How to avoid a databinding / events hell on a complex screen? - c#

This is more of an architecture / design question.
I have run into a few projects in the past written in WPF/Windows Forms, etc. that have complex screens with a lot of fields and these fields are connected to each other (their values depend on each other with some logic involved).
These projects I have taken on after they were implemented, and I found a lot of events / data bind hell - what I mean by this is that because all these fields are depending on others they have implemented INotifyPropertyChanged and other fields are being modified as a result. This causes the same fields being updated 5-6 times when the screen loads and the order in which fields are populated causes horrible bugs. (For example, Date was set before Job Type, instead of after Job Type, so I end up with a different Job Fee.)
To make matters worse, some hacks are implemented on UI events (for example, DropDown changed to update field X) while others are in the domain model that the UI binds to.
Basically, it's a huge mess, and I just want to know what the best way to implement something like this is if I was to start from scratch. Or is it a good idea to avoid such a complex screen in the first place?

I would try to keep the business logic out of the property setters as much as possible.
First of all, if several properties are needed for one calculation, I'd write one method that does the calculation, and call that method when appropriate. E.g. if all different combinations of property values make sense, one could just call the method in the setters of each property, making sure that the same code runs any time one of the properties is changed. If you only can evaluate special combinations of property values, you could either implement a command and let the user decide when to calculate the resulting changes, or you could provide feedback through validation, and only evaluate the property changes if the combination is valid. If there are several interdependent properties, I often use a "ChangeInitiator" variable to indicate what property has changed, so that it is clear in the calculation method which property is responsible for the change and which others should change as a result. Basically, this is the same as doing one part of the calculation in each property setter, but I find that it helps me to keep an overview of things if the different parts of the relationship are all in one method.
In a program I wrote once, I had some calculations running on a background thread periodically, so I would just set a flag whenever a piece of data changed that required a new calculation, and do all the updates based on a timer every second or so... that could also help you get the logic more straight, and it avoids to have the calculation run several times for one set of related changes.
With regard to change notification, I'd really try to only use it for UI data binding.

We have fairly complex UIs (including several related fields of different types in, say for example a Row in a DataGrid) and the MVVM pattern has worked pretty well for us. All the properties coming from the Model and exposed to the View that have complex logic related are "wrapped" by an equivalent property in the ViewModel, which has no Backing Field, but rather points directly to the Model:
public class SomeComplexViewModel
{
public SomeModel Model {get;set;}
public string SomeCrazyProperty
{
get
{
return Model.SomeCrazyProperty;
}
{
Model.SomeCrazyProperty = value;
//... Some crazy logic here, potentially modifying some other properties as well.
}
}
}
<TextBox Text="{Binding SomeCrazyProperty}"/>
This removes the "initial value" problem, as the initial value read by the Binding is actually the real value coming from the Model, and thus the logic placed in the Setter is executed only when needed.
Then, for dummy properties (which have no logic behind), we bind directly from the View to the Model:
<TextBox Text="{Binding Model.SomeRegularProperty}"/>
This reduces the bloat in the ViewModel.
With regard to events in the code behind, I totally avoid that. My code behind files are almost always one InitializeComponent() and nothing else.
Only View-Specific logic is placed in the code behind (such as animations stuff, etc), when it cannot be directly done in XAML, or is easier to do in code (which is not the case most of the time).
Edit:
It's important to mention that the winforms binding capabilities are a joke compared to the XAML-based ones. could that be the cause you're seeing those horrible messes in those projects?

Related

How to raise PropertyChangedEventHandler when a sub property is changed in WPF?

