Slashdot Mirror
Mathematica 6 Launched
Ed Pegg writes "Wolfram Research has just released Mathematica 6. That link, in addition to the usual 'dramatic breakthrough' material, has an amazing flash banner that simultaneously shows a thousand mathematical demonstrations all at once. The animations came from the Wolfram Demonstrations Project, a free service with 1200+ dynamically interactive examples of math, science, and physics, all with code. For the product itself, much is new or improved, with built-in math databases, improved visualizations, and more."
its version 6.06.06
An article about the demonstrations is at0 5_02_07.html
http://www.maa.org/editorial/mathgames/mathgames_
That a dollar in nickels needs $1.88 in metal to be made is surprising.
check out SAGE: http://modular.math.washington.edu/sage/
I think Mathematica is ok but I still prefer Maple. The coding seems to be more regular and fluid.
...does it run Rule 110 in Linux?
As Mathematica seems to be transitioning more and more into the realm of visualization, I wonder when Wolfram Research will add support for 3D-accelerated rendering. A lot of things I've drawn in Mathematica have been somewhat limited by the software's non-accelerated output capability.
This message printed on 100% post-consumer recycled electrons.
Can it process the number 09 F9 11 02 9D 74 E3 5B D8 41 56 C5 63 56 88 C0?
:(){
I hope this one is faster than Mathematica 5. Also they finally seem to be able to handle graphics fast, and added sliders so that you don't have to recompute. Overall, seems to kick Maple's ass =)
From the section on 3D improvements, a whole one click away from the summary-linked page:
"Seamless optimization with graphics hardware on all computer platforms."
How does it compare to software like Matlab and Maple?
"I don't know what this does but the pictures are cool"
Engineering is the art of compromise.
Quick question... after reading their PR stuff I still wonder:
/Yes, I run it in Linux
Does it still use those ugly, outdated widgets in the front-end?
"Does it run Linux in Rule 110?"
--
It still costs a bazillion dollars, which puts it far out of reach of the average person.
Mea navis aericumbens anguillis abundat
Come on, people. This looks like slashvertising if I've seen it...
Mathematica is an absolutely fantastic package. The symbol manipulation (derivaties, integration, etc.) are outstanding, the graphing is rich (though the options are obscure), and the data-manipulation is great. It's a very useful tool when doing serious engineering and science, plus it's fun to play around with.
However, I recently ran into one of those "top 10 reasons why proprietary software is annoying" situations. I hadn't used Mathematica in awhile, and wanted to go back to some old code and re-run some analysis. However in the meantime I had migrated from Windows to Linux. No problem--the install disk has the installer for Windows, Linux, and Mac all in one. Great. So I install it in Linux and then get the annoying "you must register this product to use it." (On Windows it gives you two weeks before locking out, but in Linux it won't open unless you enter the code, which changes with each new hardware installation.) The online automatic registration said I had to contact them via email. So I did. Eventually got the reactivation code. Turns out it didn't run properly on Linux. The controls were clunky and I couldn't get individuals block to execute (though I could re-execute an entire workbook.) Okay, no problem--I have a Mac laptop. So I load it up there. Then I have to go through the reactivation process again. Another email, more waiting. Eventually get it running.
My point is that I had alot of difficulty getting my (legitimately purchased) copy of Mathematica to actually work for me. I was in a hurry and just wanted to run some code quickly. This 10-minute operation turned into a 4-day ordeal, at the end of which I was very frustrated. It really reminded me the great advantages of programming in standards-based languages, that have open-source implementations. If the code had been written in python (using the Gnu Scientific Library), I would have been able to run it without hassles, and I could send the code to others, knowing that installing Python (on the OS of their choice) was always easy.
I don't want to turn this into a stereotypical OSS vs. proprietary rant... but this very recent experience with Mathematica has left a bad taste in my mouth--and I was previously very much a Mathematica evangelist!
I see that the student license still goes away when you graduate. Yuck.
--
Given enough personal experience, all stereotypes are shallow.
Mathematica has a killer engine (kernel), but a lousy UI, and it costs a shocking amount of money. Mathematica was one of the first pieces of software to scan your computer's MAC address and serial number while you entered the activation key, so it could not be installed on more than one computer (this after the $250-$1000 price tag). A student can get the castrated $250 version, but the real version is considerably pricier. Wolfram's treatment of his users is as distrustful as Micro-Suck.
