Re: Steep learning curve - or not?

On Mon, 28 Nov 2005 15:56:43 +0100, trio@xxxxxxxxxx (Donna
Richoux) said:

> Bob Cunningham <exw6sxq@xxxxxxxxxxxxx> wrote:

> > On Sun, 27 Nov 2005 16:22:01 GMT, Bob Cunningham
> > <exw6sxq@xxxxxxxxxxxxx> said:

> > [...]

> > > Here's my take. I don't think I ever saw anyone try to
> > > relate "steep learning curve" to an actual graph, but so far
> > > as I know, among the engineering and marketing people I
> > > worked with, we were all agreed on what it meant, and we
> > > used it fairly often: It meant you had a lot to learn
> > > about something and not much time to learn it.

> > > If I try to conceive of a graph that fits that definition, I
> > > get a starting point (beginning of effort and nothing
> > > learned yet) and an ending point (enough learned and time is
> > > up). The steepness would refer to the slope of a straight
> > > line drawn between those two points. It wouldn't matter
> > > much what wiggles the curve might have in between. All that
> > > mattered was knowing enough soon enough.

> > Looking at some Google hits on "steep learning curve", I can
> > understand why some people think other people think it means
> > "difficult to learn". In every hit I've looked at, "steep
> > learning curve" is used because there's a lot to learn. But
> > my concept of "steep learning curve" fits those cases
> > because in every case there must be an implicit time
> > available to do the learning. With a certain time
> > available, the more you have to learn in a given time, the
> > steeper the learning curve is.

> > I suppose it's possible, though, that some people have heard
> > "steep learning curve" used properly and have misperceived
> > its meaning to be simply "difficult to learn". Like "beg
> > the question" and "could care less", "steep learning curve"
> > would then acquire an accepted meaning that makes no sense
> > when the words are taken literally.

> Whether you take it to mean "difficult to learn" or "a lot to
> learn in a short time" doesn't matter to me. They're close
> enough to the same thing.

They're totally different. The word "steep" implies a
slope. A slope implies a first derivative of one variable
with respect to another. When you say "difficult to learn",
you have only one variable, so you have no slope, so
"steep"' has no meaning. When you say "amount learned vs
time consumed", you have two variables, so you have a slope,
and "steep" has meaning.

If I say it's 300 miles from point A to point B and it took
me six hours to travel that distance, I don't need to draw a
graph to say that my average speed was 50 miles per hour and
that if I want to make it in less time I have to drive
faster than 50. If I say it's 300 miles over narrow,
winding mountain roads covered with snow and ice, I can say
that's a difficult drive, but you can't calculate my average
speed until I tell you how long it took me to make the
drive.

Your implying that with regard to the meaning of "steep"
there's no difference between a case with only one variable
and one with two suggests that you haven't thought enough
about what "steep" means.

> Neither one relates directly to any graph that was ever
> used for any practical purpose, and therefore this
> non-existent graph cannot be the true source of the the
> expression.

So far as I know, we haven't been discussing the "true
source" of the expression "steep learning curve"; we've been
discussing different meanings it has for different people.
Maybe you've been discussing the true source, but I don't
care what the true source may be; I'm interested only in
what it means now.

A purist will tell you that it's wrong to use "aggravate" to
mean "annoy" and that it can mean only "make heavier". One
who recognizes that words can have different meanings will
tell you that both meanings of "aggravate" are in use and so
are acceptable.

"Steep learning curve" appears to represent a similar
situation. I know its meaning as my coworkers and I used it
for about fifty years. Others, like you, know of meanings
that I'm not familiar with, feel no need of, and have no
reason to learn. I'm content with the meaning I know and am
long accustomed to. I suggest that you be content with the
one you like and stop trying to tell other people they're
wrong because they know and use meanings that are different
from yours.

> I would actually like to discuss your posts item by item
> but I've got other stuff I'm supposed to do today, and my
> post wouldn't be much different from the ten or more posts
> I made some months ago, taking other people apart item by
> item.

Since I can see that you have a myopic attachment to a
single meaning of "steep learning curve", I have no doubt
that I could take your reasoning apart item by item. I
wouldn't bother to do so, though, because your underlying
premise, that there is only one acceptable meaning for an
English phrase, is all that's needed to thoroughly discredit
your position. No further taking apart is needed.

> These imaginary graphs that supposedly
> plot "amount assigned to learn" against "time it takes," or
> "time" against "amount learned," or "amount assigned to
> learn" against "cumulative amount learned," or "effort
> required" against "results," or whatever people want to
> dream up,

> (l) do not follow the basic requirements of graphic
> practice (like I say, I can't go into why not, again)

The only indispensable requirement for drawing a graph is
that you have two variables such that the value of one can
be found from the value of the other. In the simplest case
the change will be linear, so the graph is a straight line.
That's the case where all you're interested in is the time
available to learn something and the time available to learn
it.

> (2) do not yield the required result of "the steep part
> represents the difficult part,"

You're assuming that that is invariably the "required
result". Others are free to say what the required result is
in their application.

> or (3) in the rare case a tortured graph
> can be invented to fit, has never actually been employed
> by anyone (witness, no such graph can be found anywhere
> except in people's minds.)

If you can say that a straight line between two points is a
"tortured graph", you must have a different perception of
what "tortured" means from the one I have.

If you know of a graph someone has bothered to draw to
illustrate that driving 300 miles in six hours shows an
average speed of 50 miles per hour and that to make it in
less time you have to drive faster than 50, please tell us
where it is so we can look at it.

> On the other hand, there actually have been real graphs,
> displaying genuine data, that have been called "learning
> curves" for a century or so, and they're *different*. The
> learning is neither the x-variable nor the y-variable, it
> is what is demonstrated by the shape of the curve. When
> it is steep (steeply dropping), the subject is learning
> and producing more results (or faster or cheaper). When
> it levels off, they're not learning and produce the same
> results.

That is a meaning of "steep learning curve" that I'm not
familiar with, can foresee no personal need for, and have no
desire to remember. I'll stick with the one I fully
understand and have used effectively. But I won't deny that
you have the right to embrace whatever meaning pleases you.

You probably will continue to be like Miss Thistlebottom
heaping scorn on those who use "decimate" to mean "destroy
most of".

(Note the highly significant absence of a comma after
"Thistlebottom".)
.

Relevant Pages

  • Re: Candidate for worst English word: biweekly
    ... Since I never understood the steep curve, ... "steep learning curve" means that you have to learn more in less ... Try drawing the graph you just imagined and described to us, ... The x axis is time and the y ...
    (alt.usage.english)
  • Re: A true pre-emptive multi-tasking operating system
    ... If a "learning curve" is what you get on a graph of knowledge versus ... then why would a steep learning curve be a bad thing? ... It's the relationship between experience and efficiency, ...
    (comp.os.vms)
  • Re: Dense vs. Continuous
    ... I was trying to ascertain what the difference was between "dense" and "continuous". ... The graph would be a continuous line ... has no clear meaning in this context. ... But, suppose I defined the x-axis to be the set of rationals, not the set of reals. ...
    (sci.math)
  • cdt glossary [Graph]
    ... relation why do they define a graph of a function f as ... In cdt it is the latter meaning that is mostly used (when ... possibly confusing definitions of 'function'] ... Just helping to avoid boring repetitons of real cdt miscommunications ...
    (comp.databases.theory)
  • Re: Meaning of -:0 in unsigned infinity solve working with limits?
    ... The graph is undefined for x<=0. ... The calculator ... the meaning in previous posts. ...
    (comp.sys.hp48)