New York Times Article On R!
The New York Times has published an article on R titled, Data Analysts Captivated by R’s Power. My favorite part of the article is the quote from the SAS spokesperson;
“I think it addresses a niche market for high-end data analysts that want free, readily available code," said Anne H. Milley, director of technology product marketing at SAS. She adds, “We have customers who build engines for aircraft. I am happy they are not using freeware when I get on a jet.”
I think the SAS spokesperson made a mistake with this cheap-shot comment. Because R publishes how it operates internally, I have the ability (and right) to see how R operates. There is no way for me to get the same kind of information regarding SAS' mathematical procedures. They are proprietary. They are a secret.
I would rather fly on a plane built by a company that actively reviews the quality of it's software and KNOWS the answers their computers give them are correct. Of course, all computer software has bugs and it is entirely possible that there is an undiscovered bug in R, but the same can be said about SAS, SPSS or any other piece of statistical software. The difference is your right to look at R's code, and evaluate it for your self.
I wish the author had also interviewed someone from SPSS. Newer versions of SPSS include an optional plug-in that make it possible to access R through SPSS. I am not aware of any such product or tool from SAS (but I am NOT a SAS expert). SPSS is spending a lot of time and energy trying to improve the programmability of it's product (pure SPSS macros suck). SPSS has chosen Python and R for this purpose and I assume that a spokesperson from SPSS would have been much more complimentary to R, and the open-source community in general.
This is a fantastic site. I
This is a fantastic site. I really enjoyed to visit the site. Thanks the service provider. colesterol alto
SAS and aircraft mfg
In response to overdetermined's request for specifics of what aircraft engine manufacturers do with SAS:
SAS has several aerospace manufacturing companies as clients, one who manufactures jet engines, another jet aircraft and still others aircraft components. They do a lot of things with SAS regarding supply chain control and other business functions, but I believe the question is more about design and fabrication, i.e. reliability issues.
Most of the reliability applications with SAS revolve around failure analysis, MTBF and that sort of thing. SAS business unit JMP is very heavily involved in aerospace engineering applications. At their web site (http://www.jmp.com/industries/manufacturing/index.shtml) you can see how their analytics, data mining and custom design applications are used for manufacturing process improvement, defect analysis and the like. JMP has two named customer stories (http://www.jmp.com/software/success/aerospace.shtml) on its website that discuss aerospace applications at Honeywell and Georgia Tech. There are several other prominent aircraft and engine manufacturers who have declined to be named.
Please don't use this response to belabor the quote from my colleague about the reliability of jets designed on open source code. She made a statement that was admittedly much stronger than she intended. Sometimes that happens when a single quote results from a much longer media interview.
Re: SAS and aircraft mfg
I promise. I will not use this to beat a dead horse. I was really just curious as to how statistics would be used in the manufacturing process and your links answer that question. I don't know too much about how they build jet engines. It's interesting. Oddly enough, we don't build very many jet engines at this site.
NYT and R
I am in SAS marketing and certainly come to this discussion with that bias. Your comments about the difference between proprietary and open source software are impossible to argue with, but I think both communities have something to offer. Check out this follow-on post by SAS' Anne Milley who was quoted in the Times article. It is at http://blogs.sas.com/sascom/
Re: NYT and R
First, I want to thank you for your interesting comment and the link to Anne Milley's comment on the NY Times article. I also apologize for taking so long to reply to you. I only recently noticed your comment here on the site. Indeed, her blog post sounds much more reasonable than the quote in the New York Times but that is always the risk when talking to a reporter.
I also agree with you that both communities have something to offer. I sat down and thought about this quote (and the content you linked to, here) and my concern is a very simple one. SAS competes with several products, including R. I am sure you can provide me with good arguments as to why your product is better than SPSS, Stata, or some other statistical package. Having never used SAS, I am in no place to judge the merits of your claims, but I would be shocked if any SAS representative made a joke about being glad they weren't on a plane built (in-part) with SPSS or Stata. (BTW, I'm not an engineer, what do they use SAS for when building a plane? I'm just curious.) R is a very good product, and in spite of the fact it is completely free, it manages to provide a comprehensive suite of tools that many statisticians find compelling. Some even contribute their time to making R better. The community gets a little irked when comments like this are made. I don't want to speak for the R community (I am not a developer in the project) but I do think the community would be open to a real competitive analysis between the two products. I am sure the SAS is better at some things. I am equally sure that R is better at some things. I would be interested in learning more about the relative strengths and weaknesses, but as a consumer of open-source tools, I am immediately turned off when I hear cheap shots about freeware that no decent rep would have made (publicly) about a competing proprietary product. I know I would have responded differently if the SAS rep had claimed that SAS was a superior product because it could do:
The additional modules/packages (!install.views) for R are worth discussing. You are right, these are not of uniform quality. Some are clearly better than others. But, this is similar to the situation found with other products, it's just a little harder to find the material. As a SPSS user, I can assure you that there are many freely downloadable programs for SPSS and these are not of uniform quality either. But, they are a quick Google Search away. I am sure the same is true for the SAS community. Anytime a user decides to use an add-on to a program like R, SAS or SPSS; they should first review the package themselves (including test data-sets) to make sure that the results are appropriate and accurate. As an open project, R is able to make it easier for users to find and use these resources. But yes, users do need to be aware of what they are doing when they use these resources.
Damn
With apologies to DD for copying him and double posting this.
But, this is a great article and deserves the attention.
Heh
Just be glad that I didn't comment my own post with "First!!11eleven!".
DD
Dirty D writes about polling, analytics, data and whatever else may cross his mind as being neat. Feel free to contact him by email : D I R T Y D AT O V E R D E T E R M I N E D DOT N E T.
More rows please
I track new posts on OD.net via my iGoogle homepage. I will have to increase the number of lines it shows me, so I don't miss things on days when there is a lot of new material posted to the site.