07-06-2017, 10:03 PM

07-07-2017, 02:46 AM

That's a really interesting plot. It comes at an especially interesting time for me, because just this weekend I was looking at what the licensing schemes for Mathematica (which I used a little in college) and Maple (which I used a lot in college and loved and bought a copy) were. Both now have a hobbyist category, which they didn't last time I looked. I was really looking for a CAS, for some stuff I want to do. Mathematica was always the aspirational purchase for me; something about it being bundled on the NeXT and my father pointing out that if he had that much money, he'd buy my mother a car...

I did end up getting a hobbyist license for Maple. I wasn't totally sure, and then a combination of nostalgia and a very nice sales rep pushed me over the edge...

But back to the plot. I'm not surprised, especially for research computation. I've been at few research places in the last ten years, and everybody's moved from Mathematica or Matlab (mostly Matlab; they were computer vision folks for the most part, and then some machine learning) to Octave, then to SciPy. I don't know what's happening in the universities. I have a niece going to ASU in the fall to start her undergrad studies in physics, and I am very curious to talk with her about her classes this Christmas.

I did end up getting a hobbyist license for Maple. I wasn't totally sure, and then a combination of nostalgia and a very nice sales rep pushed me over the edge...

But back to the plot. I'm not surprised, especially for research computation. I've been at few research places in the last ten years, and everybody's moved from Mathematica or Matlab (mostly Matlab; they were computer vision folks for the most part, and then some machine learning) to Octave, then to SciPy. I don't know what's happening in the universities. I have a niece going to ASU in the fall to start her undergrad studies in physics, and I am very curious to talk with her about her classes this Christmas.

07-07-2017, 08:29 AM

Ray, thanks for your enlightening answer.

If you search for MatLab/Octave's interest, it seems they are slowly fading too.

The Raspberry Pi "educational" license is a great way to test whether Mathematica is for you. In my case, I have used it for some months to learn Mathematica, and check how well this tool could help us to solve some model simulations at work. Although Mathematica in the Raspberry is painstakingly slow, it helps you to get introduced to this tool. I am very sure it is the top tool for educational purposes and CAS, so I recently approached my (chinese) boss to get some budget for a license, and... I was asked to search for a FOSS solution with a lower â‚¬ntry barrier. It now seems to me an Anaconda Business Subscription might better fit our budget.

BTW, the interest in Mathematica shows the seasonal spikes of the educational demand, but SciPy's growth looks more steady, less trenched in the educational market?

[OFF TOPIC] 73&DX (when I was young, I was addicted to 20 m EA-Pacific long path "very very very" early in the morning ;O)

If you search for MatLab/Octave's interest, it seems they are slowly fading too.

The Raspberry Pi "educational" license is a great way to test whether Mathematica is for you. In my case, I have used it for some months to learn Mathematica, and check how well this tool could help us to solve some model simulations at work. Although Mathematica in the Raspberry is painstakingly slow, it helps you to get introduced to this tool. I am very sure it is the top tool for educational purposes and CAS, so I recently approached my (chinese) boss to get some budget for a license, and... I was asked to search for a FOSS solution with a lower â‚¬ntry barrier. It now seems to me an Anaconda Business Subscription might better fit our budget.

BTW, the interest in Mathematica shows the seasonal spikes of the educational demand, but SciPy's growth looks more steady, less trenched in the educational market?

[OFF TOPIC] 73&DX (when I was young, I was addicted to 20 m EA-Pacific long path "very very very" early in the morning ;O)

07-07-2017, 09:24 AM

For working on high accuracy mathematics algorithms, mathematica has proven essential. Wolfram Alpha usually nearly always sufficient. Alpha is also great for simplification of formula (I'm sure I could have done this by hand when I was back in under and post graduate but I've forgotten way too much and the computer doesn't make careless errors).

A purely numeric package isn't going to be as useful -- modern calculator designs are working beyond the standard numeric precisions (43S, Free42, newRPL).