I have a property called PriceChangeInPercentWeekly which has other properties like LastOpenPrice, LastClosePrice, etc.
I am changing these continuously. But how can I raise PropertyChangedEventHandler for PriceChangeInPercentWeekly when a sub-property is changed?
When PriceChangeInPercentWeekly itself is changed, I raised it using my custom SetField function. But how can I do this for sub-properties. Because sub-properties do not know the instanced class, right?
PriceChange priceChangeInPercentWeekly;
public PriceChange PriceChangeInPercentWeekly
{
get => this.priceChangeInPercentWeekly;
set => SetField ( ref this.priceChangeInPercentWeekly, value,
"PriceChangeInPercentWeekly"
);
}
Based on the comments, it sounds like you have a issue with properly binding in a collection case. When dealing with collections in MVVM, there are actually 3 seperate kinds of bindings/Change Notifications you need. If any one is missing, you end up getting issues with non-updates pretty quickly:
Whatever class is holding LastOpenPrice, LastClosePrice, PriceChange and the like - it needs change notificaiton on each and every property. Whatever template you use to display that class, needs to properly bind to each of those properties.
You need change notificaiton if anything is added or removed from the Collection. That is the only change notification ObservableCollection<ClassThatHoldsThosePrices> takes care off.
Whatever property exposes the ObservableList<ClassThatHoldsThosePrices> or its Collection View needs change notificaiton as well. ObservableCollections do not support bulk additions. With a big changes you usually have to build a new list in the background, then replace the whole collection.
At this point it usually becomes a question of properly setting up (or automatically having set up) the 3 kinds of bindings. The proper MVVM pattern is not a easy thing to learn and it can not tell if you did it properly, so it is hard to say where the issues could be. About 8 years ago I wrote a short introducing which I think should still be okay-ish for learning the basics: https://social.msdn.microsoft.com/Forums/vstudio/en-US/b1a8bf14-4acd-4d77-9df8-bdb95b02dbe2/lets-talk-about-mvvm?forum=wpf

Is there a way to bind properties together?

I've recently needed to synchronize values from a higher up view model (bound to a custom tab content control) to the base one (bound to the host view) for the CanExecute delegate in the host view to use. The base model has the instance of the top one, among others.
The only thing that I can think of for it to know when a value changes higher up is to subscribe to the PropertyChanged event. But, that seems excessive considering how many times that event would get fired for all of the other properties. It also doesn't feel right for MVVM (but I may be wrong).
Right now, I'm setting it all in a central method in the manager class where the magic happens to make sure the values match up:
If setDealable Then multilegViewModel.IsDealable = isDealable
multilegViewModel.IsIndicative = (Not isDealable)
' [...]
tktViewModel.IsCommandOtherEnabled = (Not isDealable)
tktViewModel.IsCommandBuyEnabled = multilegViewModel.IsBuyButtonEnabled
tktViewModel.IsCommandSellEnabled = multilegViewModel.IsSellButtonEnabled
tktViewModel.IsDealable = isDealable
tktViewModel.IsIndicative = (Not isDealable)
' [...]
But, smaller sets of the "multilegViewModel" properties are being set elsewhere, so I have to find them all and add the copy over. The risk is having another developer leave out such a pairing somewhere.
So, does anyone have ideas other than one view model subscribing to the other to ensure that the values always get set?
Note: The dual language tags are on purpose. It's a mixed language solution, such as the manager being VB.NET, but the models are C#, so I accept suggestions in either one.
UPDATE: I've changed my approach and greatly simplified it, so my initial reason for needing this is no longer valid. But, I may have an edge case or two that could still benefit from this.

