Wolfram releases the free Wolfram Engine for Developers

[Ackbach]

Wolfram made what I would call a gigantic move: they released the Wolfram Engine for Developers - for zero cost!

The idea of the Engine is primarily that you can call and run Wolfram Langauge commands from lots of different sources. I am not the kind of person to automate Mathematica, although I'm sure many others are. For me, the attraction is that you can download it for Windows, Mac, or Linux (might even run on a Raspberry Pi), install it, and run full-blown Mathematica commands on your own computer!

To be sure, the Wolfram Development Platform, now simply renamed to the Wolfram Cloud, has allowed free accounts, and you can execute Wolfram Langauge commands there in a sort of notebook. But the limitation was always that if you had a command taking a long time to execute, it might abort because of not having enough server time. Now that limitation is gone: you decide how long you want it to execute!

You still don't get palettes and other nice display-type things, as those are reserved for Mathematica proper. Moreover, the Wolfram Engine has an even more primitive display than the Cloud. However, you can always ask for the output in $\LaTeX$ format by doing

//TeXForm

at the end of a command.

I predict this will greatly enhance usage of the Wolfram Langauge and Mathematica, as it has always been so amazingly expensive, particularly for commercial users.

Highly recommended!

Note that you are limited to two copies per Wolfram ID.

 

[Wrichik Basu]

One of the main reasons why I chose to learn Matlab before Mathematica is the high price of the latter. The Matlab mobile app is free, and the cost of the software is still affordable. But Mathematica Student is out-of-bounds. (Matlab also has free alternatives, but that is a different story.)

Thanks for the information. It seems that I can finally learn Mathematica too.

 

[Ackbach]

Mathematica/Wolfram Langauge absolutely rules the symbolic world. I think MATLAB is probably the best for numerical calculations, as it makes matrix syntax so easy (it "thinks" in vectors). These days, though, if Mathematica can't do a symbolic integral, I'm usually going to say that it can't be done.

 

[Dale]

This is awesome! I have been a Mathematica user for 23 years now. I am stunned that they are giving it away now.

 

[Greg Bernhardt]

Wolfram is a business, I wonder what their angle is on this? Have they moved on to private custom business/science applications?

 

[Ackbach]

Wolfram is a business, I wonder what their angle is on this? Have they moved on to private custom business/science applications?

It's the old Joel Spolsky routine: 1. Get users for some free portion of your service, because without users you have nothing. 2. Offer some premium payed services related to the free portion, but always keep part of it free.

 

[PAllen]

Mathematica/Wolfram Langauge absolutely rules the symbolic world. I think MATLAB is probably the best for numerical calculations, as it makes matrix syntax so easy (it "thinks" in vectors). These days, though, if Mathematica can't do a symbolic integral, I'm usually going to say that it can't be done.

That last is not true. Not long ago I had an integral that Mathematica couldn’t solve. I did what I considered the most obvious substitution. Mathematica could then solve the rest. I corresponded with them, and they said Mathematica does not actually do substitutions. Instead, it uses a series of very general reductions that “almost always” obviate the need for substitutions or integration by parts. However, once in a while, there are integrals like the one I had, where it can’t get started without help. So never forget substitution, by parts, or partial fractions. You still need to know them.

 

[Ackbach]

That last is not true. Not long ago I had an integral that Mathematica couldn’t solve. I did what I considered the most obvious substitution. Mathematica could then solve the rest. I corresponded with them, and they said Mathematica does not actually do substitutions. Instead, it uses a series of very general reductions that “almost always” obviate the need for substitutions or integration by parts. However, once in a while, there are integrals like the one I had, where it can’t get started without help. So never forget substitution, by parts, or partial fractions. You still need to know them.

I did say "usually", and non-universal generalizations are not disproved by a counterexample. You're quite right that we still need to know our integration techniques; I usually have told my students that they need to know what a technology is doing before they're allowed to use it.

 

[MathematicalPhysicist]

Mathematica/Wolfram Langauge absolutely rules the symbolic world. I think MATLAB is probably the best for numerical calculations, as it makes matrix syntax so easy (it "thinks" in vectors). These days, though, if Mathematica can't do a symbolic integral, I'm usually going to say that it can't be done.

Maple handle the job as well.

I have used both and I much prefer maple for its affordability.
I once asked how much does cost mathematica for a student per year, and they gave me the price of 1500 shekels while maple and MATLAB each per year cost 400.

