Hi folks,
Quoting Jan Kim <[email protected]> on Mon, 9 May 2016 11:11:12 +0100:
Dear All,
On Fri, May 06, 2016 at 10:20:41AM -0400, Ivan Gonzalez wrote:
Hi,
I do a similar thing. I usually start by running a simple statement
(print("Hello") or something alike) from the Python interpreter, then from
IPython, and then from the Jupyter notebook. (This was more important
before the renaming of IPython notebook to Jupyter).
I've used the same approach in the past (not in official SWC workshops
to be pedantic, but in teching computing / Linux to scientists in general).
I'd type some Python line-by-line into the plain vanilla interpreter,
pointing out that "now we have a different prompt, the Python prompt,
so we have to type Python code to be understood". Then (well before starting
with program flow control, because getting indentation right when typing
interactively is not a skill I really possess) I'd write the line(s)
known by then into a script, and go through the cycle of edit -- save
-- run -- check results a number of times quite explicitly and without
introducing new stuff that could be distracting at this point, to get
everyone familiar and comfortable with that cycle.
From my own experience, that cycle is one of those things that may appear
trivial to the initiated (most instructors), but can make a real difference
in terms of enabling learners to go on learning after the workshop.
One observation I have about using the numpy based intro is that this
can be confusing to participants. I got the question "what do these plots
mean?" after we had been through a couple of them.
Generally, not all particpants are not familiar with the concept of plotting
a matrix of numerical values using some arbitrary colour scheme. Without
some practice / familiarity / mental model of that, they may end up with
dual cognitive demands of having to match the colour-coded plot to the
time-courses (with the peak in the middle), and trying to figure out the
indexing and slicing (some of which is rather fancy and not standard
Python).
For these folks, introducing Python via a bunch of simple statements may
be more beneficial.
It would also be more in the spirit as to what is the stated goal of
the lesson, which isn't to teach numpy or Python:
"Our real goal isn?t to teach you Python, but to teach you the basic
concepts that all programming depends on." [1][2]
If I was a learner who has to write C or Fortran, I'd have difficulty
seeing the point, whereas I'd see the point of good variable naming,
pulling duplicated code into functions, meaningful comments etc.
I also agree with your moving to the edit-save-run approach as soon as
possible, again because it's more consistent with how programming is
done in C, C++, Fortran and Java, but also because typing loops or
conditionals at the Python prompt quickly becomes painfully error prone.
cheers,
mike
[1] http://swcarpentry.github.io/python-novice-inflammation/
[2] No more than the revision control lesson is primarily to teach Git
and GitHub, for example.
Best regards, Jan
Every time I open and close the corresponding tool, noting the differences
when pressing "Enter". This allows me to stress the difference between
Python, the language, and python, the program, and the rest of the tools.
This is useful if later in the class if I have to open/restart the
notebook. It good to note here what's Anaconda too.
Then I introduce arithmetic operators with a simple statement (say 1+1) and
then variable/names to save the result of the operation (say two = 1 + 1)
and briefly talk about the memory, objects and names.
This is very quick ~10 mins, but allows you to set some vocabulary (object,
name, statement, operator, ...) and some idea of what's happening
memory-wise which you may need later.
Then I follow the lesson importing numpy. I take advantage of the previous
definitions and usually say here that numpy is just another name which
refers to a bunch of code that is now loaded in memory and that was
installed as part of Anaconda.
Best,
Ivan
On Thu, May 5, 2016 at 5:50 PM, Byron Smith <[email protected]> wrote:
After my first time teaching python-novice, I decided to always skip the
first section. Instead I work it in later while showing the use of
packages.
Although I understand why the Python lesson is structured the way it is,
and I like the idea of an application-first approach, this particular
implementation has been difficult for learners in my experience. I look
forward to seeing the new lesson as it takes shape.
Instead I have had great success with the more traditional, "bottom-up"
approach. Luckily, that's what you end up with when you remove section one
from the current lesson.
While my experience contradicts the general SWC belief that material needs
to be motivated early, I've found learners to already believe that they
should learn Python, making a focus on motivation less important. This is
probably not the case for the shell and git lessons.
-Byron
On Thu, May 5, 2016 at 4:35 PM, Maxime Boissonneault <
[email protected]> wrote:
Hi Alexandre,
I find that starting with Numpy is great, because it allows to do
interesting things without the need for introducing weird and boring
concepts like loops, variables or conditionals.
Students immediately can relate to things they know like seeing a
picture, or seeing a table of numbers.
Starting with abstract concepts like variable naming, loops,
conditionals, etc, is bound to bore students, which is never a good thing
at the very beginning of a class.
Cheers, and have a good first time!
--
---------------------------------
Maxime Boissonneault
Analyste de calcul - Calcul Qu??bec, Universit?? Laval
Pr??sident - Comit?? de coordination du soutien ?? la recherche
de Calcul Qu??bec
Team lead - Research Support National Team, Compute Canada
Instructeur Software Carpentry
Ph. D. en physique
On 2016-05-05 15:16, Alexandre Manhaes Savio wrote:
Hi,
I am preparing for the first time the python-novice-inflammation lesson.
I am also going to add in the beginning instructions on the most
popular initial options on how to actually run the python code.
I am struck on how the novice lesson directly starts with Numpy
without explaining too much variable naming and assignment, for
instance. In my case I will begin with students with very little
programming experience.
I am also aware that Greg is preparing another lesson for python
but there is some work there to be usable.
Could you please write me your opinion and experience about how
you prepare the beginning of the python novice lesson?
Thank you very much!
Best,
Alex
Alexandre Manh??es Savio
PhD, Medical Imaging, Machine Learning
Klinikum rechts der Isar, TUM, M??nchenalexsavio.github.io |
email: [email protected]
_______________________________________________
Discuss mailing
[email protected]http://lists.software-carpentry.org/mailman/listinfo/discuss_lists.software-carpentry.org
_______________________________________________
Discuss mailing list
[email protected]
http://lists.software-carpentry.org/mailman/listinfo/discuss_lists.software-carpentry.org
_______________________________________________
Discuss mailing list
[email protected]
http://lists.software-carpentry.org/mailman/listinfo/discuss_lists.software-carpentry.org
_______________________________________________
Discuss mailing list
[email protected]
http://lists.software-carpentry.org/mailman/listinfo/discuss_lists.software-carpentry.org
--
+- Jan T. Kim -------------------------------------------------------+
| email: [email protected] |
| WWW: http://www.jtkim.dreamhosters.com/ |
*-----=< hierarchical systems are for files, not for humans >=-----*
_______________________________________________
Discuss mailing list
[email protected]
http://lists.software-carpentry.org/mailman/listinfo/discuss_lists.software-carpentry.org
------------------------------------------------------------
Dr. Michael (Mike) Jackson [email protected]
Software Architect Tel: +44 (0)131 650 5141
EPCC, The University of Edinburgh http://www.epcc.ed.ac.uk
Software Sustainability Institute http://www.software.ac.uk
--
The University of Edinburgh is a charitable body, registered in
Scotland, with registration number SC005336.
_______________________________________________
Discuss mailing list
[email protected]
http://lists.software-carpentry.org/mailman/listinfo/discuss_lists.software-carpentry.org
