[R-SIG-Mac] learning R

Philippe Grosjean phgrosjean at sciviews.org
Mon Dec 13 09:56:16 CET 2010

Ah! That is interesting!

For me, the learning curve is like the energy required for a chemical or 
a biochemical reaction to occur. Thus, on the X-axis, you could have the 
amount of R learned/assimilated, and on the Y-axis, you have the 
energy/effort/time (or whatever measure of learning effort you like to 
use). Thus, a steep learning curve means you have to provide a lot of 
effort to reach a little bit of learning. A flatter learning curve 
allows you to progress faster with little effort (like running on a flat 
ground versus climbing a mountain, to reuse John's metaphor).

Still with the (bio)chemical reaction analogy in mind, anything that 
makes R more "digest" (a good tutorial, a good GUI, etc.) is like an 
enzyme that allows for the (bio)chemical reaction to occur with less energy.

Otherwise, I am always amazed that people could use a metaphor like 
here, a reference to the shape of a curve on a graph, without even 
knowing what exactly are the X- and Y-axes. Shame on us!


  ) ) ) ) )
( ( ( ( (    Prof. Philippe Grosjean
  ) ) ) ) )
( ( ( ( (    Numerical Ecology of Aquatic Systems
  ) ) ) ) )   Mons University, Belgium
( ( ( ( (

On 13/12/10 00:36, William Revelle wrote:
> I share with Carl about the misuse of the term learning curve.
> The original derivation was from learning theory where one plotted
> number of correct responses on the y axis against trial number on the x
> axis. Steep learning curves thus implied rapid learning (of easy
> material). Flat learning curves implied slow learning (of difficult
> material).
> When I complain about this misuse of the term to my psychological
> colleagues, they smile, agree with me that their usage was incorrect,
> and then suggest I go back to worrying about how to code things in R.
> When I teach R, I suggest that yes, the learning curve is steep, smile
> and then point out that therefore it is easy to learn. Realistically, as
> is true of learning any skill, the curve is negatively accelerated with
> asymptotic learning achieved only by the wizards of Core R.
> Bill
> At 8:42 AM +1100 12/13/10, John Maindonald wrote:
>> Surely what is envisaged is the sheer effort involved in climbing
>> a step mountain side. It does not have a graph in mind. If one
>> wants to change the metaphor and turn it into a graph, it is not
>> at all obvious what the horizontal axis ought to be, though
>> various rather strained interpretations can be proposed.
>> There is a further aspect to the metaphor that deserves attention.
>> The reward for negotiating the steep learning curve is to reach
>> a great height, where marvellous vistas spread out before the
>> climber!
>> John Maindonald email: john.maindonald at anu.edu.au
>> phone : +61 2 (6125)3473 fax : +61 2(6125)5549
>> Centre for Mathematics & Its Applications, Room 1194,
>> John Dedman Mathematical Sciences Building (Building 27)
>> Australian National University, Canberra ACT 0200.
>> http://www.maths.anu.edu.au/~johnm
>> On 11/12/2010, at 8:25 AM, Rolf Turner wrote:
>>> I agree with you completely about ``begging the question''. The
>>> nearly universal misuse of this expression drives me crazy. I'm
>>> not so sure about ``steep learning curve'' however. My impression
>>> is that this phrase has *always* been used to convey the idea that
>>> a subject area is difficult to learn, whence to use it (as you suggest)
>>> in the sense that the subject area can be learned quickly would be to
>>> change the original meaning of the phrase. That would be undesirable,
>>> even given that the original meaning is counter-intuitive.
>>> I recall having heard/read a ``justification'' for the original meaning
>>> to the effect that what is envisaged is plotting effort expended on
>>> the *y* axis and knowledge level on the *x* axis. Thus a steep learning
>>> curve would entail expending a great deal of effort for a small increase
>>> in knowledge.
>>> I agree that this is a silly choice of axes --- I certainly wouldn't
>>> make
>>> such a choice. But I don't suppose that there's any law against it.
>>> cheers,
>>> Rolf Turner
>>> On 11/12/2010, at 4:22 AM, Carl Witthoft wrote:
>>>> Next to "begging the question," the phrase "steep learning curve" is
>>>> probably the most misused cliche out there.
>>>> A 'learning curve' represents knowledge (or understanding) as a
>>>> function
>>>> of time. THerefore, the steeper the better.
>>>> Please help save the English language from descent into Humpty-Dumpty
>>>> land, and train your colleagues in the correct usage of both these
>>>> terms.
>>>> Carl
>>>>> Message: 2 Date: Thu, 9 Dec 2010 09:51:27 -0800 From: Payam
>>>>> Minoofar<payam.minoofar at meissner.com> To:
>>>>> "r-sig-mac at r-project.org"<r-sig-mac at r-project.org> Cc:
>>>>> "deniz.kellecioglu at gmail.com"<deniz.kellecioglu at gmail.com> Subject:
>>>>> [R-SIG-Mac] R for Mac, good enough?
>>>>> Message-ID:<53DF393B-2037-4B0D-890F-8DBAA1BA1F55 at meissner.com>
>>>>> Content-Type: text/plain; charset="us-ascii"
>>>>> The power of R is virtually unmatched, and R for Mac works extremely
>>>>> well.
>>>>> The learning curve is steep, however, and documentation is difficult
>> >>> to grasp, even though it is abundantly available. I am more partial
>>>>> to a commercial data analysis package with which I grew up, but I
>>>>> have done enough work with R on the mac platform to recommend it
>>>>> highly.
>>>> _______________________________________________
>>>> R-SIG-Mac mailing list
>>>> R-SIG-Mac at r-project.org
>>>> https://stat.ethz.ch/mailman/listinfo/r-sig-mac
>>> _______________________________________________
>>> R-SIG-Mac mailing list
>>> R-SIG-Mac at r-project.org
>>> https://stat.ethz.ch/mailman/listinfo/r-sig-mac
>> _______________________________________________
>> R-SIG-Mac mailing list
>> R-SIG-Mac at r-project.org
>> https://stat.ethz.ch/mailman/listinfo/r-sig-mac
> _______________________________________________
> R-SIG-Mac mailing list
> R-SIG-Mac at r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-mac

More information about the R-SIG-Mac mailing list