Silverlight LINQtoSQL: one big dataclass, or several small ones? - silverlight

I'm new to Silverlight, but being dumped right into the fray - good way to learn I suppose :o)
Anyway, the webapp I'm working on has a relatively complex database structure that represents various object types that are linked to each other, and I was wondering 2 things:
1- What is the recommended approach when it comes to dataclasses? Have just one big dataclass, or try and separate it into several smaller dataclasses, keeping in mind they will need to reference each other?
2- If the recommended approach is to have several dataclasses, how do you define the inter-dataclasses references?
I'm asking because I did a small test. In my DB (simplified here, real model is more complex but that's not important), I have a table "Orders" and a table "Parameters". "Orders" has a foreign key on "Parameters". What I did is create 2 dataclasses.
The first one, ParamClass, were I dropped the "Parameters" table only, so I can have a nice "parameter" class. I then created a simple service to add basic SELECT and INSERT functionality.
The second one, OrdersClass, where I dropped both tables, so that the relation between the tables would automatically create a "EntityRef<parameter>" variable inside the "order" class. I then removed the "parameters" class that was automatically created in the OrdersClass dataclass, since the class has already been declared in the ParamClass dataclass. Again I created a small service to test it.
So far so good, it builds happily. The problem is that when I try to handle things on the application code, I added service references for both dataclasses, but it is not happy doing something like:
OrdersServiceReference.order myOrder = new OrdersServiceReference.order();
myOrder.parameter = new ParamServiceReference.parameter(); //<-PROBLEM IS HERE
It comlpains that it cannot implicitly convert from type 'MytestDC.ParamServiceReference.parameter' to 'MytestDC.OrdersServiceReference.parameter'
Do I somehow need to declare some sort of reference to ParamClass from OrdersClass, or how do I "convert" one to the other?
Is this even a recommended and efficient way of doing this?
Since it's a team-project, I initially wanted to separate the dataclasses so that they (and their services) can be easily checked out by one member without checking out the whole entire dataclass.
Any help appreciated!
PS: using Silverlight 4, in case that's important

Based on the widely accepted Single Responsability Principle (SRP), a class should always be responsible for one task, and one task only.
That pretty much invalidates your "one big dataclass" approach.
I would always recommend smaller, more manageable bits that can be combined, instead of one humonguous class that does everything (except brew coffee for you).
Resources for the SRP:
Wikipedia on SRP
OODesign: Single Responsibility Principle
ObjectMentor: list of articles on good app design - which has a few links to PDF documents, like this one on SRP written by Robert C. Martin - the "guru" on proper OO design

OK, some more research let me to this: it is not simple to separate classes from a relational model using LINQtoSQL. I ended up switching to an Entity Framework approach, which itself doesn't deal with it gracefully (see here and there, for example), but at least it solved another major problem I had with LINQtoSQL.
There are other ORMs out there that are apparently much more capable at this (NHibernate comes up often in recommendations), unfortunately, I don't have time to investigate them now, being under such a tight deadline.
As for the referencing, it was quite simple, change the line to:
myOrder.parameter = new OrderServiceReference.parameter();
even though I removed the declaration from that dataclass.
Hope this helps someone!