Complex WPF UserControl in MVVM application

Having taken over maintenance of an existing WPF application, I was horrified to discover that two of the Views and ViewModels had large blocks of near-identical code. Obviously, I want to refactor this so they can both reuse a single block of functionality, but I'm not sure how best to go about it, architecturally.
The identikit code deals with processing UI data from a tab. However I split this, it is essential that the code in the other tabs (which is different in the two cases) has access to the properties and objects of the tab I need to split out.
To further complicate matters, the replicated code needs database access. We've got a repository object that handles this. Normally when creating new objects, I've been making them testable by passing a copy of the repository into the constructor. However, if I do that in this case I'll have two copies of the repository object - one in the ViewModel, one in the split out code - needing to handle the same data, which is going to cause concurrency issues.
My first thought was to make a UserControl for this, but the more I think about this, the more problematic the two issues above seem to be.
The other option I've considered is just to make a Helper class to do some of the identical processing. But that's only going to partially solve the problem as some identical UI code (raising property changed events, XAML, etc) is still going to be in both Views/ViewModels.
What's the best approach here? Is there a way I can get past the repository/access issues and make a UserControl? Or is than an alternative based on Interfaces or Inheritance I haven't considered?
EDIT - Code was asked for. It's a bit complex to give a comprehensive example, but here's a snippet from each VM:
public void CheckOrderHist(int months)
{
var endDate = DateTime.Today.AddMonths(months);
Dictionary<OrderHistory, bool> orders = new Dictionary<OrderHistory, bool>();
this.ordersToExclude.Clear();
foreach (var kvp in rep.OrderHistories.GetRecent(months))
{
if (kvp.Key.MailingDate >= endDate)
{
orders.Add(kvp.Key, true);
this.ordersToExclude.Add(((OrderHistory)kvp.Key).OrderID);
}
else
{
orders.Add(kvp.Key, false);
}
}
BuildOrderExclusionsOnCount(); //code in this is near-identical across VM's too
OrderHistoryMonths = Math.Abs(months); //OrderHistoryMonths is a property on the ViewModel
OnPropertyChanged("MajorityOrderBoolean");
}
And in the other VM:
private void CheckOrderHist(int months)
{
var endDate = DateTime.Today.AddMonths(-months);
ObservableCollection<Tuple<OrderHistory, bool>> orders = new ObservableCollection<Tuple<OrderHistory, bool>>();
this.ordersToExclude.Clear();
foreach (var tuple in rep.OrderHistories.GetRecent(-months))
{
if (tuple.Item1.MailingDate >= endDate)
{
orders.Add(new Tuple<OrderHistory,bool>(tuple.Item1, true));
this.ordersToExclude.Add(tuple.Item1.OrderID);
}
else
{
orders.Add(new Tuple<OrderHistory, bool>(tuple.Item1, false));
}
}
BuildOrderExclusionsOnCount(); //code in this is near-identical across VM's too
OrderHistoryMonths = months; //OrderHistoryMonths is a property on the ViewModel
OnPropertyChanged("OrderHistories");
OnPropertyChanged("GroupedOrders");
}
This illustrates the problem nicely - the function is essentially the same, but one uses a Dictionary and the other a Tuple (there's no good reason for this - they both need a Tuple really, for ease of ordering). And one arbitrarily takes a negative int parameter, and the other a positive.
Both contain different OnPropertyChanged events, and will use different copies of the repository object, making it hard to properly separate them using a Helper class. Yet putting it in a UserControl would isolate them from OrderHistoryMonths on the main ViewModel.
If I'm hearing the current comments right, the best solution here is to farm out the main ForEach loop to a helper class, and just put up with the rest of the duplication?
By all means, extract common logic where possible to a new 'helper' class that each ViewModel can construct; this is the standard pattern of re-use through composition. The code you've shown in your question is a good candidate for this kind of refactoring.
As far as boilerplate, though, it's a bit trickier. This is something that is difficult to address in general and must be examined on a case-by-case basis. There are various ways to simplify property changed notification, for instance (helper methods encapsulating property updates, AOP, etc.) but these are generally part of your MVVM framework and embraced application-wide. As far as XAML duplication, you can often use Styles, Data Templates and Value Converters to improve things, but again, it requires a careful analysis of your particular code base to identify the patterns that may merit this treatment. If you have more specific examples that you think are clear duplicates, but aren't sure how to refactor, those may make good questions.

Why should I use an automatically implemented property instead of a field?

Between these two:
With Property:
class WithProperty
{
public string MyString {get; set;}
}
With Field:
class WithField
{
public string MyString;
}
Apparently I'm supposed to pick the first one. Why?
I've heard the argument that the point here is to allow interface changes, but
if I have the second one, and change it to the first one, no other code should
ever have to change. When recompiled everything's just going to point to the
property instead.
Am I missing something important here?
The most important difference is the fact, that if you use a field, and later need to change it to a property (say, to enforce some validation), then all libraries calling your code will need to be recompiled. It's true that you can compile the exact same code if the name stays the same - but the consumers of your code will still need to be recompiled. This is because the IL generated to get the value is different between a field and a property. If it already is a property, you can make a change without forcing consumers of your code to change.
This may or may not be an issue for you. But the property is almost the same amount of code, and is considered best practice. I would always go for the property.
The property can be changed later if you need to add validation or other logic without breaking other assemblies.
Also, the property can be used with databinding.
The important part you are missing is the gravity of this statement:
When recompiled
When your code point to a field and you change it to point to a property of the same name, the C# itself doesn't change, but the resulting IL does - it generates a method call to the getter or setter as appropriate.
Not every app has all of it's pieces contained in a single distributed unit. Many apps rely on interfaces for pluggability/expandability. If you have an app with an interface to a field and you want to change it to a property to take advantage of the power of properties, the app has to be recompiled and redistributed. You might as well just make it a property in the first place.
With a property, you can easily extend it to include new logic.
For example, if you need to add validation logic to the set.
This article goes into several additional reasons why you should prefer properties:
http://csharpindepth.com/Articles/Chapter8/PropertiesMatter.aspx

