Saturday, April 26, 2014

Why are R users so damn Stingy?!

YES R!

Looking at rapporter's recent blog post on "R activity Around the World", I am shocked by how few users actually monetarily support the R Foundation.  Looking at the US for instance there are only 27 donars representing a little more than 0.1% of those registered on R users groups (of which there is nearly 27,000 members) which is a small estimate of the user base which is hard to estimate.

To get a better idea of the user base according to the same report it states that over 8 million packages have been downloads from the US.  Imaging that each active user may download on average 50 packages this gives a US user base of no less than 160,000 active users as an extreme lower bound.  This notation of large usage is complemented by Robert A. Muenchen's long running article tracking usage statistics of statistical software.

Yet we know that R users represent some of the highest earning professional skills we know that if users were not using R they would likely end up paying for proprietary software which costs thousands of dollars in order to do the same tasks.  And still the number of actual donors is abysmal.

So Why Do R Users not Contribute to the Foundation?

1. Nobody knows what the foundation does!  

The R foundation lists three things under its purpose:
  • Provide support for the R project and other innovations in statistical computing. We believe that R has become a mature and valuable tool and we would like to ensure its continued development and the development of future innovations in software for statistical and computational research.
  • Provide a reference point for individuals, instititutions or commercial enterprises that want to support or interact with the R development community.
  • Hold and administer the copyright of R software and documentation.
The first two points (1. provide general support for R and 2. provide public face for R) are critical points for which R has continued to be handicapped in gaining popularity I believe largely because the foundation has been underfunded and inactive.  I believe that being a board member on the R-Foundation should be seen with similar prestige as other massive open source projects such as Firefox, Apache, and GNOME.

Yet when I called the foundation to talk with someone, I cannot even be sure I got the write number (despite it being listed on the foundation's webpage).  It is clear to me that the foundation has taken an extreme back position in promoting the public image of R, leaving it almost entirely up to the user base, other foundations, and corporations to promote the language.

This has worked fine, yet the roles outlined above are important roles which should not be left by the wayside.

2. It is pain in the %$# to contribute to the foundation!  

In the process of writing this post I  attempted to call the phone number on the R-Project Foundation website.  The person I got was not happy to talk to me and well did not seem interested in talking to me at all.  When I looked at contributing to the foundation I found a PDF form that was supposed to be printed off and mailed with check or credit card information to the foundation!  What decade are we in?

I am surprised even 27 people in the US gave to the foundation.  Besides the R-Project website interface clearly have not undergone any major renovations in the last ten years.  Who uses frames anymore?  Why would anybody fill out a pdf document to mail in when the standard for professional websites is to provide a secure interface for making payments online?

3. Projects funds usually cover software budgets

And since R is free nobody factors in the software cost to their budget.  Professionals never want to pay for something out of their own pocket which could be paid for out of their project budget.  However, R is so clearly free that it is impossible for a project to allocate a donation to the continued support of R even though the program administrators might be very willing to provide such funds.

I therefore propose an optional annual "Maintenance Fee" that will provide businesses and institutions with a (expense account) justification for funding R.  Such a service could come with priority support on R mailing lists or forums with maybe three tears (Gold- $1000, Silver- $500, and Bronze $100, maybe).  Users could post their status when asking questions and other users who respect the donors willingness to pay to support the R-Project will be more generous with their time when answering such questions.  Such a system would allow for projects grants and funds to channel some small portion of their resources to help support the continued existence of R.

4. Users do not like paying to a single service

This is something that find I particularly difficult.  My logic goes, why give to R when there are so many people here in need in Mozambique?  But how do I decide which organization to give to say Free The Girls an organization which provide an alternative source of income for prostitutes or Massana an organization which provides food and education to street kids in Mozambique (I personally know board members of both these organizations and they are excellent people who serve faithfully).  Then I must wonder how much to give and in what increments etc.  Long and short of it, I give much less than I intend to and when I do give it is usually for a friend raising money for this thing or that thing.

I am therefore suggesting that if you are like me then please consider giving money through flattr.  It is an organization which acts as a clearinghouse of donations.  You give a fixed amount to flattr each month and flattr redistributes 90% of those funds to the organizations that you have chosen.  The other 10% it keeps for itself.  It seems to me that this is an excellent way for users of R to fund R as well as other initiatives which seem worthy.