Related

Game player object design

I'm supporting a server for an online card game and while thinking about refactoring it into a better state I have found myself unable to decide what is a proper object model for my needs.
I have a Player class which has a lot of attributes. The first problem is just that - the class is too big. The second problem is that I don't know how to refactor it. I will list some of the attributes and issues with these.
Some attributes are very tightly bound to a player: nick, email, last login &c. These, I suppose, are to be kept directly in the player class and in the same table in the DB.
Now, some attributes are a little more difficult, like money and gold amount. The problem with these is that they are historically stored in a different table, there might be some more currencies later on and that they MUST be synched into the database at their own pace.
Third category of attributes are loosely coupled to the player, like status string, experience, achievements, statistics &c. These are stored in different tables in the DB and MUST be stored, retrieved, cached and synchronized at their own pace.
Note that one of the big problems here is that we have to implement relatively complex database synchronization schemes because we have a lot of online players and our game is soft-realtime and we have to make load on the DB as low as possible.
My questions are:
How to determine which attributes to store within a player class and which not to? Say, experience, nickname, money amount?
When one has some attributes that may be grouped together like (strength, agility, endurance, &c.) and (handItem, headItem, feetItem, weapon) when they should be grouped and when not?
What to do with complex database synchronization schemes? Make a separate model for each attribute that needs to be synched independently or make some DataManager classes to take them apart and work with them?
What to do with the need for a class to have several different "data representations" for external consumers? Like XML, Json, another XML for some external service, human-readable string, &c.
I'm sorry if my questions are bogus, I'm not really good at OOP design, I'm more an FP guy. And my English is not very good =).
There is no "limit" to what you can store in a player class. As long as it is concerning him and him only, it should be in his class. But one thing you should consider is to make several player classes. The idea is : if you don't need is, don't query it. You may have PlayerView_Small, PlayerBuying, PlayerFighting, PlayerSettings (depending on your game, they may not be fulfilling the exact same purpose)... This way for each "need" of info on a player, you only load the player data you need, and can handle it properly. Also, you may use inheritance if some class is only a more detailed version of the other.
If you are talking about the class, it may be in a sub-class PlayerAttributes of which an instance is contained into PlayerFighting and PlayerView_Detailed. In the database, it might be interesting to store it as a string (conveniently outputted by our class, and accepted in constructor), to avoid having too much fields, but you will lose the sorting ability. That's probably not a problem in our case, but might be in some others.
Blank for now, I don't understand where there is synchronization, will edit when informed.
In your PlayerViewDetailInfo(or in your PlayerAllData depending what you need), you place some methods such as ToXmlClient1(), ToJson(), ToHumanReadableString() (although that might be a bit confusing to the eye, you should consider HTML^^). The class having the method should be the class with the least (but sufficient to provide the answer) data. When requested, you load the Player... which has the method giving the correct output, and you write it directly in the response.

Self Tracking Entities Traffic Optimization

I'm working on a personal project using WPF with Entity Framework and Self Tracking Entities. I have a WCF web service which exposes some methods for the CRUD operations. Today I decided to do some tests and to see what actually travels over this service and even though I expected something like this, I got really disappointed. The problem is that for a simple update (or delete) operation for just one object - lets say Category I send to the server the whole object graph, including all of its parent categories, their items, child categories and their items, etc. I my case it was a 170 KB xml file on a really small database (2 main categories and about 20 total and about 60 items). I can't imagine what will happen if I have a really big database.
I tried to google for some articles concerning traffic optimization with STE, but with no success, so I decided to ask here if somebody has done something similar, knows some good practices, etc.
One of the possible ways I came out with is to get the data I need per object with more service calls:
return context.Categories.ToList();//only the categories
...
return context.Items.ToList();//only the items
Instead of:
return context.Categories.Include("Items").ToList();
This way the categories and the items will be separated and when making changes or deleting some objects the data sent over the wire will be less.
Has any of you faced a similar problem and how did you solve it or did you solve it?
We've encountered similiar challenges. First of all, as you already mentioned, is to keep the entities as small as possible (as dictated by the desired client functionality). And second, when sending entities back over the wire to be persisted: strip all navigation properties (nested objects) when they haven't changed. This sounds very simple but is not at all trivial. What we do is to recursively dig into the entities present in trackable collections of say the "topmost" entity (and their trackable collections, and theirs, and...) and remove them when their ChangeTracking state is "Unchanged". But be carefull with this, because in some cases you still need these entities because they have been removed or added to trackable collections of their parent entity (so then you shouldn't remove them).
This, what we call "StripEntity", is also mentioned (not with any code sample or whatsoever) in Julie Lerman's - Programming Entity Framework.
And although it might not be as efficient as a more purist kind of approach, the use of STE's saves a lot of code for queries against the database. We are not in need for optimal performance in a high traffic situation, so STE's suit our needs and takes away a lot of code to communicate with the database. You have to decide for your situation what the "best" solution is. Good luck!
You can find an Entity Framework project item at http://selftrackingentity.codeplex.com/. With version 0.9.8, I added a method called GetObjectGraphChanges() that returns an optimized entity object graph with only objects that have changes.
Also, there are two helper methods: EstimateObjectGraphSize() and EstimateObjectGraphChangeSize(). The first method returns the estimate size of the whole entity object along with its object graph; and the later returns the estimate size of the optimized entity object graph with only object that have changes. With these two helper methods, you can decide whether it makes sense to call GetObjectGraphChanges() or not.

