Brought To You By the Letter R: Microsoft Acquiring Revolution Analytics 105
theodp writes Maybe Bill Gates' Summer Reading this year will include The Art of R Programming. Pushing further into Big Data, Microsoft on Friday announced it's buying Revolution Analytics, the top commercial provider of software and services for the open-source R programming language for statistical computing and predictive analytics. "By leveraging Revolution Analytics technology and services," blogged Microsoft's Joseph Sirosh, "we will empower enterprises, R developers and data scientists to more easily and cost effectively build applications and analytics solutions at scale." Revolution Analytics' David Smith added, "Now, Microsoft might seem like a strange bedfellow for an open-source company [RedHat:Linux as Revolution Analytics:R], but the company continues to make great strides in the open-source arena recently." Now that it has Microsoft's blessing, is it finally time for AP Statistics to switch its computational vehicle to R?
Coming soon: Visual R (Score:2, Interesting)
Drag and drop integrals.
Re: (Score:3)
Re:Coming soon: Visual R (Score:4, Informative)
Why oh Why (Score:2, Insightful)
Why good things are always acquired by douchebag companies and ruined to the ground? First Java, now this.
Re:Why oh Why (Score:4, Funny)
Why good things are always acquired by douchebag companies and ruined to the ground? First Java, now this.
Shouldn't you also direct your ire at the people from R who decide that selling the company was a good idea. Do you really think that MS went to them and said
That's a nice company you have there. It'd be a shame if you didn't sell it to us
Re: (Score:3)
Why good things are always acquired by douchebag companies and ruined to the ground? First Java, now this.
Shouldn't you also direct your ire at the people from R who decide that selling the company was a good idea. Do you really think that MS went to them and said
That's a nice company you have there. It'd be a shame if you didn't sell it to us
It's not unlikely.
Re:Why oh Why (Score:5, Interesting)
Considering that this company neither owns the R language or holds any copyright over it it seems highly unlikely.
Re: Why oh Why (Score:1)
"The people from R?"
R is, and remains, open source. This is a company that sells R support and bolt-on applications based on R.
Nothing has happened to the R language.
Re:Why oh Why (Score:5, Informative)
Microsoft acquired a company that provides commercial services for R. It does not own the R project. The R Project is a GNU project and there's no way in hell that the FSF would have sold R to Microsoft.
Re: (Score:2)
It is GPL so they really couldn't sell it to anyone anyways. They could I suppose but it would be meaningless and a foolish move on anyone who would shell out the money for it.
Re: (Score:3)
It is GPL so they really couldn't sell it to anyone anyways.
That's false. Assuming you are the sole copyright holder to all said code you can do with it as you please. See MySQL.
Re: (Score:2)
That's false. Assuming you are the sole copyright holder to all said code you can do with it as you please. See MySQL.
They can sell it, but they can't sell it out from under you. Besides the obvious downside of having to build support all over again, the only down side to forking is not being able to change the license in the future — which is an up side for the user base.
Re: (Score:2)
They can sell it, but they can't sell it out from under you.
Who can't sell what out from under me? Sole copyright owners can sell their code at any time out from under anyone. It's their legal right.
Re: (Score:2)
They can sell it, but they can't sell it out from under you.
Who can't sell what out from under me? Sole copyright owners can sell their code at any time out from under anyone. It's their legal right.
You cannot change the license after the fact.
You can take your GPL project closed, but anyone who obtained your code under GPL license prior to that can continue to use, modify, and redistribute that code under terms of said license.
Re: (Score:3)
This is about R. That would be difficult to do with a list of contributors this long not impossible, well, yeah, probably impossible.
http://www.r-project.org/contr... [r-project.org]
Re: (Score:2)
Just to add, Microsoft cooperates with lots of projects like this. They are comfortable with this model.
However, Revolution Analytics has lots of the stuff that makes large computations viable like multithreaded / big data parallel R. And that wasn't GNU.
Re: (Score:1)
but those libraries are in violation of the gpl. they fail the modularity plugin test.
that they exist makes it less likely R developers will implement parallel support in trunk using another approach. which sucks and means the 3rd party company effectively controlls the direction that part of the development will take.
Re: (Score:1)
Given than 'parallel' is in the packages included in the vanilla R distribution and maintained by the R-core team I don't think know what you are talking about...