Why can't the FOSS community beat Wolfram at this? Octave, Maxima, Yacas; they all fail miserably in comparison. The UI for Yacas is so idiotic that the function that transposes a matrix is Transpose[], a nine-character entry for an operation that a real mathematician may use a few hundred times in a given program. At least Mathematica is smart enough to use T (or at least it was when I last used it, at 4.0). Why can't we do better than this?
The best UI of any CAS was the UI for the built-in graphing calculator for Mac OS 9. The current version, NuCalc, is available for Mac and Windows, but it is proprietary, and there is no plan for a Linux/UNIX version. The FOSS community can put a UI like NuCalc over a Maxima engine, use MathML and/or LaTeX for the syntax (like LaTeX input, MathML output). Use code from GNU TeXmacs for the UI, but include the beautiful way that NuCalc simplifies fractions and radicals (and algebraic equations) by clicking on them with the mouse. Brilliant. And possible. Future generations of math and physics and engineering grad students will thank us.
"Indeed, it is wise never to consider any form of electronic data as final." --Arnold Robbins
title says it all.
000010011111100100010001000000101001110101110100 1110001101011011 11011000010000010101011011000101011000110101011010 00100011000000
1011 1010 1101 1100 0000 1111 1111 1110 1110
So It seems they've finally caught up to the 21st century....
--
Given enough personal experience, all stereotypes are shallow.
I hope they fixed the horrible bugs with the exponential integrals. IIRC this one was completely wrong
} ]
Integrate[Exp[-a*x+i*b*x]/(x^2+y^2),{x,0,infinity
Maple can't do them either, so it's not like I'm just bagging Mathematica. Exponential integrals have a branch cut in the complex plane and the programmers never seemed understand it. Not that Mathematica was capable of simplifying the resulting sum of logarithms, because it wasn't, but at least it could give you something correct.
Here's wishing for the best from a program that doesn't get supported with bug patches. I reported this years ago. Yeah it's a bug, but no it won't get fixed in my copy. Why would I upgrade otherwise?
For other open source options, see Comparison of computer algebra systems on Wikipedia.
It seems to crash and display a DMCA notice. What the hell is a "processing key" anyhow?
and Hart?! I always knew maths was evil.
In[1]:= RealDigits[Pi, 13256278887989457651018865901401704640, 10]
Out[1]= {{3, 1876991704476780800150180628325083636,
9189391413076810481201219232926836119,
5384063602086586924188333802024283231,
2513453230179012144646634684477494849,
3068214336797764299997264261540154903,
10748763127124397137593816210347368190,
4318173104433516756252170067941446829,
5632687311927575967371556397320072633,
3878443159091356825173975449160258985}, 1}
Why? Because when it does its symbolic algebra thing, it largely acts as a black box. You've got no idea how it got its answers. So you can't rely on it.
So, if you're using it to figure out any symbolic algebra out that's part of research that you're later going to publish, at best it's useful for finding things which you then have to show by hand anyway.
Any sufficiently advanced technology is indistinguishable from a rigged demo
--Andy Finkel (J. Klass?)
"Why can't the FOSS community beat Wolfram at this?"
It's hard to find highly-qualified people willing to work for free. Yes, I know some people get paid to write FOSS, but it happens only in those cases where some other means of earning money is possible. This is the exception rather than the rule.
"It still costs a bazillion dollars, which puts it far out of reach of the average person."
Cost is irrelevent if one lacks the knowledge to make use of such a tool. Those who need it can afford it.
While you're checking out Mathematica, consider taking a look at the major open source computer algebra projects:
Axiom: http://wiki.axiom-developer.org/ (formerly known as Scratchpad) was developed at IBM as a commercial system, sold to NAG, and released a few years ago as an open source program.
and
Maxima: http://maxima.sf.net/ (descended from the pre-commercial Macsyma codebase) was maintained by William Schelter for many years and he obtained permission to release it as open source. Sadly, he passed away a few years later but the Maxima project has grown and now has many active contributors.
They won't have the glitzy graphics or army of specialized packages Mathematica boasts, but they also don't cost $1500 and (theoretically) can be audited for correctness all the way down to their foundations. I regard the latter as very important for people trying to do scientific research with computer algebra tools, and what's more no commercial company is required for their survival (the story of Macsyma is a very good object lesson.)