Since I was unable to contact the R-Foundation I have set up a flattr account in there name which people can donate to using the following button:

RFoundation Flattr this

As soon as I am contacted by a verified Foundation Member, I will transfer over complete control of the flattr account.  (Yes it is strange that there is no verification step to ensure that creators are actually the ones who set up the flattr accounts)

But Why Give to the R-Foundation?

I have frequently wondered why it is that despite a super abundance of resources R maintains its reputation as a language which has a steep learning curve.  I personally attribute this reputation primarily to the horrible user interface that R new users routinely encounter when going to R-Project.org.  It is frankly embarrassing to be an R user when the platform is so bad at representing itself.

Likewise the foundation clearly needs to have some resources to fund staff members.  This staff could focus on developing resources in order to provide basic support to media, business, universities, students, etc.  The R-Foundation compares its existence to the Apache Foundation and the GNOME Foundation yet despite the tremendous success of R, it has no official public image to speak of if I can gauge from the webpage or the failed 5 minute phone conversation I had with the official number.  I believe, all users of R will benefit by the language representing itself more professionally.

With additional funding the foundation could also provide additional support to making R user conferences appear more professional as well support the development of the R-Journal and other R other publications.

However, the primary goal of the foundation which could be furthered through the support of a wider donor base is the continued development of resources to facilitate the use of R by existing users as well as continuing to develop new tools for new R users.

Thank for reading! A good rule might be to think about how much you would be willing to pay to use R if it were proprietary then give say 5% of that.

If you are a frequent reader of my blog please consider flattring me! In the last year I have made 3 dollars and 62 cents from people flattring my blog :)

Econometrics by Simulation Flattr this

14 comments :

  1. Yes, I know. I am also not a supporting member of the R-Foundation up to this point.

    ReplyDelete
  2. It is indeed a shame that donating is not just a few clicks. There have been some r-help threads on this so there is demand for an easier way to donate:
    http://r.789695.n4.nabble.com/Make-a-donation-via-PayPal-was-R-Monetary-support-to-the-R-project-td1747936.html
    http://r.789695.n4.nabble.com/Re-Monetary-support-to-the-R-project-Was-Re-Executable-for-Production-Use-td1585186.html#none

    ReplyDelete
  3. Good post. This needs to be addressed by the foundation, and I would be surprised if their isn't some action around improving the ease of use around a donation system/process. Moreover, I think it is important to raise awareness, and it that regard I say again: Good post!! We should all consider donating to the foundation. You have inspired me to do so.

    ReplyDelete
  4. Yeah, I really wanted to donate like 2 years ago and was discouraged by that PDF form. However, I want to caution you re the damages which confusion can bring to most users.

    Also, R users being the highest paid professionals doesn't really mean much. If you want donations, you don't need the average to be really high. You just need the rich to be really rich. Not sure where R stands in terms of how many users brought their companies to NASDAQ.

    ReplyDelete
    Replies
    1. Could you please clarify what you mean by "caution"? I just don't understand what you are saying.

      Delete
  5. I am a little sad that you seem so annoyed by the fact that the R Foundation is so laid back on the matter of PR and competing for donations. For me, it's a matter of credibility. The moment R starts having "modern" web pages, a built-in advanced GUI and a "donate to us visa paypal!1!11!" button, people will migrate from R to something else. Julia, perhaps. I certainly would.

    I am a scientist, and I happen to think R is a tool from scientists for scientists. Even though (especially advanced) R users are admittedly stingy - and yes, I'm afraid do get ripleyed when I write a mail to r-help: these people are (mostly) academics with a tenured position, developing R as a part of their work(, mostly). So "shut up and take my money" is not really the attitude I would follow here, personally.

    Also, at least where I work, scientists would also have serious trouble taking extra money for their work - at least from an administrative point of view. I'm quite positive it would not be worth the effort.

    Also, while R is quite popular among researchers, I have serious and well-founded doubt it's a real asset in your CV outside of academia. Sorry to say, but when I ask my friends outside academia, no one really works with R. Maybe, it's the causality question all over again? Analysts earning the top buck are *also* able to use R, because they've got all the superskills needed for top grade payment?