Django Advantage forms.Form vs forms.ModelForm

There is a question very similar to this but I wanted to ask it in a different way.
I am a very customized guy, but I do like to take shortcuts at times. So here it goes.
I do find these two classes very similar although one "helps" the programmer to write code faster or have less code/repeating code. Connecting Models to Forms sounds like an obvious thing to do. One thing that is not particularly clear in the docs using a ModelForm. What happens if you need to add extra fields that are not in the Model or some way connected to another Model?
I guess you could subclass that out and make it work, but does that really help you save time than just manually doing it with a Form?
So next question may not have a definite answer if I do subclass it out, and use ModelForm. Is ModelForm particularly faster than Form? Does it still use the same Update techniques or is binding significantly faster in one or the other?
If you want a form across two models, you got a couple options:
1) create two modelforms, save each individually when posted, and if one of the two depends on the other (i.e. foreignkey), set that in your view before saving.
2) try Django's inline formset: http://docs.djangoproject.com/en/dev/topics/forms/modelforms/#using-an-inline-formset-in-a-view
3) Add non-model fields to your modelform. On a ModelForm, you can add fields that are not tied to your model. They are available in cleaned_data as any other field would, be but are simply ignored when the model is saved.
One advantage that ModelForm's have over Form's is you can specify the ordering of fields (searching for how to order Form fields brought to your post incidentally). Obvious other advantages are you don't have to rewrite your model saving code

Using Flyweight Pattern in database-driven application

Can anyone please give me any example of situation in a database-driven application where I should use Flyweight pattern?
How can I know that, I should use flyweight pattern at a point in my application?
I have learned flyweight pattern. But not able to understand an appropriate place in my database-driven business applications to use it.
Except for a very specialized database application, the Flyweight might be used by your application, but probably not for any class that represents an entity which is persisted in your database. Flyweight is used when there otherwise might be a need for so many instantiations of a class that if you instantiated one every discrete time you needed it performance would suffer. So instead, you instantiate a much smaller number of them and reuse them for each required instance by just changing data values for each use. This would be useful in a situation where, for example, you might have to instantiate thousands of such classes each second, which is generally not the case for entities persisted in a database.
You should apply any pattern when it naturally suggests itself as a solution to a concrete problem - not go looking for places in your application where you can apply a given pattern.
Flyweight's purpose is to address memory issues, so it only makes sense to apply it after you have profiled an application and determined that you have a ton of identical instances.
Colors and Brushes from the Base Class Library come to mind as examples.
Since a very important part of Flyweight is that the shared implementation is immutable, good candidates in a data-driven application would be what Domain-Driven Design refers to as Value Objects - but it only becomes relevant if you have a lot of identical values.
[Not a DB guy so this is my best guess]
The real bonus to the flyweight pattern is that you can reuse data if you need to; Another example is word processing where ideally you would have an object per "character" in your document, but that wuld eat up way too much memory so the flyweight memory lets you only store one of each unique value that you need.
A second (and perhaps simplest) way to look at it is like object pooling, only you're pooling on a "per-field" level as opposed to a "per-object" level.
In fact, now that i think about it, it's not unlike using a (comparatively small) chunk of memory in c(++) so store some raw data which you do pointer manipulation to get stuff out of.
[See this wikpedia article].

Is it a bad practice to have multiple classes in the same file?

