GUI Paradigms (1): Redux (Flutter/React) translated to MATLAB

For GUI development, we often start with controls (or widgets) that user interact with and it’ll emit/run the callback function we registered for the event when the event happens.

Most of the time we just want to read the state/properties of certain controls, process them, and update other controls to show the result. Model-View-Controller (MVC) puts strict boundaries between interaction, data processing and display.

The most common schematic for MVC is a circle showing the cycle of Controller->Model->View, but in practice, it’s the controller that’s the brains. The view can simultaneously accept user interactions, such as a editable text box or a list. The model usually don’t directly update the view directly on its own like the idealized diagram.

MVC with User Action
From https://www.educative.io/blog/mvc-tutorial

With MVC, basically we are concentrating the control’s callbacks to the controller object instead of just letting each control’s callback interact with the data store (model) and view in an unstructured way.


When learning Flutter, I was exposed to the Redux pattern (which came from React). Because the tutorials was designed around the language features of Dart, the documentation kind of obscured the essence of the idea (why do we want to do this) as it dwelt on the framework (structure can be refactored into a package). The docs talked a lot about boundaries but wasn’t clearly why they have to be meticulously observed, which I’ll get to later.

The core inspiration in Redux/BLoC is taking advantage of the concept of ‘listening to a data object for changes’ (instead of UI controls/widget events)!

Instead of having the UI control’s callback directly change other UI control’s state (e.g. for display), we design a state vector/dictionary/struct/class that holds contents (state variables) that we care. It doesn’t have to map 1-1 to input events or 1-1 to output display controls.

When an user interaction (input) event emitted a callback, the control’s callback do whatever it needs to produce the value(s) for the relevant state variable(s) and change the state vector. The changed state vector will trigger the listener that scans for what needs to be updated to reflect the new state and change the states of the appropriate view UI controls.

This way the input UI controls’ callbacks do not have to micromanage what output UI controls to update, so it can focus on the business logic that generates the content pool that will be picked up by the view UI controls to display the results. In Redux, you are free to design your state variables to match more closely to the input data from UI controls or output/view controls’ state. I personally prefer a state vector design that is closer to the output view than input controls.


The intuition above is not the complete/exact Redux, especially with Dart/Flutter/React. We also have to to keep the state in ONE place and make the order of state changes (thus behavior) predictable!

  • Actions and reducers are separate. Every input control fires a event (action signal) and we’ll wait until the reducers (registered to the actions) to pick it up during dispatch() instead of jumping on it. This way there’s only ONE place that can change states. Leave all the side effects in the control callback where you generate the action. No side effects (like changing other controls) allowed in reducers!
  • Reducers do not update the state in place (it’s read only). Always generate a new state vector to replace the old one (for performance, we’ll replace the state vector if we verified the contents actually changed). This will make timing predictable (you are stepping through state changes one by one)

In Javascript, there isn’t really a listener actively listening state variable changes. Dispatch (which will be called every time the user interacts using control) just runs through all the listeners registered at the very end after it has dispatched all the reducers. In MATLAB, you can optionally set the state vector to be Observable and attach the change listener callback instead of explicitly calling it within dispatch.

https://gist.github.com/gaearon/ffd88b0e4f00b22c3159

Here is an example of a MATLAB class that captures the spirit of Redux. I added a 2 second delay to emulate long operations and used enableDisableFig() to avoid dealing with queuing user interactions while it’s going through a long operation.

classdef ReduxStoreDemo < handle

    % Should be made private later
    properties (SetAccess = private, SetObservable)
        state % {count}
    end
        
    methods (Static)
        % Made static so reducer cannot call dispatch and indirectly do
        % side effect or create loops
        function state = reducer(state, action)        
            % Can use str2fun(action) here or use a function map
            switch action
                case 'increment'
                    fprintf('Wait 2 secs before incrementing\n');
                    pause(2)
                    state.count = state.count + 1;
                    fprintf('Incremented\n');
            end                
        end        
    end
    
    % We keep all the side-effect generating operations (such as
    % temporarily changing states in the GUI) in dispatch() so
    % there's only ONE PLACE where state can change
    methods
        function dispatch(obj, action, src, evt)
            % Disable all figures during an interaction
            figures = findobj(groot, 'type', 'figure');
            old_fig_states = arrayfun(@(f) enableDisableFig(f, 'off'), figures);      
            src.String = 'Wait ...';
            
                new_state = ReduxStoreDemo.reducer(obj.state, action);
                