Re: Why oh Why (Score:2)
So they're buying a GPL test case in order to try to invalidate it with their army of lawyers and lobbyists to scuttle the Android alliance so they can FUD their way in with WinPhone 10, all while looking like an OLAP services enhancement?
Re: (Score:2)
No they aren't. The AC was wrong. It is a lousy GPL test case since the people with standing to sue Revolution have multiple times already said they don't think what they are doing is a violation of the GPL.
Re: (Score:2)
That's not what they said. What they said was that they considered modules to be data not code and thus not covered by the GPLed i.e. no linking occurred. An explicit statement from the copyright holder that action X is not a copyright violation is a very strong endorsement. Better yet of course would be an explicit written and signed license permitting it, but the statements could and would be considered by the court in a lawsuit.
Re: (Score:2)
The R core team indicated that in their opinion any module code was just data for the GPL-R engine. Also those modules in theory could work with S. So given that Revolution publishes their changes to base R, I think it is a stretch to say they are a GPL violation.
Re: (Score:1)
Oracle != Microsoft
It's pretty obvious why Microsoft has been within the top few spots of the computing industry for so long. Beats the hell out of me how Oracle is still around.
Re: (Score:2)
Oracle invented the commercially viable relational database. That's not a small innovation.
Oracle has consistently pushed the database world with new technologies that really did make a difference on computation speeds.
Lately Oracle has been one of the few vendors to have a a broad range of large enterprise software designed to work together.
Re: Why oh Why (Score:2)
That's nice.
Isn't Oracle's table/column names still limited to 30 characters? I would think a modern, enterprise-worthy database could handle a bit more.
Re: (Score:2)
That's nice.
Isn't Oracle's table/column names still limited to 30 characters? I would think a modern, enterprise-worthy database could handle a bit more.
Even MySQL supports 64-character identifiers for databases and tables.
Re: (Score:2)
That's not Oracle that's ANSI/ISO SQL Standard. Complain to them. Though 30 characters seems pretty long to me.
Re: (Score:2)
That's not Oracle that's ANSI/ISO SQL Standard. Complain to them.
From what I can find, the standard length is 18. That appears to be a minimum, not a maximum.
Regardless, everyone else that matters supports more. Why not Oracle?
30 characters seems pretty long to me.
[Insert obligatory 640K quote here]
It's plenty if you have standards to abbreviate everything and remove all vowels. Of course, when your NMNG_CNVNTN_RQRS_UNDRSCR_CHRCTRS
oops too long
Re: (Score:3)
I'm seeing the same 30 characters for Teradata and Sybase. When I look at the 2008 SQL standard (last version I own) I get totally lost in the notation and I'm just not that motivated, I'm going to take their word for it. As for everyone else that matters I'd say those two matter.
As for it being big enough. Table names can have synonyms and be accessed functionally via. PLSQL. Oracle itself tends to use table names like X12A with another table that uses a descriptor. If you want documentation Oracle p
Re: Why oh Why (Score:1, Interesting)
Java was never good. Unless youre talking about the one in indonesia. Or the one in my mug every morning.
Re: (Score:1)
Easy != good
Re: (Score:1)
Re: (Score:3)
and how many products that are actually good have you seen ruined by companies like Microsoft?
Either ruined or forced into stasis which eventually ruined them. Wolfpack fit in the first category, Visio the second...
Re:Why oh Why (Score:4, Informative)
Why good things are always acquired by douchebag companies and ruined to the ground? First Java, now this.
First, I'd repeat the observation made by many that Revolution Analystics doesn't own R; it simply provides commercial support.
As an R user in business, this seems like good news. Microsoft has been promoting R for some time as an analytic layer to sit over its databases, but people in business are a conservative bunch. I've spoken to many associates in other businesses, and the main reasons that they prefer to continue with SAS is that support, training and consultancy are far more readily available for SAS than R. 'Supported by Microsoft' is a label that may persuade some to shift, especially if it's supported by a genuine expansion of commercial R support.
Re: (Score:2)
There's one other aspect to it that is obvious when you see who the people on MS side talking about this are. It is, effectively, an acquisition by the Azure ML division - the sole purpose of which is to get the "big data" people to come to Azure and pay for burning CPU cycles.