I used to have one class for one file. For example car.cs has the class car. But as I program more classes, I would like to add them to the same file. For example car.cs has the class car and the door class, etc.
My question is good for Java, C#, PHP or any other programming language. Should I try not having multiple classes in the same file or is it ok?
I think you should try to keep your code to 1 class per file.
I suggest this because it will be easier to find your class later. Also, it will work better with your source control system (if a file changes, then you know that a particular class has changed).
The only time I think it's correct to use more than one class per file is when you are using internal classes... but internal classes are inside another class, and thus can be left inside the same file. The inner classes roles are strongly related to the outer classes, so placing them in the same file is fine.
In Java, one public class per file is the way the language works. A group of Java files can be collected into a package.
In Python, however, files are "modules", and typically have a number of closely related classes. A Python package is a directory, just like a Java package.
This gives Python an extra level of grouping between class and package.
There is no one right answer that is language-agnostic. It varies with the language.
One class per file is a good rule, but it's appropriate to make some exceptions. For instance, if I'm working in a project where most classes have associated collection types, often I'll keep the class and its collection in the same file, e.g.:
public class Customer { /* whatever */ }
public class CustomerCollection : List<Customer> { /* whatever */ }
The best rule of thumb is to keep one class per file except when that starts to make things harder rather than easier. Since Visual Studio's Find in Files is so effective, you probably won't have to spend much time looking through the file structure anyway.
No I don't think it's an entirely bad practice. What I mean by that is in general it's best to have a separate file per class, but there are definitely good exception cases where it's better to have a bunch of classes in one file. A good example of this is a group of Exception classes, if you have a few dozen of these for a given group does it really make sense to have separate a separate file for each two liner class? I would argue not. In this case having a group of exceptions in one class is much less cumbersome and simple IMHO.
I've found that whenever I try to combine multiple types into a single file, I always end going back and separating them simply because it makes them easier to find. Whenever I combine, there is always ultimately a moment where I'm trying to figure out wtf I defined type x.
So now, my personal rule is that each individual type (except maybe for child classes, by which a mean a class inside a class, not an inherited class) gets its own file.
Since your IDE Provides you with a "Navigate to" functionality and you have some control over namespacing within your classes then the below benefits of having multiple classes within the same file are quite worth it for me.
Parent - Child Classes
In many cases i find it quite helpful to have Inherited classes within their Base Class file.
It's quite easy then to see which properties and methods your child class inherits and the file provides a faster overview of the overall functionality.
Public: Small - Helper - DTO Classes
When you need several plain and small classes for a specific functionality i find it quite redundant to have a file with all the references and includes for just a 4-8 Liner class.....
Code navigation is also easier just scrolling over one file instead of switching between 10 files...Its also easier to refactor when you have to edit just one reference instead of 10.....
Overall breaking the Iron rule of 1 class per file provides some extra freedom to organize your code.
What happens then, really depends on your IDE, Language,Team Communication and Organizing Skills.
But if you want that freedom why sacrifice it for an iron rule?
The rule I always go by is to have one main class in a file with the same name. I may or may not include helper classes in that file depending on how tightly they're coupled with the file's main class. Are the support classes standalone, or are they useful on their own? For example, if a method in a class needs a special comparison for sorting some objects, it doesn't bother me a bit to bundle the comparison functor class into the same file as the method that uses it. I wouldn't expect to use it elsewhere and it doesn't make sense for it to be on its own.
If you are working on a team, keeping classes in separate files make it easier to control the source and reduces chances of conflicts (multiple developers changing the same file at the same time). I think it makes it easier to find the code you are looking for as well.
It can be bad from the perspective of future development and maintainability. It is much easier to remember where the Car class is if you have a Car.cs class. Where would you look for the Widget class if Widget.cs does not exist? Is it a car widget? Is it an engine widget? Oh maybe it's a bagel widget.
The only time I consider file locations is when I have to create new classes. Otherwise I never navigate by file structure. I Use "go to class" or "go to definition".
I know this is somewhat of a training issue; freeing yourself from the physical file structure of projects requires practice. It's very rewarding though ;)
If it feels good to put them in the same file, be my guest. Cant do that with public classes in java though ;)
You should refrain from doing so, unless you have a good reason.
One file with several small related classes can be more readable than several files.
For example, when using 'case classes', to simulate union types, there is a strong relationship between each class.
Using the same file for multiple classes has the advantage of grouping them together visually for the reader.
In your case, a car and a door do not seem related at all, and finding the door class in the car.cs file would be unexpected, so don't.
As a rule of thumb, one class/one file is the way to go. I often keep several interface definitions in one file, though. Several classes in one file? Only if they are very closely related somehow, and very small (< 5 methods and members)
As is true so much of the time in programming, it depends greatly on the situation.
For instance, what is the cohesiveness of the classes in question? Are they tightly coupled? Are they completely orthogonal? Are they related in functionality?
It would not be out of line for a web framework to supply a general purpose widgets.whatever file containing BaseWidget, TextWidget, CharWidget, etc.
A user of the framework would not be out of line in defining a more_widgets file to contain the additional widgets they derive from the framework widgets for their specific domain space.
When the classes are orthogonal, and have nothing to do with each other, the grouping into a single file would indeed be artificial. Assume an application to manage a robotic factory that builds cars. A file called parts containing CarParts and RobotParts would be senseless... there is not likely to be much of a relation between the ordering of spare parts for maintenance and the parts that the factory manufactures. Such a joining would add no information or knowledge about the system you are designing.
Perhaps the best rule of thumb is don't constrain your choices by a rule of thumb. Rules of thumb are created for a first cut analysis, or to constrain the choices of those who are not capable of making good choices. I think most programmers would like to believe they are capable of making good decisions.
The Smalltalk answer is: you should not have files (for programming). They make versioning and navigation painful.
One class per file is simpler to maintain and much more clear for anyone else looking at your code. It is also mandatory, or very restricted in some languages.
In Java for instance, you cannot create multiple top level classes per file, they have to be in separate files where the classname and filename are the same.
(C#) Another exception (to one file per class) I'm thinking of is having List in the same file as MyClass. Where I envisage using this is in reporting. Having an extra file just for the List seems a bit excessive.

Resources