Properties vs Methods

Quick question: When do you decide to use properties (in C#) and when do you decide to use methods?
We are busy having this debate and have found some areas where it is debatable whether we should use a property or a method. One example is this:
public void SetLabel(string text)
{
Label.Text = text;
}
In the example, Label is a control on a ASPX page. Is there a principle that can govern the decision (in this case) whether to make this a method or a property.
I'll accept the answer that is most general and comprehensive, but that also touches on the example that I have given.
From the Choosing Between Properties and Methods section of Design Guidelines for Developing Class Libraries:
In general, methods represent actions and properties represent data. Properties are meant to be used like fields, meaning that properties should not be computationally complex or produce side effects. When it does not violate the following guidelines, consider using a property, rather than a method, because less experienced developers find properties easier to use.
Yes, if all you're doing is getting and setting, use a property.
If you're doing something complex that may affect several data members, a method is more appropriate. Or if your getter takes parameters or your setter takes more than a value parameter.
In the middle is a grey area where the line can be a little blurred. There is no hard and fast rule and different people will sometimes disagree whether something should be a property or a method. The important thing is just to be (relatively) consistent with how you do it (or how your team does it).
They are largely interchangeable but a property signals to the user that the implementation is relatively "simple". Oh and the syntax is a little cleaner.
Generally speaking, my philosophy is that if you start writing a method name that begins with get or set and takes zero or one parameter (respectively) then it's a prime candidate for a property.
Searching through MSDN, I found a reference on Properties vs Methods that provides some great guidelines for creating methods:
The operation is a conversion, such as Object.ToString.
The operation is expensive enough that you want to communicate to the
user that they should consider caching
the result.
Obtaining a property value using the get accessor would have an observable
side effect.
Calling the member twice in succession produces different results.
The order of execution is important. Note that a type's properties should
be able to be set and retrieved in any
order.
The member is static but returns a value that can be changed.
The member returns an array. Properties that return arrays can be
very misleading. Usually it is
necessary to return a copy of the
internal array so that the user cannot
change internal state. This, coupled
with the fact that a user can easily
assume it is an indexed property,
leads to inefficient code.
If you're setting an actual property of your object then you use a property.
If you're performing a task / functionality then you use a method.
In your example, it is a definite property being set.
If however, your functionality was to AppendToLabel then you would use a method.
Properties are a way to inject or retrieve data from an object. They create an abstraction over variables or data within a class. They are analogous to getters and setters in Java.
Methods encapsulate an operation.
In general I use properties to expose single bits of data, or small calculations on a class, like sales tax. Which is derived from the number of items and their cost in a shopping cart.
I use methods when I create an operation, like retrieving data from the database. Any operation that has moving parts, is a candidate for a method.
In your code example I would wrap it in a property if I need to access it outside it's containing class:
public Label Title
{
get{ return titleLabel;}
set{ titleLabel = value;}
}
Setting the text:
Title.Text = "Properties vs Methods";
If I was only setting the Text property of the Label this is how I would do it:
public string Title
{
get{ return titleLabel.Text;}
set{ titleLabel.Text = value;}
}
Setting the text:
Title = "Properties vs Methods";
Symantically properties are attributes of your objects.
Methods are behaviors of your object.
Label is an attribute and it makes more sense to make it a property.
In terms of Object Oriented Programming you should have a clear understanding of what is part of behavior and what is merely an attribute.
Car { Color, Model, Brand }
A car has Color, Model and Brand attributes therefore it does not make sense to have a method SetColor or SetModel because symantically we do not ask Car to set its own color.
So if you map the property/method case to the real life object or look at it from symantic view point, your confusion will really go away.
You need only look at the very name... "Property". What does it mean? The dictionary defines it in many ways, but in this case "an essential or distinctive attribute or quality of a thing" fits best.
Think about the purpose of the action. Are you, in fact, altering or retrieving "an essential or distinctive attribute"? In your example, you are using a function to set a property of a textbox. That seems kind of silly, does it not?
Properties really are functions. They all compile down to getXXX() and setXXX(). It just hides them in syntactic sugar, but it's sugar that provides a semantic meaning to the process.
Think about properties like attributes. A car has many attributes. Color, MPG, Model, etc.. Not all properties are setable, some are calculatable.
Meanwhile, a Method is an action. GetColor should be a property. GetFile() should be a function. Another rule of thumb is, if it doesn't change the state of the object, then it should be a function. For example, CalculatePiToNthDigit(n) should be a function, because it's not actually changing the state of the Math object it's attached to.
