandrR I don't know what you're reading but it reminds me of the joke about distinguishing grizzly bear scat...

rcollyer

@OleksandrR This is oddly relevant to that.

12 mins ago, by rcollyer

@OleksandrR Good luck! And, I hope you're prepared for the snake wrestling portion of the defense!

Mr.Wizard

Thanks.

R.M

@MrWizard there's plenty of that too in a defense

rcollyer

@RM don't forget the poo flinging!

Mr.Wizard

3:00 PM

"Although if you get a poisonous snake, it often means that there was a problem with the formatting of your bibliography." lol

Oleksandr R.

@rcollyer wow, that is terrible. Both to fake the data, and then to admit to it. My data aren't easy to fake, but fortunately as I run the final analysis pass (while also making the figures...) I'm beginning to realise that actually these experiments I did were actually self-consistent and could, with effort, be used to tell a story. Bit concerned about the analysis itself though as it's almost entirely empirical, laser plasmas being rather complicated and difficult to understand.

rcollyer

@OleksandrR For me, I discovered a bug in pictures that are up on the ArXiv, which changes this enough that I have to rewrite that section.

Also, relevant. wow!

Oleksandr R.

@rcollyer wow, shocking! (Both of those links.) One would hope that higher standards would apply when it comes to anaesthetizing people or convicting them of crimes.

rcollyer

@OleksandrR one would hope. The anesthesia guy worked very hard to hide his fraud, but the procedures at the FBI seem to be related to inadequate vetting of the procedures. And in those cases, there is more of a time pressure to get things done, so I can understand (but don't like) why it would happen. Of course, it means an expensive review of a huge number of cases, and a fair amount of people's lives destroyed. :|

Got to run. Later all.

tkott

hi folks, I'm trying to put together a good set of line styles using AbsoluteDashing for 10 unique lines. 10 would be ideal, but I'd take 5 or more. Does anyone have a good reference for doing this in MMA? I could always use Origin as a template, but I figured this was a good place to start

R.M

3:15 PM

10 lines? o_O

Oleksandr R.

@tkott No idea, sorry--I'd probably use PlotMarkers or choose a different style of plot. But if you figure it out for 20 lines, I'll be interested to hear about it... I have some datasets that need to be plotted with 20 curves on the same axes.

R.M

Is that common in your field?

Oleksandr R.

@RM referring to me? Well, yes and no. I'm doing an experiment that I'm not aware too many other people have ever done before. There are other ways of presenting the data, but will not be as clear.

tkott

well, I did say "ideally" 10. They are generally spaced enough that a rotating set of 5 or so would be fine

as for "common" I would say probably not

R.M

@OleksandrR It was to both of you, I guess... I've usually been told that any more than 4 is a no-no and colours do not count (i.e., don't use colours to distinguish the lines)

tkott

3:25 PM

I've tried plotting things as a 2D density plot, or contour, and it ends up being more confusing than just putting things onto the same plot.

Oleksandr R.

@RM Currently I'm using the default ColorData[1, n] colors and the default PlotMarkers to distinguish between 5 lines on the same plot. Fortunately the lines don't (or shouldn't) cross, and there is a clear trend, so visually it isn't as bad as you might think.

R.M

@OleksandrR Oh, yes. In that case I can see that being not that much of a problem. A lot of my plots have had lines cross each other multiple times, so it could become hard to follow

rcollyer

3:39 PM

@RM that is debatable (the colors thing). I've seen data where colors are the absolutely perfect way to differentiate the lines; each of the 15 - 20 data sets were at different temperatures, and considering our visual acuity is usually pretty good, it would have illustrated the trends very well. Obviously, it should also be combined with differentiating the line type, but colors should not be out of the question.

R.M

@rcollyer I agree, but it depends on the journal. Some specifically disallow colours, unless if you're also willing to pay colour printing charges (default it is b&W print and colour online)

I don't mean disallow colours... colours are allowed, but you can't say "compare blue curve vs red curve" or anything like that in the text. It has to be something like "dashed curve vs solid curve" or if you've placed labels next to the curves, with that

Oleksandr R.

Not only do some journals charge extra for color printing, but also it annoys colorblind people. The latter is the more important consideration IMO.

R.M

Yes, that too.

acl

@OleksandrR so you're in the UK then? or is it viva voce elsewhere?

Oleksandr R.

@acl yes, I'm in the UK. I presume other countries have the viva voce as well, but I don't know which.

acl

3:46 PM

@OleksandrR I do not think it is called viva voce in Belgium, the Netherlands, Germany and Greece, at least, not colloquially

Heike

4:02 PM

@MrWizard This reminds me of an old game show on television. During the finale of the show one contestant had to navigate a maze while wearing a blindfold (and large wellington boots) while the other one gave directions.

Gustavo Bandeira

Guys, what function can I use to multiply all the elements of a list.

Like {1,3,6} -> 1*3*6

Heike

@GustavoBandeira Times@@{1,3,6}

Gustavo Bandeira

Thanks, I made it another way:
l = Select[Range[0, 10], PrimeQ]
Product[l[[x]], {x, 1, Count[l, _]}]

Heike

@GustavoBandeira That should work as well (although I expect @@ to be faster).

Also, the usual way to count the number of elements in a list is to use Length.

Gustavo Bandeira

@Heike Yep, I didn't use it because I didn't know how to use @@

@Heike Doesn't Count works for it?