Maxima is the more "engineering" oriented of the two systems and will probably make more sense to Mathematica inclined users - it can use gnuplot, run on Windows and has a decent GUI called wxMaxima: http://wxmaxima.sf.net./ Axiom is more oriented towards being "strong" mathematically - it takes more getting used to and has very ambitious goals for long term mathematical research. It is attempting to become a literate program in the tradition of Knuth's TeX system. It doesn't currently have the interfaces to familiar tools the way Maxima does.
Both systems are already very powerful and while there are many bugs to work out progress is being made. If you're shopping around for a CAS and are interested in open source systems, I highly recommend checking them out.
(Bias disclosure - I have been a (minor) member of the Maxima project and am currently interested in/doing a little work on/with Axiom, in case the URL in my info doesn't give it away.)
"I object to doing things that computers can do." -- Olin Shivers, lispers.org
Sigh, it's not that at all.
The reason why no-one has bothered to make free software for this niche is that it is so fucking boring.
How we know is more important than what we know.
Is $1000 actually expensive? I can imagine that reproducing Mathematica in a months time would be a bit of a trick, so for an individual, maybe it isn't that expensive. If someone who makes a decent US salary used a license for a couple of years, it would only have to save them a couple of weeks to be worth it(so it could increase their productivity by ~1% and be a net win).
And I understand that if it were Open or Free that it could be the product of many free months of effort and be a win for its users, but the notion that hundreds of dollars for software is always unreasonable(it may not ever be preferable...) is a bit tiresome.
Nerd rage is the funniest rage.
Also the number of people who are able to contribute significantly to such project is very limited. This is on the border of several areas - pure abstract mathematics, computer science and engineering. How many qualified LISP programmers you can find nearby? How many of them are also good at abstract algebra? Very few such people exist in the world, partially because mathematicians tend to hate computer tools in their abstract work. That's why most of the CAS systems were created by physicists in response to their practical needs.
Apart from Maxima, which is free software, if somebody wants to contribute, please have a look also at Axiom CAS http://savannah.nongnu.org/projects/axiom, which IMHO gives a very nice and strict approach.
Another C++ based project is GiNaC http://www.ginac.de/
I clicked on the Demonstration Project link, then browsed through the list of demos and decided to try the Monty Hall Problem demo.
It brings me to a flash application which lets me experiment with the problem by clicking on doors and then seeing where the prize is. Actually, it doesn't. It gives me two options: I can download a "live version", or I can watch a demo of someone else clicking on doors and seeing where the prize is. Hello? This is flash, it's already interactive! Gah..
Repton.
They say that only an experienced wizard can do the tengu shuffle.
My favorite part of their site is the quote from the FAQ:
"Q:Is FreeMat 100% compatible with MATLAB? What about IDL?"
"A:No. FreeMat supports roughly 95% (a made up statistic) of the features in MATLAB."
Mathematicians making jokes about made up statistics, hee hee :-)
Please help metamoderate.
When I checked the price, it was $2500. To many people, that's a lot of money.
Mea navis aericumbens anguillis abundat
Points taken. However, the student version of Mathematica is "only" $150, not $250 USD, (for the 5.x versions, anyway) and it isn't crippled in any way except for the printing of "Mathematica for Students" when one prints a notebook. ALL of the functionality of the full professional version was there. Of course, I don't know (yet) about this new 6.0 version.
You've got a mind like a steel trap, inflexible and rusty.
Doh, wrong story, please mod -R -1 ../The illegal SVG vector paths!/ or rm -R ../The illegal SVG vector paths!/
Yeah, it's a lot of money. It's way more money than I would spend to 'have it around', but if I was doing stuff that made me think "Gee, if only I had Mathematica" almost everyday, I would give it some serious thought. I just don't see any reason to focus(so much) on the fact that the scarcity is artificial when the value is often astronomical.
Nerd rage is the funniest rage.
I just spent 30 minutes trying to find/download/compile a working linux svg viewer just to be greeted by the site of the goatse man's gaping asshole?! thanks a lot cocksucker. I hope you get raped by the GNAA.
oss can't even get a _reliable_ clone of f---ing matlab, which is ALSO braindead enough to have transpose() instead of t().