                % Don't waste cycles updating nops 
                if( ~isequal(new_state, obj.state) )
                    % MATLAB already have listeners attached.
                    % So no need to scan listeners like React Redux            
                    obj.state = new_state;
                end                
                
            % Re-enable figure obj.controls after it's done
            arrayfun(@(f, os) enableDisableFig(f, os), figures, old_fig_states);                        
            src.String = 'Increment';
        end
    end
    
    methods
        function obj = ReduxStoreDemo()
            figure();            
            obj.state.count = 0;                        
            
            h_1x  = uicontrol('style', 'text', 'String', '1x Box', ...
                              'Units', 'Normalized', ...
                              'Position', [0.1 0.3, 0.2, 0.1], ...
                              'HorizontalAlignment', 'left');                          
            addlistener(obj, 'state', ...
                        'PostSet', @(varargin) obj.update_count_1x( h_1x , varargin{:})); 
                        
            uicontrol('style', 'pushbutton', 'String', 'Increment', ...
                      'Units', 'Normalized', ...
                      'Position', [0.1 0.1, 0.15, 0.1], ...
                      'Callback', @(varargin) obj.dispatch('increment', varargin{:}));                             
            
            % Force trigger the listeners to reflect the initial state
            obj.state = obj.state;
        end
    end
    
    %% These are 'renders' registered when the uiobj.controls are created
    % Should stick to reading off the state. Do not call dispatch here
    % (just leave it for the next action to pick up the consequentials)
    methods
        % The (src, event) is useless for listeners because it's not the 
        % uicontrol handle but the state property's metainfo (access modifiers, etc)
        function update_count_1x(obj, hObj, varargin)            
            hObj.String = num2str(obj.state.count);
        end        
    end
    
end

Loading

Dart Language: late binding in lambda (no capture syntax)

Dart’s documentation at the time of writing (2021-10-27) is not as detailed as I hoped for so I don’t know if the Lambda (anonymous function) is late-binding or early-binding.

For those who don’t know what late/early-binding is:

Bindingearlylate
Expressionclosed (closure)open (not a closure)
Self-containedyesno
Free symbols
(relevant variables
in outer workspace)
bound/captured at creation
(no free symbols)
left untouched until evaluated/executed
(has free symbols)
Snapshotduring creationduring execution

Late-binding is almost a always a gotcha as it’s not natural. Makes great Python interview problems since the combination of ‘everything is a reference‘ and ‘reference to unnamed objects‘ spells trouble with lazy evaluation.

I’d say unlike MATLAB, which has a more mature thinking that is not willing to trade a slow-but-right program for a fast-but-wrong program, Python tries to squeeze all the new cool conceptual gadgets without worrying about how to avoid certain toxic combinations.

I wrote a small program to find out Dart’s lambda is late-bound just like Python:

void main()
{
  int y=3;
    var f = (int x) => y+x;	
    // can also write it as 
    // f(int x) => y+x;
  y=1000;	

  print( f(10) );
}
The result showed 1010, which means Dart use open lambdas (late-binding)

LanguageBinding rules for lambda (anonymous functions)
DartLate binding
Need partial application (discussed below) to enforce early binding.
PythonLate binding.
Can capture (bind early) by endowing running (free) variables (symbols) with defaults based on workspace variables
C++11Early binding (no access to caller workspace variables not captured within the lambda)
MATLABEarly binding (universal snapshot, like [=] capture with C++11 lambdas)

I tried to see if there’s a ‘capture’ syntax in Dart like in C+11 but I couldn’t find any.

I also tried to see if I can endow a free symbol with a default (using an outer workspace variable) in Dart, but the answer is no because unlike Python, default value expressions are required to be constexpr (literals or variables that cannot change during runtime) in Dart.

So far the only way to do early binding is with Partial Application (currying is different: it’s doing partial application of multiple variables by breaking it into a cascade of function compositions when you partially substitute one free variable/symbol at a time):

  1. Add an extra outer function layer (I call it the capture/binder wrapper) putting all free variables you want to bind as input arguments (which shadows the variables names outside the lambda expression if you chose not to rename the parameter variables to avoid conceptual confusion),
  2. then immediately EVALUATE the wrapper with the with the outer workspace variables (free variables) to capture them into the functor (closed lambda, or simply closure) which binder wrapper spits out.