And for calculations in PDE, linear algebra, analysis I don't see any point in using mathematica.

 

[Dale]

My experience with Maple is about 20 years ago, but it was very unstable back then. It would crash frequently while the then-current version of Mathematica was reliable. As a result I just didn’t trust Maple from a quality standpoint then, and so I never used it since.

 

[fluidistic]

Mathematica/Wolfram Langauge absolutely rules the symbolic world. I think MATLAB is probably the best for numerical calculations, as it makes matrix syntax so easy (it "thinks" in vectors). These days, though, if Mathematica can't do a symbolic integral, I'm usually going to say that it can't be done.

There are many performance tests done on several CAS, some results are published in the google group https://groups.google.com/forum/#!forum/sci.math.symbolic. Time taken to solve thousands of ODE's, integrals, etc. is reported with the % of completion for several versions of these CAS.
So if you look deep enough, you should find probably hundreds of integrals mathematica isn't able to perform that other CAS are able to.

 

[MathematicalPhysicist]

My experience with Maple is about 20 years ago, but it was very unstable back then. It would crash frequently while the then-current version of Mathematica was reliable. As a result I just didn’t trust Maple from a quality standpoint then, and so I never used it since.

The big question is what Mathematica can do that Maple can't?

 

[Ackbach]

The big question is what Mathematica can do that Maple can't?

From the Maple side.
From the Mathematica side.

As a side note, the comparison isn't really what this post is about. I've linked to the free engine, you can download it or not. This post isn't primarily about comparing Mathematica to Maple, MATLAB, or Mathcad or anything else.

 

[eys_physics]

Thanks for sharing.

 

[heff001]

IT-ish blurb -> Since AWS, Microsoft are staying Cloud centric (meaning good old ETL, Data Warehousing, Internet focused...), "Open Source" stuff will fade away. The more companies that provide gifts, such as this, means they will be used more and therefore will likely ride along with the IT gaints (i.e. Wolfram Cloud Notebook).

 

[Office_Shredder]

Today I learned I can get Matlab on my phone! Amazing.

 

[benorin]

I’m no longer a student and I’m not in academia at all. Maple costs $6200 for me. I wanted to do some research on computer-assisted solving of Selberg integrals. Does anyone know a CAS I can actually afford as a non-student/non-professor entity? I’m poor, maybe I could afford a few hundred $ at most.

I emailed a grad student of Doron Zeilberger who I’m hoping will take up the computer-assisted solving mantle for me. I think he uses Maple. His name was [name redacted for privacy reasons by the Mentors]. I know squat about how to program Maple except for Calculus and a DE/linear algebra I did in school 20 years ago.

I’ve emailed several profs look for a co-author now. So far no joy.

 

[PAllen]

I’m no longer a student and I’m not in academia at all. Maple costs $6200 for me. I wanted to do some research on computer-assisted solving of Selberg integrals. Does anyone know a CAS I can actually afford as a non-student/non-professor entity? I’m poor, maybe I could afford a few hundred $ at most.

I emailed a grad student of Doron Zeilberger who I’m hoping will take up the computer-assisted solving mantle for me. I think he uses Maple. His name was [name redacted for privacy reasons by the Mentors]. I know squat about how to program Maple except for Calculus and a DE/linear algebra I did in school 20 years ago.

I’ve emailed several profs look for a co-author now. So far no joy.

The personal edition, per their website just now, is $239. The professional, nonacademic, is $2390. Is it possible the personal would be ok for you? There are free/opensource symbolic and numeric math packages, as well (of course, they usually have a little more up front time investment learning to use them).

 

[Dale]

There are free/opensource symbolic and numeric math packages, as well

I tried Sage Math. It seemed relatively capable. It wasn’t as fully featured as Mathematica, but surprisingly good for free

 

[benorin]

Oh, I must have read the licensing page incorrectly. Wasn’t the personal not cleared to be used for research?

 

[PAllen]

Oh, I must have read the licensing page incorrectly. Wasn’t the personal not cleared to be used for research?

No, you are correct. The issue is ”what is research?” I think if you want to refer to Maple (for acknowledgment) or provide code for it to back up a paper publicly published in some way, that is not allowed. So, if you hope to publish something based on use of Maple, then personal is no good.

 

[PAllen]

Re Maple again, you could buy personal and do whatever work you want. Only if you end up with something to publish, upgrade to professional. This sort of upgrade is explicitly allowed in the license. However, if you dont’t ever want to spend $2390, then Maple isn’t a good choice.