And we’ll throw in RStan and PyStan for free!
Well, if you’re an individual (not an IT department) at an academic institution, Revolution Analytics will match that price. Our software has been free to academics for almost 4 years.
Excellent—thanks for letting us know, that’s good news.
Apologies if this is a stupid question, but other than speed / parallelization is there a good reason to choose Revolution R over plain old R?
I tried reading your website but wasn’t sure.
This “academic use only” thing is popular. I also see it used frequently by academics or academic institutions who want to reserve the right to sell it to those they think might pay for it. You also see it in special pricing or licensing for students. You also see some licesnes, like Intel’s MKL compiler for C++ that are free for not-for-profit uses, or research uses.
The two main issues I have with this gambit are:
1. I might not be an academic forever. It drove me crazy when I moved from academia to industry and all of the “free” things I was using all of a sudden became paywalled or very expensive. If you’re a student, you should carefully think about whether you want to be able to use the software you wrote when you’re not a student any more. And think twice about submitting to paywalled journals — they may look free when you’re on the campus network, but they’re hard for those on the outside to read.
2. Code I write using “free to academia” licenses isn’t free to share with non-academics. There are lots of researchers at places like Google, Microsoft, etc., and contrary to popular beliefs, they don’t have piles of gold in the corner of their offices to spend on these things. Another way of saying this is that the code you write is not free in Stallman’s sense of the word “free”.
Is Revolution R open source? I still have no idea how they’re managing this scheme with a basis in GPL software — I do know they said they ran it by a lawyer who said it was OK, but in my experiences, if you ask three lawyers something you get six answers.
Open source is not synonymous with free, by the way. My previous full-time gig was at Alias-i, a natural language processing software company, and we released our LingPipe software under a royalty-free open-source software license for research purposes, but it wasn’t free in any of the senses understood by the open-source software community. We got a lot of grief from the open-source die-hards over that license. I’m just saying I understand Revolution R’s desire to make a buck off their work.
Stan, by the way, is also not free in Stallman’s sense of the word. The core C++ code and the command-line interface are licensed under the BSD, which imposes no GPL-style copyleft restrictions. RStan is GPL-ed due to links with R. PyStan is GPL-ed due to philosophical scruples on the part of its lead developer.
The issue Stallman has with non-GPL licenses is that someone may take your code, extend it, and then not share the extensions with you. Hmm, sounds like what Revolution R is doing with R, so maybe the GPL isn’t strong enough? Without being able to get other people’s extensions, versions of your code become no longer free. Stallman makes a good point here.
There’s a really nice discussion in the comments section of a Slashdot article of all places, FSF’s Stallman calls LLVM a terrible setback. I’m just glad LLVM rocks — we use it all over for Stan with the clang++ compiler.
What do think of Wolfram’s gambit to make Mathematica free (bundled) to run in a Rasbery Pie?
It grants “a non-exclusive license to use the Product solely for personal or educational purposes on a Model A or Model B Raspberry Pi computer.” See: https://www.wolfram.com/legal/agreements/wolfram-mathematica-raspberry-pi.html
And presumably the Pi isn’t so fast? That’d also make it a kind of crippleware.
I’m just not that interested in using something where I can’t write programs that are free for other people to execute.
i understand bob’s comment, “I’m just saying I understand Revolution R’s desire to make a buck off their work.”
but i also wanted to link back to an old post on christian’s blog. it is by ross ihaka.
“I also think one other change is necessary. The license will need to a better job of protecting work donated to the commons than GPL2 seems to have done. I’m not willing to have any more of my work purloined by the likes of Revolution Analytics, so I’ll be looking for better protection from the license (and being a lot more careful about who I work with).”
I’m familiar with that post. As I said, I have no idea how they’re managing what they’re doing and still respecting the GPL. But one has better things in life to do than worry about other people’s intellectual property.
We’re giving Stan away and would be happy if other people could make a buck off of it. I’m getting paid a salary out of public grant money and if I wanted to go back into the commercial software world, I’d do just that.
I really don’t like how universities have gotten into the intellectual property business, but that’s another blog post…
I see it as price discrimination. Ability & willingness to pay varies a lot between academics & industry.
Another way to look at it is that student / faculty sales are loss leaders to push market share / product penetration / free advertising.
The first point is known int he business world as “market segmentation”. Like international editions of textbooks or “crippleware” versions of CPUs. I just loved Hal Varian’s book Information Rules, which explains market segmentation and why crippleware made sense for Intel, as well as many other topics on marketing and econ of interest to software developers.
As to the second point, when I worked at Bell Labs, we didn’t have any more of a budget for this kind of thing than when I was a professor at Carnegie Mellon. When I worked at a two-person company, we had even less budget. Yet I still find Columbia buys site licenses to MS Office and still negotiates deals on MATLAB. I know academics use Stata, which also isn’t free.
The third point is the motivation for my warnings to students. Don’t get locked into something you won’t be able to afford later!
I got one of the individual academic Revolution R licenses and promptly never used it for exactly these reasons. Now I work at Redfin, where I will use R. Glad I didn’t start depending on something that I’d have to ask Redfin to shell out thousands of dollars a year to make work.
“Code I write using “free to academia” licenses isn’t free to share with non-academics. There are lots of researchers at places like Google, Microsoft, etc., and contrary to popular beliefs, they don’t have piles of gold in the corner of their offices to spend on these things. Another way of saying this is that the code you write is not free in Stallman’s sense of the word “free”.”
This is such a great point. The investment I’ve made in free software (R, emacs, gnu toolchain, maxima, jags, stan, etc) is paying off extremely well in industry. It makes such a difference to be able to just go to the download page, get your tool and solve the problem rather than having to spend political capital harassing managers, making purchase orders, and trying to convince teams to adopt/invest in the software you want.
I also agree with bob about the raspberry pi mathematica offer – I too tend to be skeptical of half-free / educational offer software bait. I already got burned once working in matlab’s ecosystem when I had “free” access to it. In retrospect, could have been better spent investing in python or something else.
Isn’t this a bit like saying “We’ll sell you our washing machine $999 cheaper than that lawn mower?”
Stan & R aren’t really close substitutes are they?
That’s right. Someone might well want both Stan and Revolution R, or only Stan, or only Revolution R, or neither.
But Stan is a thousand times more expensive than R from CRAN.
Mail (will not be published)