Pauli

A purely numeric package isn't going to be as useful -- modern calculator designs are working beyond the standard numeric precisions (43S, Free42, newRPL).

Pauli

07-07-2017, 07:06 PM

Luigi,

Nice to meet a fellow ham!

I actually debated Mathematica on the Pi --- I don't have a Pi, and it would be an excuse to set one up. But clearing my desk so I'd have a place to plug it into the one HDMI monitor we have seemed like more time and work (yes, my office is that much of a disaster) than sending the Maplesoft folks a couple of hundred dollars. (My wife would likely disagree... ) I may still do that.

I chose Maple primarily for nostalgia, secondarily for the hopes I'd remember it better because I used it more, and finally because although dated I have a stack of books on using it I can refer to.

You make a great point about the lack of seasonal spikes in SciPy. That's really interesting --- I hadn't noticed that. Of course you're right, Mathematica would show clear trends there.

A couple of astronomers I know also rave about SciPy and NumPy in their work --- it wouldn't occur to them to even use a paid solution at this point. The costs don't outweigh the benefits, and they use Python for a lot of the rest of their computing and pipeline construction. That may be typical in other professions as well.

Paul, that's a good observation about Alpha. I should really play with it more. Another advantage to Alpha is that it's server side --- you don't need anything special to ask it questions on Android, iOS, or Chrome. Good for students and may well become the next "pocket calculator" for serious math stuff on our cell phones, what with ubiquitous[/i] WiFi and cell coverage at most of the places we do math.

Nice to meet a fellow ham!

I actually debated Mathematica on the Pi --- I don't have a Pi, and it would be an excuse to set one up. But clearing my desk so I'd have a place to plug it into the one HDMI monitor we have seemed like more time and work (yes, my office is that much of a disaster) than sending the Maplesoft folks a couple of hundred dollars. (My wife would likely disagree... ) I may still do that.

I chose Maple primarily for nostalgia, secondarily for the hopes I'd remember it better because I used it more, and finally because although dated I have a stack of books on using it I can refer to.

You make a great point about the lack of seasonal spikes in SciPy. That's really interesting --- I hadn't noticed that. Of course you're right, Mathematica would show clear trends there.

A couple of astronomers I know also rave about SciPy and NumPy in their work --- it wouldn't occur to them to even use a paid solution at this point. The costs don't outweigh the benefits, and they use Python for a lot of the rest of their computing and pipeline construction. That may be typical in other professions as well.

Paul, that's a good observation about Alpha. I should really play with it more. Another advantage to Alpha is that it's server side --- you don't need anything special to ask it questions on Android, iOS, or Chrome. Good for students and may well become the next "pocket calculator" for serious math stuff on our cell phones, what with ubiquitous[/i] WiFi and cell coverage at most of the places we do math.

07-20-2017, 03:07 PM

SciPy is open source, which often has two major advantages: cost, and the ability to trace your computations via source code.

http://conscientiousprogrammer.com/blog/...g-answers/

http://conscientiousprogrammer.com/blog/...g-answers/

07-20-2017, 03:20 PM

Thanks for the information.

07-21-2017, 12:56 PM

(07-06-2017 10:03 PM)Luigi Vampa Wrote: [ -> ]Google Trends: interest in SciPy vs. Wolfram Mathematica

Opinions are welcome. Thanks in advance.

Dangerous conclusions ... if your omit the word "Wolfram" and just ask for "Mathematica" you see that interest in "Mathematica" are still about 25 times as high as in "SciPy".

On the other hand, if so more and more people are asking for information about "SciPy" this could mean that the documentation is poor while Mathematica's is good (indeed they are very good). I guess that the queries for "Microsoft Office" programs jumped up considerably when they omitted the printed and converted the locally installed documentation into worthless stuff some years ago.

Also interesting: if you ad the word "Matlab" to the comparison. All the others are dwarfed down to a single line in comparison.

Now add "HP prime" to the comparison with "NumPy" - whoa! Prime kills SciPy?

Statistics is a difficult beast to understand.

Martin