Message Boards Message Boards

Why is Wolfram Language free to use, but Mathematica is not?

Posted 4 years ago

Question:

The differences of Mathematica and Wolfram Engine/Wolfram Language are discussed at other places. Like:

Even though I still can't grasp concretely what the differences are.
My Question here is why there are different payment models for Mathematica, WolframCloud and WolframEngine?
It seems like you gain practically very little by paying for Mathematica. (Mostly interface-wise)
I would think of Mathematica to be kind of like an IDE.
But nobody would pay hundreds of dollars just for an IDE.
So what am I missing about Mathematica that makes it reasonable to pay for it?
Why didn't the new business model for distributing Wolfram Language change the business model of Mathematica?
And why isn't Mathematica free to use?

My Background:

I heard of Mathematica ever since I started studying physics 2 years ago and I was always kind of sad that it is so expensive.
I used WolframAlpha for many calculations while studying, but I always found it quite rigid.
Often it just wasn't enough to solve a problem I had.

So I searched for alternatives and since I love Python I started to learn more about SageMath.
But quickly I figured it just isn't as well designed as I'd like it to be.

So I searched for alternatives again and stumbled upon WolframCloud and Wolfram Engine.
I've been using it for a few days now and it's just what I imagined Mathematica to be. (and more)
And I feel like I won't ever have to pay for Mathematica, because I could just use Wolfram Language for free.

POSTED BY: Nico G
4 Replies

I don't understand why people think that Mathematica is expensive. Sure, a full commercial license is not cheap, but if you have one of these licenses, you (probably) have a job and the time saved by using Mathematica makes it worth every penny.

I have been using Mathematica since version 1.1, with a commercial license. Granted, the initial cost was about $800 (from MacConnection, if anyone else remembers those times), and the cost for premier service over the years has been much more.

I did a quick price check, and the cost for a student or home-use license is in the neighborhood of $1 (US) a day. Premier service is less than that.

What you are paying for is, essentially, the cost for the R&D and new ideas that are in Mathematica/Wolfram language that are not part of (most) of the 'free' software out there. Wolfram Language is not like languages like c or Fortran, which have a small core, and the user is expected to add (by purchase or writing code) all the extra bits. When I was writing code in c, (before Xcode), what I purchased was the compiler, linker, and IDE (in CodeWarrior or Lightspeed c, for example), not the language itself. With Mathematica, the core language is not really separable from the IDE -- or as it is called, the notebook interface, plus all the access to data (Knowledge Base) and libraries.

In essence, Mathematica gives me a team of collaborators to work on my projects. I have people who can solve PDEs and Machine Learning problems. Do you think that any of the fancy mathematical algorithms in Wolfram Language would even exist if it were not for the efforts of Wolfram Research. (The same can be said I am sure for the nifties in Maple and MatLab.) This type of functionality is not in R or Python. I am happy to pay my license fee so that these people can continue their work.

My opinion is that if you can afford to buy a computer capable of running Mathematica, the incremental cost -- a dollar a day -- to use Mathematica is cheap. Plus, there is always the Raspberry pi and the free versions of Mathematica on-line, although someone is probably paying more for internet access than it would cost to have a desktop version.

I am also aware of the digital divide in the US (and elsewhere) where a substantial percentage of the population cannot afford a computer or internet access. Wolfram Research cannot wave a magic wand and make these problems disappear, and it is my understanding that the Wolfram Foundation is working to reduce the digital divide.

I really do not understand why people think software should be free. Someone has to write it, test it, document it, etc. People working on bits and pieces of a system (linux, for example) in their spare time, are providing a service, but they cannot hold a candle to what a team of professionals who are paid to work on a project can do. Perhaps people think that clothes should be free, or food. Think of what you do: would be ok if the product of your work was free, which would necessarily mean that you would be providing the sweat of your brow for nothing.

Just because the marginal cost of software is (almost) nothing, does not mean that the fixed costs are similarly small. As software become more sophisticated, the fixed costs for any software, let alone anything a complex as Mathematica is huge, and continues to get more expensive.

Anyway, my 2 cents.

Actually, Mathematica is already de facto free for education thanks to its inclusion with the Raspberry Pi and the emergence of Pi versions that are powerful enough to run Mathematica (A Pi4 with 4GB or 8GB of RAM can do).

It is probably the cheapest fully fledged computer one can get, and is sufficient until the end of High School, except maybe for the most CS/Math inclined students, and for these ones, the school could still throw an additional pool of Remote Kernels in a server !

Another advantage of the Pi is that it is relatively easy to distribute a specific build of the OS for the school, with all the required softwares, file server and email access etc... all securely packaged. Then one can let students be free to use any screen/keyboard/mouse combination that they like (laptop, tablet, mobile phone, even TV sets at home) to access it, within school premises or at home, with VNC or RDP without compromising the school network.

POSTED BY: charles monneron

I am STEM educator at a small private school in the United States and have been a user of Mathematica and now the Wolfram Language for many years. Over the past 4 years I have been trying to integrate the Wolfram Language into our curriculum but the administration will not commit to purchasing a site license. The stated reasons are:

  1. There are free options which can accomplish the computational thinking goals in our curriculum just as effectively as the Wolfram Language. For example Python at Carnegie Melon, Java at Khan Academy, etc. or a general workflow using Jupyter Notebooks with associated libraries.
  2. I am the only person that really knows WL here.