:)
oss has, however, gsl and pdl (perl data language, turns perl (and thus perl shell) into a quasimatlab) and etc. etc. etc. This is what oss seems to be good at (and damned good I add), at least from the available empirical evidence. The only almost-user-friendly general-use scientific software is R. The only example of user-friendly input engine (along the lines of wysiwig input, &c.) seems to be openoffice which is an opened-up project by a bunch of traditionally-closed-source engineers. Notably, it does have trouble playing along with standard open-source development methods. Hmm... coincidence?
The thing is, everyone who needs friendly input seems to be able to afford a commercial package (usu. cheap for students), or learn to use gsl/c++ or whatever a bit lower-level. The thing is, once you learn enough programming to put together a friendly package, you've gone way past needing a friendly package anymore.
Wow! That is one cool number, alright! Are you a member of the Math Professionals Association of America?
Open Source: I'll show you mine if you show me yours.
Thats the fucking reason why missing _payment_ is a limiting factor.
Because, you know, people are amazingly good at doing boring stuff as long as it gets them a shitload of money on their bank account...
HI O WISE PRINCE. WHT TOOK U SO DAM LONG?
"The reason why no-one has bothered to make free software for this niche is that it is so fucking boring."
Apparently not to the people who keep demanding that it be free.
1) I wouldn't say anyone is good at doing things that are boring
2) Mathematica was written by people who find this stuff the most interesting shit in the world.
The problem is not that there is no-one who finds it interesting enough to write.. the problem is that there isn't enough people who find it interesting - and the result is useful to people who are not interested in writing it. So the people who find this crazy interesting jump at the chance to write it and tell everyone they know what they are doing (who just look at them like they're talking about stamp collecting) and then someone comes along and says "hey, ya know, we can sell this."
That's how scarcity happens.
How we know is more important than what we know.
I like to tell myself that given enough time I could rewrite almost any application that I have ever used. I say almost because of one crystal clear exception: Mathematica. The power of that program is incredible. I'm not too keen on the user interface (it seems backwards to anyone raised on MATLAB), but nevertheless what Wolfram has accomplished with Mathematica is awe-inspiring.
This account verified sig-free since..., uh, never mind.
http://reference.wolfram.com/mathematica/guide/Sum maryOfNewFeaturesIn60.html
"Full-featured source-level debugger, including breakpoints, watchpoints and stepping."
If you haven't used Mathematica, you have no idea how badly the debugging sucks (prior to the new version.)
Quattuor res in hoc mundo sanctae sunt: libri, liberi, libertas et liberalitas.
I think the Student version is only valid until you leave school; how they tell this, I'm not sure, but given how aggressive they are about requiring revalidation for trivial hardware changes, I wouldn't assume that you could just keep using it forever -- eventually you might need to get it revalidated and they might ask for proof that you're still a student after 10 years.
"Ladies and gentlemen, my killbot features Lotus Notes and a machine gun. It is the finest available."
In Matlab and its OSS clones such as Octave, transpose of matrix A is written A' which is a very effective notation.
As a mathematician, and more to the point an applied mathematician, I can confirm that indeed the FOSS offerings are usually significantly inferior to proprietary solutions.
Maxima, though theoretically powerful, lacks the sane(r) syntax of either Mathematica or Maple. It's also buggier and lacks the sheer breadth of abilities that Mathematica has on hand. I'm not as familiar with Yacas, but if I remember correctly it suffers from much the same flaws. Put simply, there is no open source symbolic language that can realistically compete.
The exception is Octave. Vanilla Octave is only slightly behind Matlab in its capabilities, and indeed performance. However, with the addition of capabilities offered by the Octave-Forge packages, and VTK visualisation via Octaviz(If you can get it to compile), Octave IMHO actually surpasses Matlab in ability.
Octave is an example of how FOSS can do better. The core package can be built on and expanded by others. This has yet to happen for symbolic algebra and calculus packages, probably due to the inherently arcane nature of Maxima and Yacas' syntax. The great FOSS symbolic system has yet to be written, but once it is, I have little doubt that it will be up to and past the standard of Mathematica as time goes by.
Basically, my post is a plea for someone, somewhere to start writing the One True Computer Algebra System form scratch. I'd do it myself, but I'm too busy trying to get my Maxima code to run.
May the Maths Be with you!