Re: Why oh Why (Score:3)
Re: (Score:2)
Hey, in case it makes you feel any better, Oracle has also incorporated R inside it's database.
So, Oracle is involved with both Java and R.
Maybe both companies can start fighting and introduce incompatible versions! Just like the old days.
buy the competition (Score:1)
This is just another E^3 move.
Re:buy the competition (Score:4, Insightful)
They haven't bought R. R is a GNU Project and still is even after this acquisition of a third party company.
Re: (Score:3)
It's ancient history, but when Microsoft put some money into perl-on-Windows [slashdot.org] development, there were a lot of ruffled feathers and panicky headlines.
It didn't amount to anything even close to "taking over perl", even during the nastier stretch of Microsoft's "embrace and extend" era, but asking people to remember things that happened so long ago is obviously too much.
Re: (Score:2)
And this won't amount to anything either. The company they acquired neither controls the R Project nor holds any copyrights to it.
Re: (Score:2)
Re: (Score:2)
In all honesty, I don't know where it is exactly, but I'm confident that it's where it would've been anyway had Microsoft done absolutely nothing. I'll blame any usage drop solidly on the rise of PHP, Python and maybe Ruby.
Re: What programming language do pirates use? (Score:2)
Re: (Score:3)
"In that respect, sheer size begets evil deeds for some reason."
I have a corporation and so before creating it I studied some aspects of the corporate structure before creating it. That and observation has brought me to the conclusion that there are two factors, at least in the US, which turn corporations into sociopaths:
1) The only real mandate they have is to funnel money to the stockholders, and in this day in age the most powerful stock holders are the CEOs and BODs
2) Avoidance of responsibility is ensh
Re:Not the first time (Score:5, Informative)
Microsoft acquired a company who develops a forked version of R. R itself is a GNU project and is not owned or controlled by either Microsoft or the company they bought. You're hyperventilating over nothing.
Re: (Score:1)
On the contrary, now the forked version of R has more resources than the open source version. Time to freak out.
Re: (Score:2)
The forked version is still GPL.
Re: (Score:2)
Have you considered that, if it did get past the lawyers, then they are respecting it to the extent the law requires them to?
Re: (Score:2)
Lack of social ability at Microsoft (Score:1)
In The Official Microsoft Blog [microsoft.com] there is a lot of corporate-speak, of the kind used by people with no actual interest in a subject who nevertheless want to be considere
Re: (Score:3)
Microsoft is large. Very F**** large. Their development tool division, while it has had some hiccups over the years, overall has been pretty good, devs liked them and they were always pushing to embrace open source. The rest of the company, not so much.
So things like this look weird depending on where you're looking from. If you look at Microsoft the company that makes Windows and Office, this is awkward, they're trying too hard, etc.
If you look at it from Microsoft the company that makes C#, has been pushi
Re: (Score:2)
One thing of note is that this particular acquisition is not DevDiv, it's Azure ML. But Azure ML is, in some ways, even more F/OSS friendly - at least I don't know anyone else in MS running Linux servers in production for user-facing services, and it's where a lot of ex-MSR guys (like, from those labs that were closed) ended up. It's also where all the Python stuff now is.
Then again, after Satya's takeover, there was a strong push from top down to stop treating open source in general and Linux in particular
Re: (Score:2)
Who cares? R is a lousy language, anyway (Score:2)
"is it finally time for AP Statistics to switch its computational vehicle to R?"
No. Absolutely not. R is not a reasonable language for computing: http://r.cs.purdue.edu/pub/eco... [purdue.edu]
Re:Who cares? R is a lousy language, anyway (Score:5, Informative)
I think that's being too harsh. As the paper described in its conclusions of the 3 groups who make use of R, the largest and primary group is the users, people who don't do programming in R, but rather make use of it for generating and displaying statistics in an interactive environment. R is a much better language to work I think if one has to access to RStudio, the gui frontend to R.
Is R a good general purpose programming language in the sense of other programming languages such as C/C++, python, perl, shell scripting, etc.? No, I think it's clear it's not a good general purpose language, but for what it focuses on, namely make it easy to do statistical computations it's hard to beat the language.
For statistical analysis the only competitor I see for it is a mixture of ipython notebook + python statistical modules such as pandas, numpy, scipy, pymc, sklearn,statsmodel, pystan, etc.