Heike

4:17 PM

@GustavoBandeira It does but Count compares every element in the list with the matching condition which makes it a lot slower that Length.

Gustavo Bandeira

@Heike Oh, ok. Thanks for the advice.

Heike

@GustavoBandeira @@ is a useful function to learn. a@@b is short for Apply[a, b, {1}], so if you're not comfortable with the infix notation you could use that instead.

Gustavo Bandeira

@Heike I was reading about it now, I got afraid because it has lots of pattern signs - i still dunno how to use the patterns, but i'll learn with Schifrin's book.

Heike

Pattern signs? Do you mean slots (i.e. #)?

Gustavo Bandeira

@Heike Yes.

@Heike # /@ @@ and all that

CHM

4:28 PM

/@ and @@ are shorthands, not patterns. Map and Apply, respectively.

Gustavo Bandeira

@CHM I imagined that, It was just a naive assumption. Haha

CHM

Hehe

Heike

And # are slots which are used in pure functions. Something like (#1+ #2^3)& is equivalent to Function[{a, b}, a+b^3].

Gustavo Bandeira

@Heike Is it possible to do (#x+ #y^3)& so that the x and y can change?

Heike

4:47 PM

@GustavoBandeira Why would you want to do that?

Gustavo Bandeira

@Heike I have no idea.

@Heike I imagined something like: [{a, b, c, d, e, f, g, h, i, ...}, a+b^3].

Heike

@GustavoBandeira You need something in front of the square brackets. Do you mean Function[{a,b,...}, a+b^3]?

Gustavo Bandeira

@Heike Yes.

@Heike But I noticed now it's pretty dumb.

rcollyer

@OleksandrR and @RM I agree with that sentiment, but if done correctly, it will work even there. As to printing in color, the journals I've worked with allow print to be in B&W while online is in color. So, it is a mix of both color and symbols. But, presentation wise, I'd go with color any day.

Heike

@GustavoBandeira It's not that dumb. The number of arguments of a pure function isn't fixed, so you can do (#1+#2^3)&[1,2,3,4] is you like. The third and fourth argument are simply ignored in this case.

Gustavo Bandeira

4:57 PM

@Heike Let me elaborate this

Heike

@OleksandrR I agree. I wasn't really aware of issues with colourblindness until I collaborated with someone who is red-green colourblind. Since then I try to avoid plotting green and red curves in the same plot. Apparently red on on a dark background is hard to distinguish too for red-green colourblind people.

Gustavo Bandeira

5:17 PM

I was thinking something like this:
While[x <= 5,
AppendTo[l, (# x + # (x = x + 1)^2) &[1, 2, 3, 4, 5, 6, 7, 8]]; x++]

Oleksandr R.

@GustavoBandeira what's that supposed to achieve? # is a shorthand for #1 and refers only to the argument 1 in the first place.

Gustavo Bandeira

@OleksandrR It would perform the operation and append it's value to a list, then loop for a determined number of times.

Oleksandr R.

@GustavoBandeira Well, I appreciate what it actually does (calculate in a very obscure and inefficient way Table[4 x^2 + 2 x - 1, {x, 1, 3}]) but is that what you really wanted here?

Gustavo Bandeira

@OleksandrR Not actually, I was just thinking about an application for, and as my Mathematica skills are in a "While and Append" phase, that's the only thing I could imagine.

Oleksandr R.

@GustavoBandeira okay, but my point is, your function has 8 arguments and it never touches the last 7 of them. Is that intentional?

Gustavo Bandeira

5:29 PM

@OleksandrR Yes. I just wanted to ilustrate what I thought, in a possible application, I would use all - if needed.

Oleksandr R.

@GustavoBandeira oh, I see. Okay. Well, I would suggest that in cases like this you can often use Listable functions to good effect: e.g. x = Range[3]; 4 x^2 + 2 x - 1.

Heike

Instead of #3 you can also write Slot[3], so if you want to vary which argument to use you could do Slot[x]

Gustavo Bandeira

@Heike Yeah, i kinda noticed that the #x is not very functional

Oleksandr R.

@Heike have you ever used a variable-numbered Slot? I haven't, but it seems like it could lead to some interesting tricks.

Heike

@OleksandrR I haven't. It also doesn't really seem to work as I thought.

Gustavo Bandeira

5:33 PM

@OleksandrR Yep. Thanks

Oleksandr R.

@Heike how so? (n \[Function] Slot[n] &)[2][1, 2, 3] seems to "do what it says on the tin"?

Heike

@OleksandrR Interesting. I tried something like (Function[n, Slot[n]] &)[2][1, 2, 3] but that doesn't work.

Ah, silly me, that should be Function[n, Slot[n]&][2][1, 2, 3]

Rojo

6:14 PM

@OleksandrR That's interesting, humm

Stack Exchange

Wolfram Blog: Digging into River-Aquifer Interactions with Mathematica

posted on July 11, 2012 by Wolfram Blog Team

As a PhD candidate in civil engineering, Diego Oviedo-Salcedo needed a computational environment that he could use to not only explore the abstract concepts within his civil engineering research, but also to present and communicate his findings to his advisor, peers, and decision-makers. His solution: Mathematica. Mathematica‘s enhanced built-in statistical analysis capabilities allow Ovi…