This is maybe rambling a bit, but it really boils down to deciding what your objects are, and what they represent. If you can't figure out if it should be a property or function, maybe it doesn't matter which.
Properties should only be simple set and get one liners. Anything more and it should really be moved to a method. Complex code should always be in methods.
I only use properties for variable access, i.e. getting and setting individual variables, or getting and setting data in controls. As soon as any kind of data manipulation is needed/performed, I use methods.
As a matter of design Properties represent Data or Attributes of class object, While methods are actions or behaviors of class object.
In .Net, world there are other implications of using Properties:
Properties are used in Databinding, while get_ / set_ methods are not.
XML serialization user properties as natural mechanism of serilization.
Properties are accessed by PropertyGrid control and intern ICustomTypeDescriptor, which can be used effectively if you are writing a custom library.
Properties are controlled by Attributes, one can use it wisely to design Aspect Oriented softwares.
Misconceptions (IMHO) about Properties' usage:
Used to expose small calculations: ControlDesigner.SelectionRules's get block runs into 72 lines!!
Used to expose internal Data structures: Even if a property does not map to an internal data member, one can use it as property, if its an attribute of your class. Viceversa, even if its an attribute of your class properties are not advisable, to return array like data members (instead methods are used to return deep copy of members.)
In the example here it could have been written, with more business meaning as:
public String Title
{
set { Label.Text = text; }
}
Also big plus for Properties is that value of property can be seen in Visual Studio during debugging.
I prefer to use properties for add/set methods with 1 parameter. If parameters are more, use methods.
Properties are really nice because they are accessible in the visual designer of visual studio, provided they have access.
They use be used were you are merely setting and getting and perhaps some validation that does not access a significant amount of code. Be careful because creating complex objects during validation is not simple.
Anything else methods are the preferred way.
It's not just about semantics. Using properties inappropriate start having weirdness occur in the visual studio visual designer.
For instance I was getting a configuration value within a property of a class. The configuration class actually opens a file and runs an sql query to get the value of that configuration. This caused problems in my application where the configuration file would get opened and locked by visual studio itself rather than my application because was not only reading but writing the configuration value (via the setter method). To fix this I just had to change it to a method.
Here is a good set of guidelines for when to use properties vs methods from Bill Wagner
Use a Property when all these are true:
The getters should be simple and thus unlikely to throw exceptions. Note that this implies no network (or database) access. Either might fail, and therefore would throw an exception.
They should not have dependencies on each other. Note that this would include setting one property and having it affect another. (For example, setting the FirstName property would affect a read-only FullName property that composed the first name + last name properties implies such a dependency )
They should be settable in any order
The getter does not have an observable side effect Note this guideline doesn't preclude some forms of lazy evaluation in a property.
The method must always return immediately. (Note that this precludes a property that makes a database access call, web service call, or other similar operation).
Use a method if the member returns an array.
Repeated calls to the getter (without intervening code) should return the same value.
Repeated calls to the setter (with the same value) should yield no difference from a single call.
The get should not return a reference to internal data structures (See item 23). A method could return a deep copy, and could avoid this issue.
*Taken from my answer to a duplicate question.
This is simple.
1: use property when you want your data should be validated before storing in field. So in this way property provides encapsulation for your fields. Because if you leave your fields public end user may assign any value which may or may not be valid as per your business requirement like age should be greater than 18. So before value is store corresponding field we need to check its validity. In this way properties represent data.
2: Use method when you want perform some action like you are supplying some data as parameter and your method is doing some processing on the basis of supplied values and returning processed value as output. Or you want to change value of some field by this calculation. "In this way method represents action".
I come from java an i used get.. set.. method for a while.
When i write code, i don't ask to my self: "accessing this data is simple or require a heavy process?" because things can change (today retrive this property is simple, tomonrow can require some or heavy process).
Today i have a method SetAge(int age) tomonrow i will have also method SetAge(date birthdate) that calculate the age using the birthdate.
I was very disappointed that the compiler transform property in get and set but don't consider my Get... and Set.. methods as the same.

Categories