Re: (Score:3, Informative)
Re: (Score:2)
Re: (Score:2)
And Python and R can work together (rpy2), with Python being the glue.
Re: (Score:2)
R does have lots of statistical packages available, some of which are not available for Python. But those are often for obsolete statistical methods. When it comes to modern statistics, i.e. large scale Bayesian methods, non-parametric methods, simulations, neural networks, etc., then R is quite deficient in good packages relative to Python. In part, the reason why R is deficient in such packages is because the R language itself is actually pretty poor.
Re: (Score:1)
Re: (Score:2)
Yes, and that combo is head and shoulders above R.
Re: (Score:1)
I think that's being too harsh. As the paper described in its conclusions of the 3 groups who make use of R, the largest and primary group is the users, people who don't do programming in R, but rather make use of it for generating and displaying statistics in an interactive environment.
In other words, it's a good programming language if you don't do any programming, but just feed parameters to analysis and plotting packages. So no, it would not be a good "computational vehicle" for AP Statistics. That is, not if the students are actually learning anything in AP Statistics.
obsolete (Score:2)
R was a great tool relative to other statistical computing tools until maybe a decade ago. It's still better than Matlab, but that's not saying much. There are better options these days, like for example SciPy with Pandas.
SSAS/SSRS (Score:2)
Re: (Score:2, Informative)
Not true. Revolution's version of R is forked from the original version of R, which is and will remain free software (both libre, and gratis).
Re: (Score:2)
The bit about Skype is also not quite true. MS have done horrid/idiotic things with the UI, especially in the mobile versions, but I use the Linux desktop version almost daily and it works just fine for its intended purpose.
Re: (Score:2)
Except that, for example,
- video skype and desktop sharing is not available on Linux with more than two participants
- the Linux GUI is a confused mess
Re: (Score:3, Insightful)
You hear it here first, R the open source programming language wont run on linux as from the next revision.
Why would the GNU Project stop developing R for Linux because Microsoft bought up some other company that in no way controls or holds copyright to the R source code? In what universe does that make sense?
A bit like skype, linux version doesnt really work much ever since that shit company bought that as well.
It's actually the opposite. The Linux client was much shittier before it was bought by Microsoft. It languished far behind other OSes with respect to bug fixes and new features.
Re: (Score:2)
What makes you think GNU has anything to do with R (other than writing its license)?
Re: (Score:2)
There are numerous sources to support that. Two that probably hold some weight are:
http://www.gnu.org/manual/blur... [gnu.org] and http://www.gnu.org/software/so... [gnu.org], both of which list R as a GNU package.
Re: (Score:2)
That only means that the FSF has chosen the package for the mythical "GNU operating system". It's no different from Debian or RedHat making an "R" package. Well, it is different in that the original "GNU operating system" remained eternal vaporware and "the GNU operating system" now consists of a haphazard mix of relabeled and restricted Linux distributions created by others.
Unfortunately, the FSF has a nasty habit of implying that they deserve credit for software whose creation they had nothing to do with.
Re: (Score:2)
http://www.r-project.org/ [r-project.org] also states that "R is a language and environment for statistical computing and graphics. It is a GNU project which is similar to the S language and environment which was developed at Bell Laboratories (formerly AT&T, now Lucent Technologies) by John Chambers and colleagues." So obviously the GNU project itself doesn't do a lot of actual development, though I would expect that they provide some administrative support in some form (perhaps in similar manner that the FSF does fo
Re: (Score:2)
What "administrative support" do you think the FSF provides for "many open source projects"? All they ever seem to want to do is for you to transfer your copyright to them based on bogus justifications.
As far as the R Project is concerned, I don't see them listed as benefactors or supporting institution:
http://www.r-project.org/found... [r-project.org]
Furthermore, the R
Re: (Score:2)
That's pretty good evidence that there is some relationship, you are right.
However, R is developed by a team out of the University of Wisconsin Statistics Department, lead by Doug Bates. I believe that the team at UW, not GNU, makes decisions about the direction for R's future development.
Nevertheless, your original point stands and I agree with you that I don't see that team moving from linux any time soon.
Re: (Score:2)
Not possible, systemd already has the interpreter built in.