I hope that Wolfram Technologies continues to innovate and thrive as a business and I don't know what the business would look like if they were to freely provide the software for education purposes. Maybe corporate and government contracts would not be enough for the company to thrive.

My opinion is that if Wolfram were to freely provide the software for education, the exposure of this powerful language and programming paradigm would drastically increase the number of people that routinely use it. At some point the reasons stated above would completely go away.

For #1, the WL could be the free popular choice used to train computational thinking/coding. Everybody is trying to create a stand alone coding environment for the purpose of education including institutions like CMU and online platforms like Khan Academy. Wolfram Technologies has had this idea since launch 30 years ago and has indeed developed a most capable version of this environment. My opinion is that the choice to use WL would be the obvious choice. Additionally, the solution to reason #1 would auto fix #2.

I understand that Wolfram Tech is taking measures to increase awareness like WL for developers and the ability to incorporate WL into Jupyter Notebooks. However, WL users like me know that the full experience and power of the language comes from the interplay between front end and engine. So for me WL means both Mathematica FE and the Engine and that is the exposure I want my students to have. IMO, that is primarily what sets WL apart from everything else.

Well, I could take this conversation in many directions, but I think I'll just stop.

JM


Fast forward 10 months:

This year I exclusively used the Wolfram Language in back to back semester long Introductory programming courses. I had the students create their own free account and we used Stephen's textbook "An Elementary Introduction to the Wolfram Language" as our guide. All of the computations were done in the free version of the cloud and students were able to post their notebooks for my review on Schoology, our learning management system. The project was largely successful. My impression as to what is pedagogically possible for free has been changed. The students ran into very few limitations and these limitations did not surface until late in the course(web deployments). We did several small group projects including the building of a Scrabble helper bot and a code breaker for the Caesar cypher.(adapted from the AI League). My school demographics include over 20 countries represented and a wide range of exposure to programming ideas and languages. Some of the student feedback:

  • The Wolfram language is quite "intuitive". "The functions usually do what you think they might do and this helped me learn things quicker"
  • "The documentation is really useful."
  • "I wish we spent more time building user interfaces and games."
  • "EIWL is too focused on math"

We covered Lessons 1-20, 23-32, 34, 36, 38-40, 44. Well, tomorrow we talk about Import/Export(lesson 44) but as is usually the case this will not be the first time we use these functions. I supplemented EIWL with Project Euler challenges from time to time and we needed to Import and curate data earlier.

I am formally a math/physics teacher and if I get the opportunity to do more coding classes in the future, I will not hesitate to use the WL again.

POSTED BY: John Martin
Posted 4 years ago

Nico:

WARNING: Below is my opinion based on my personal experiences. I know that we are all different. In can only speak for myself.

These types of discussions help me to put things into perspective as I put proposals together for how to spend our limited resources. I hope others (including Wolfram) will share their experiences and thoughts as well.

Thanks for the great question Nico.

This is a great question and one I have struggled with with many vendors.

For me, it keeps coming back to the following question:

Does this vendor's business model work for me?

I have to say that I have passed on many great products because the vendor's business model would cost much more than the value I realize given my current use cases. And to be honest, for many years, there were no budgets for tools, so I was often stuck with what we had and what we could get for free. I do not mean to bash any particular vendor, so I'm not going to name specific products or their makers.

While I would love for all of the tools that Wolfram makes were less expensive, I want Wolfram to do more than just succeed. I want them to thrive, to innovate, to continue making my life easier and more productive (if a bit frustrating at times).

I think it is great that the Wolfram Engine is now free. This is a big deal for my use cases. A free engine will allow me to deploy Wolfram-based workflows as easily, and at the same cost, that I can deploy workflows with PowerShell, R, and other tools.

I don't mind paying (though I'd rather pay less) for Mathematica because Mathematica allows me to do much more than just design standard workflows. With Mathematica, I can:

  • Do exploratory analytics in almost real-time with my business partners. For most questions, I no longer have to tell them that I'll have to get back to them later that day or week.
  • Work transparently with business partners without that having to hand them 10 pages of SQL and an Excel workbook full of formulas that even if they could read, the don't have the time to do so.
  • Build simple to complex cross-platform desktop, web and mobile apps that will delight users in hours, not days or weeks or months.

For most of my use cases, putting my data in the cloud is not an option. But I can imagine that for many use cases that would be a great lower-cost alternative. But, of course, you are then reliant on having internet connectivity, which is something that is not always true.

Unfortunately, we are not always able to afford a product just because the benefits will more than cover the costs. We have to be able to afford the product to begin with and we have to sell the product to the decision makers.

For my current challenges, I think Wolfram has struck a good balance:

  • Power Users and Developers need Mathematica (and we only have a few of those)
  • We can deploy/publish solutions via email, PDF, Excel, Wolfram Player, Wolfram Cloud (when allowed)
  • We can automate the publishing results to other tools such as Power BI, Tableau, Google Sheets, Office, etc.
  • A mostly consistent and evolving environment that doesn't require me to learn and keep up with a dozen or more opinionated packages that are all changing (and breaking things) on their own schedules.

I'm looking forward to reading how others feel on this topic.

Stay safe.

Mike

POSTED BY: Mike Besso
Reply to this discussion
Community posts can be styled and formatted using the Markdown syntax.
Reply Preview
Attachments
Remove
or Discard

Group Abstract Group Abstract