Terseness in a programming language is not a virtue - if it was we'd all be programming in APL. IIRC it was Wirth who said that programming productivity is not a function of typing speed. And really, I'm sure the average mathematician can master a decent text editor well enough to do a substitution with a regular expression... or does Yacas force the user to only use a brain damaged editor?
The tyrant will always find a pretext for his tyranny - Aesop
Warez teh warez version? Torrent? Monkeyshine? This is slashdot so I know you know.
pure and simple
I have been using TeXmacs as a front end for Maxima, which is a very nice program but a bit heavy-weight for the machine I'm running on. I'll have to checkout SAGE.
> Also the number of people who are able to contribute
> significantly to such project is very limited.
There are enough people available, but it takes significant work to find and organize them. During the last two years over 40 people have contributed significantly to SAGE http://www.sagemath.org/, which is a GPL'd mathematics software package that combines Maxima, Python, GAP, PARI, and other systems with lots of new code and interfaces to Mathematica, Maple, MATLAB, etc. Also, SAGE has a modern web-browser based GUI.
> This is
> on the border of several areas - pure abstract mathematics,
> computer science and engineering. How many qualified LISP
> programmers you can find nearby? How many of them are also
You point to one of the problems. That's one reason much of SAGE is in Python; the barrier to entry is much lower. Also, with SAGE we've done a huge amount of work to make it easy to transition code from Python to compiled C code (via Pyrex and SageX).
The reason why no-one has bothered to make free software for this niche is that it is so fucking boring.
If you think you're up to it, why not start by implementing the full functionality of Mathematica's FullSimplify routine (one single routine). I can guarantee you it's not boring at all, though you're unlikely to finish it this millenium unless you have a PhD and some serious experience.
Oh, neat. transpose() also works, and it's just what I reverted to when t() didn't do the trick as it does in mathematica and R.
Thanks.
So yeah, I know how to write something that would blow the doors off mathematica and kick matlab's ass... but how can I get two years to work on it without a paycheck??
And I'm not alone... that's why there's no FOSS alternative... ok well there's SciPy and it rocks... those people must be independantly wealthy or all live in communes where they grow their own food or something.
Seriously, SciPy rocks. I think it may already be better than matlab... and mathematica is in it's sights.
Damn your eyes!! I wish I could work with the SciPy people
There are 10 types of people in this world, those who can count in binary and those who can't.
How about A' for the complex conjugate transpose or A.' for the array transpose? The spelled out version is for people who want their code to be obvious to non-MATLAB programmers. Nobody actually uses it.
Not to disparage either Ed or Mathematica (both of which are amazing in their own way), but shouldn't this post have been flagged in some fashion with a note that Ed is an employee of Wolfram Research and that this is thusly at least a semi-commercial post?
-- Math Major, *ouch* ...
I'm actually working on an OSS CAS in the vein of Mathematica and other Math systems I've dealt with. It's not going to be released for some time, but what I have now (Basic support for Symbols/ Eq solving, Symbolic Polynomial Math, Arbitrary precision arithmetic, Sets) is looking pretty good.
Ceci n'est pas un Sig.
Perhaps the best thing is that existing Mathematica packages run incredibly well in the new version of Mathematica, much faster and with awesome graphics!
t(x) := transpose(x)
So yeah, I know how to write something that would blow the doors off mathematica and kick matlab's ass... but how can I get two years to work on it without a paycheck??
If you really think you can do this in two years, you should look for some investment money or hook up with a school and do a grant - you'd have all kinds of fame and recurring revenue from ancillary activities if you did it.
My guess is that Mathematica has more than two man-years into it, but they might just be using the wrong toolset, eh?
My God, it's Full of Source!
OUTSIDE_IP=$(dig +short my.ip @outsideip.net)
What I was doing sounds somewhat different to what you're doing. Mathematica gave me a very simple closed form for a summation I was attempting to do, and it does so in a oompletely opaque manner; I couldn't be confident publishing such a result until I'd proved it myself, in the absence of any understanding of how Mathematica found it (and Mathematica can be remarkably opaque in that situation).
If I've understood you correctly you're using Mathematica (or some other computer algebra system) to perform a known, easy to verify algorithm more times than a human would find practical. I wouldn't have any trouble with publishing results with such an exercise.
Any sufficiently advanced technology is indistinguishable from a rigged demo
--Andy Finkel (J. Klass?)