Getting started

Hi – I’m going to play about with the following software:

• modellus – http://modellus.fct.unl.pt/
• geogebra – http://www.geogebra.org/cms/
• Algodoo – http://www.algodoo.com/wiki/Algodoo_Play
• VnR – https://www.box.net/shared/static/akg8e4roko.zip

This is prompted by a physics specialist I support asking the best way to teach a topic that includes an equation, such as speed. So I’m looking for something that would be accessible for 14 year-olds (year 10) of all abilities, and would support their understanding of both the physics and the maths. This partly came from a discussion a while ago about equation triangles: https://www.talkphysics.org/mod/groups/topicposts.php?topic=35694&group_guid=13617 .

Anyone want to join me? Any other software to suggest trying out?

Depends on whether you want

 

a) intuition developing – so simulations-> Algodoo, or want to model – that is being able to write the rules of your imagined world…

b) assume or hope to develop algebraic competence – >Modellus

c) pre-algebra, but semi-quantiative -> VnR

 

For this target audience you ought o also consider StarLogo, or Scratch. More to follow on both.

 

Geogebra is focussed on constraint relationships, rather than time-evolution, so I guess best not used here…

This seems a good place to start a conversation about the upcoming Rugby workshop…so I’ll do that.

Here are a few models, in case we get stuck.

modelsforrugby.zip

Rubgy Workshop 2013 – descriptor

# Computational modelling workshop

Learning physics is quite hard, yet physics is rather simple.

We ask children to:

• Imagine challengingly simple worlds into existence.
• Express the rules that govern these worlds explicitly.
• Notice how the worlds evolve.
• Compare the lived-in world with these imagined worlds.

Creating explicit computational models with children provides resources to support each of these phases, so lightening the load.

Wider benefits include creating intellectually valuable entities and insights into the essential limitations of modelled forecasts.

In the workshop you will:

• Select building blocks for particular tasks
• Construct models from these blocks
• Evaluate the models and the opportunities

I’ll be open to continuing the conversation on TalkPhysics.

So here I am…conversing with myslef, as ever

Programming And Modelling

BCS national curriculum demands…

Design and write programs that include

o Sequencing: doing one step after another.

o Selection (if-then-else): doing either one thing or another.

o Repetition (Iterative loops or recursion).

o Language constructs that support abstraction: wrapping up a computation in a named abstraction, so that it can be re-used. (The most common form of abstraction is the notion of a “procedure” or “function” with parameters.)

o Some form of interaction with the program’s environment, such as input/output, or event-based programming.

Find and correct errors in their code.

Pretty much all of these are naturally a part of modelling.

However, this is not quite the kind of thing?

Newtoncooling Lua

Lua

T0 = 100

TR = 20

k = 0.07

delta_t = 2, 5, 10

n = 100

NewtonCooling = function( t ) return -k * ( t – TR ) end

function Euler( f, y0, n, h )

local y = y0

for x = 0, n, h do

print( “”, x, y )

y = y + h * f( y )

end

end

for i = 1, #delta_t do

print( "delta_t = ", delta_t[i] )

Euler( NewtonCooling, T0, n, delta_t[i] )

end

Javascript would seem to be commonly available, and a runner:

You could have the core of a model like this, which looks more or less intelligible:

 

var force;

var mass;

var acceleration;

var position;

var velocity;

var timeonClock;

var timeInterval;

 

function accumulate(AccumulatorValue,QuantityValue)

{

return QuantityValue = QuantityValue + AccumulatorValue*timeInterval;

}

 

 

// set initial values

 

force = 3;

mass = 5;

timeInterval = 0.01;

position = 0;

velocity = 0;

timeonClock = 0;

WallLocation = 7;

 

 

// find acceleration

 

acceleration = force/mass;

 

// set condition

while (position<=WallLocation){

 

// do the accumulations

timeonClock = accumulate(timeInterval, timeonClock)

velocity = accumulate(acceleration, velocity)

position = accumulate(velocity, position)

 

// store these results

 

PlotThis+= AddPoints(timeonClock,position);

 

}

 

 

 

The difficulty is that none of the charting functions commonly available seem to be easily tweakable[1] to operate with computed data – they rather assume that the data exists in the data table at the time of the interpretation of the document – otherwise we’d have a modelling environment in every javascript-enabled browser…

 

[1] That is, to seems likely that you’d have to write the data out to a file and then read it in again….

The code generates a datatable, which can then be pasted into a web page, resulting in:

 

<iframe name=”SPTframe” src=”https://dl.dropboxusercontent.com/u/11977706/ModellingTrials/chartfromlocaldatatable6.html&#8221; frameborder=”0″ scrolling=”yes” width=”680″ height=”506″ marginwidth=”0″ marginheight=”0″ ></iframe>

Or you can put things into the nearly-great jsfiddle:

Which works fine for data-tables, but not for computed data – something to do with the DOM….

And of course, you can simulate on the Canvas, it being HTNML5:

 

So I suppose we could hack out a graph from primitives on the Canvas…but that rather depends on people being interested in doing modelling…using Javascript. Is it intelligible enough? Helpful to make the link with coding?

As here is the BCS suggestion for coding:

 

BCS national curriculum demands…

 

Design and write programs that include 

o Sequencing: doing one step after another. 

o Selection (if-then-else): doing either one thing or another.

o Repetition (Iterative loops or recursion). 

o Language constructs that support abstraction: wrapping up a computation in a named abstraction, so that it can be re-used. (The most common form of abstraction is the notion of a “procedure” or “function” with parameters.) 

o Some form of interaction with the program’s environment, such as input/output, or event-based programming. 

 

Find and correct errors in their code.

 

Pretty much all of these are naturally a part of modelling.

maybe what’s needed is a way of keeping what we’re doing in focus… however its implemented:

overview.png

Effectively a guide for action. And this could be assembled from sub-bits, all of which are tested (Physics as bricolage: aka physics is simple – we keep on using the same bits over and over).

 

In fact the mapping from this to the Javascript model, embedded above, looks rather direct…

 

So sequencing, abstraction, repetition, decision-making..not so hot on the event-based input for this model.

Hi Ian – sorry, I’d opted to take a step back from this when I realised how much I’d taken on with the evidence-based teaching stuff for the IOP National Teachers CPD Conference 2013. 

 

I tried to take a look at your attached models above (modelsforrugby.zip), but I’m blocked from accessing zip files whilst in the IOP offices. Which software are they for, or are they for a range of the software you mention above? 

Well, rather than starting a new group(!) I’ve started a conversation with myself ahead of the Rugby meeting in June, where I’ve got asked to run a workshop on modelling. The plan is to continue the conversation here after that event.

 

A pair of models is for VnR and a pair for Modellus, but there’s nothing very clever there – they’re just  as a bit of scaffolding, in case….

[Author]

Posts