Note I used a different style of lambda syntax in the partial application code example below

void main()
{
  var y=3;
    f(int x) => y+x;
	
    // Partial Application
    g(int w) => (int x) => w+x;	
    // Meat: EVALUATING g() AT y saves the content of y in h()
    var h = g(y);
	
  y=1000;	

  // y is free in f 
  print( f(10) );	

  // y is not involved in g until now (y=1000)	
  print( g(y)(10) );		

  // y is bounded for h when it was still y=3
  print( h(10) );
}

Only h() captures y when it’s still 3. The snapshot won’t happen if you cascade it with other lambda (since the y remains free as they show up in the lambda expression).

You MUST evaluate it at y at the instance you want y captured. In other words, you can defer capturing y until later points in your code, though most likely you’d want to do it right after the lambda expression was declared.


As a side note, I could have used ‘y’ instead of ‘w’ as parameter in the partial application statement

g(int y) => (int x) => y + x

but the ‘y‘ inside g() has shadowed (no longer refers to) the ‘y‘ in the outer workspace!

What makes it confusing here is that quite a few authors think it’s helpful (mnemonics-wise) to use the free variable (outer scope) name you are going to inject into the dummy (argument) as the name of the dummy! This gives the false impression of the variable being free while it’s bounded (through feeding it as a parameter in the wrapper/outer function)!

Scoping rules is a nasty source of confusion in understanding lambda calculus so I decide to give it a different name ‘w‘. I’m generally dismissive of shadowing under any circumstances, to the extent that I find Python’s @ syntactic sugar for decorators shadowing the underlying function which could have been reused somewhere (because it’s implicitly doing f=g(f) instead doing h=g(f). I hate it when you see the function with name f then f was repurposed to not mean the definition of f you saw right below but you have to keep in mind that it’s actually a g(f)). See my rants here.

Notational ambiguity, even if resolvable, is NOT helpful at all here especially when there are so many level of abstractions squeezed into so few symbols! People jumping into learning the subject quickly for the first time should not have unnecessarily keep track of the obscure scoping rules in their head to resolve the ambiguities!

Loading

Using Dart’s C-style syntax to make chain of lambda functions concrete

Start with an example lambda calculus expression f\equiv\lambda x.(y+x). It is:

  • [programming view] a function with x as an input argument and it uses y from the outer workspace (called a free-variable). f(var x) { return y+x; }
  • [mathematical view] can also be written as f(x) = y+x where y is seen as a fixed/snapshotted value relative to the expression.
g(int y) => (int x) => y + x

Lambda calculus is right-associative

g(int y) => ( (int x) => y + x )

Unrolling in C-style it will give better insights to the relationships between layers

g(int y) {
  return (int x) { return y + x; };
}

Note that (int x) { return y + x; } is a functor. To emphasize this, the same code can be rewritten by assigning the name f to it and returning f:

g(int y) {
  f(int x) { return y + x; };
  return f;
}

Use the C++11 style syntax so that it doesn’t look like nested function implementation body instead of a functor nested inside a function:

g(int y) {
  var f = (int x) { return y + x; };
  return f;
}

Note that conceptually what we are doing with the wrapper cascade is indeed nested functions. However, in the wrapper, we spit out a functor (which did most of the partial work) so the user can endow/evaluating it with the last piece of needed info/input:

g(int y) {
  f(int x) { 
   return y + x;
  };
  return f;
}

More commonly as seen in Dart docs, this formatting shows a (binder/capturing) wrapper function returning its own nested function:

g(int y) {
  return (int x) { 
           return y + x;
         };
}

Loading

Styles for (Lambda) anonymous function (named or unnamed) in Dart Language

Official Dart docs is sometimes too simple to provide ultimate answers for various language questions. I discoverd an alternative syntax for named lambda here and here. In Dart,

(args) => expr

is the shorthand for

(args) { 
  return expr 
}

So the Lambda-style (C++11) names variable by assigning it to a messy functor type (inferred with var in Dart and auto in C++11):

var f = (args) => expr

Which can also be rewritten using C-style function pointer-like declaration except

  1. the body expr is included (which is not allowed in C),
  2. return must be explicit in curly braces block.
  3. arrow notation => takes whatever the expr evaluates to (which can be null for statements like print)
  4. it’s simply the function name f in Dart instead of pointer (*f) in C
[optional return type] f(args) { 
  return expr;
}

which can be simplified with the arrow operator

[optional return type] f(args) => { expr }

Loading