    Final sidenote: May I draw your attention to the flattr FAQ (https://flattr.com/support/faq) under the point unclaimed flattr? You would not even have needed to set up an account, from what I gather. However, thanks for the initiative - it'll probably make me join flattr, finally. :)

    ReplyDelete
    Replies
    1. Thanks for your thoughtful response.

      1. You raise some interesting points. I guess I thought the employment scene with R was much more promising than your comment has suggested. Did you have change to look at some of the posts about R usage cited above?

      2. As for people migrating from R because it becomes too flashy to something like Julia. Actually, I have become of the opinion that hard core gritty coder types should migrate to Julia. Julia is much faster and cleaner to work with. I believe R should become as user friendly as possible (maybe something similar to Stata) while still retaining core functionality. I know it is on the way there with 3rd party GUI's like RStudio. I do not mean to suggest that R should develop its own but rather simply improve the General User Interface on r-project.org.

      Delete
    2. Thank you very much for posting this and sending it to R-Bloggers. I printed the PDF form and will mail them a check, because I deeply appreciate the huge amount of effort that let's me enjoy R.

      But, there ought to be a much, much easier way to donate.

      Delete
    3. R has conquered academia and research environments, and it is popular in startups but it hasn't yet made much of an in-road into traditional enterprises - it lags behind SAS in finance and pharmaceutical companies, and behind MATLAB in engineering companies.

      There certainly are companies trying to push enterprise adoption of R, most notably Revolution Analytics, Tibco and Oracle. So I think R will eventually make progress in this direction regardless of the state of the r-project.org site and the ease of donating.

      Still, it seems like a relatively small simple task to clean up the site (according to the Internet Archive, the site hasn't changed much since May 2004, http://web.archive.org/web/20040415000000*/http://www.r-project.org/) and explain what the heck the R Foundation actually does. (Lot's of vague documentation that could just say: we host CRAN, we run the useR conference and we publish the R Journal.)

      Delete
  6. Interesting post, Francis. I was with you throughout, but then at the end I thought "if it ain't broke, why fix it?" I don't know what stress R core experience, but they seem to be content running things as is. If not, it would take nothing more than a blog post of needs from one of them to bring in all kinds of funding.

    I've introduced a few people to R and basically I just go through Rstudio. At first I felt it was important to explain the R behind Rstudio, but a colleague pointed out that it just confused learners. So now, if for nothing else, but to show how far we've come I still have them open R once, just to see what's behind the curtain, but other than that I actually see no need for a new user to even look at CRAN. Crazy right?

    ReplyDelete
    Replies
    1. Hey Tom,

      Well, this is an excellent point. I guess I don't know what I have to say about that except that this whole situation seems to be giving a lot of authority to a third party company/system. Not that I have any problem with RStudio or the team or the company. I think they are doing amazing things with R or for R.

      I just wonder if the public image of R should not be represented in and of itself somewhere. In addition, I think there are some core features which are widely demanded for R which will probably never be adopted because there is not an executive team directing the development.

      But I am not really sure if this is a major issue or not. Frankly, I don't really know. I think if there was a professional team employed with the mission of increasing usability and span of R through the R-Foundation then development might look different. But I don't really know. It would be up to that team to make the best choices for R and the greater community. But I really don't know what that might look like.

      I could imagine the R website looking a lot more like www.python.org or julialang.org. But the point is I don't really know what it could look like or what the effect of such a change would be. I just think that a wider support base, might really change the picture (in a positive way) of how the R-Foundation works.

      Thanks,
      Francis

      Delete
  7. Francis, thanks for raising these issues. If the goal of the foundation was to raise funds, then I think it could aggressively pursue them using several tools, few of which you mention. However, I think there might be a more altruistic motive going on here. What I loved about my introduction to R a little over a year ago (and I've been writing code in it on an almost daily basis ever since) was the absence of fear among newbies in posting questions (as there should be) and the abundance of free resources. That might go away with the tiers in payments and the priority responses to those with more resources. Taking the priority servicing idea further, they might choose to hide the currently free R-resources behind a pay-wall, which will be a great disservice to those who might want to join the bandwagon.

    ReplyDelete
    Replies
    1. Vivek, good points! Though I don't think that there is any way a paywall could ever be considered, adopted, or effected since there are already an abundance of resources available and the community which is entirely voluntary would probably outright veto the idea to death.

      Delete