all&all: 2010

Wednesday, December 1, 2010

.NET-- Explicit Cast or Use "as"?

In .NET there are 2 different ways to cast an object from one to another-- by explicitly casting like (OtherObject)startObject (I call this the explicit cast) or by using the as keyword like startObject as OtherObject. What are the advantages/drawbacks of using one or the other?

Why You Should Use "as"

Do you know what goes on behind an explicit cast? If the type to be converted does not match the wanted type at runtime, it tries to use any user defined conversions to create a completely new type. Also, if the explicit cast doesn't work it throws an error.

The "as" keyword doesn't have these drawbacks. If the cast fails, the newly created object is null; there's no error. And again, no conversion/new object creation occurs. Win-win.

"As" does have a downside though--it cannot be used on value types. Doesn't really come as a surprise though, remembering that it will not do new object creation...

Thursday, May 20, 2010

How to Add a Facebook Like Button to Your Blogger Posts

A friend of mine that owns a Pittsburgh-based custom t-shirts and screen printing business asked me the other day if I could research adding Facebook like buttons to his posts. Here is the result of said research:

Step One: Sign Up for Facebook Developer Access

I'm hoping you already have a facebook login. Using said login, head over here and fill out the form supplying your page name and url.

After that is done it will give you the application ID-- you're going to need this for the next step.

Editing Your Blogger Template

Ok so sign in to your blogger admin and navigate to layout > edit html. Make sure you check the "Expand Widget Templates" checkbox.

Now comes the tricky/nerdy inserting code part. Inside of that box where the code is, you'll see at the top of it a tag that starts with <html. We need to add xmlns:fb='http://www.facebook.com/2008/fbml' inside of that tag (which means we need to put it inbetween the < and > symbols for those of you not too html savvy). . So your final <html> tag should look similar to this:

<html expr:dir='data:blog.languageDirection' xmlns='http://www.w3.org/1999/xhtml' xmlns:b='http://www.google.com/2005/gml/b' xmlns:data='http://www.google.com/2005/gml/data' xmlns:expr='http://www.google.com/2005/gml/expr' xmlns:fb='http://www.facebook.com/2008/fbml'>

The important part is that you got that xmlns:fb facebook part. The next step is to scroll down and look for the <body> tag. In the picture below you'll see the body tag, and then the code you need to insert below it.

You're going to substitute my app id for your app id--which should be a string of numbers--no symbols, letters or anything else. Your final code should look like so:

<div id='fb-root'/>
<script>
  window.fbAsyncInit = function() {
    FB.init({appId: YOUR-APP-ID, status: true, cookie: true,
             xfbml: true});
  };
  (function() {
    var e = document.createElement('script'); e.async = true;
    e.src = document.location.protocol +
      '//connect.facebook.net/en_US/all.js';
    document.getElementById('fb-root').appendChild(e);
  }());
</script>

Remember, substitute YOUR-APP-ID for your appication ID you got when you signed up above. We're almost there, one more change! Now we need to find
<div class="post-footer">. We're going to add the tag that begins with fb:like in there.

So your final code there should be similar to:

<div class="post-footer">
<fb:like action='like' colorscheme='light' expr:href='data:post.url' layout='standard' show-faces='false' width='450'/>
</div>

That's it! All done! Be sure to like this post below ;-)

Monday, May 3, 2010

Making Cross Domain jQuery AJAX Calls

Today's web browsers do not allow web pages to make cross-domain ajax calls. By this i mean that if you are at www.ajax.com and try to make an Ajax call ( an HTTP request using the XmlHttpRequest object) to www.other-domain.com the browser would not allow this to happen. Why? For security purposes that i cannot currently name.

However, at some point you get to a project where you're interfacing this third-party site that needs to talk to your main site, or some other similar situation where the only way you're going to get the data you need from point a to point b is with some javascript magic. Here is how to accomplish it:

How to get/post data using jQuery/javascript (JSONP)

The short answer: its not ajax at all, its JSONP. Yes, JSONP is not Ajax. I just learned this today. Like i said earlier, browsers do not allow XHR/Ajax cross-browser requests. JSONP avoids this by making a request for a script file-- no Ajax at all. Let me explain:

Instead of accessing the XHR object, the browser first creates a new script tag to inject into the HTML DOM
The script tag's URL is set to the URL you're looking to get/post(using HTTP GET) data to
The script tag is injected into the page, causing...
The request is sent to the server, even if its cross-domain
The server returns the data in the form of a javascript function call
The browser receives the data and executes the function call

jQuery code

//use a get to post a querystring value via HTTP GET to an asp.net webhandler
$.getJSON("http://www.example.com/get-post.ashx?var1=var1value&callback=?",function(data){
     //really no need to do anything here, we're just posting data
     //but this will output success
     alert(data.return);
});

asp.net generic handler/webhandler code

using System;
using System.Collections.Generic;
using System.Linq;
using System.Web;
using System.Web.Services;

namespace jsonp_test
{
    /// 
    /// Summary description for $codebehindclassname$
    /// 
    [WebService(Namespace = "http://tempuri.org/")]
    [WebServiceBinding(ConformsTo = WsiProfiles.BasicProfile1_1)]
    public class get_post : IHttpHandler
    {

        public void ProcessRequest(HttpContext context)
        {
            string callback = "";
            try
            {
                if (!string.IsNullOrEmpty(context.Request["callback"]))
                {
                    if (!string.IsNullOrEmpty(context.Request["var1"]))
                        SaveData(context.Request["var1"]);
                    callback = context.Request["callback"];

                    context.Response.Write(callback + "({ \"return\": \"Success\" })");
                }
            }
            catch (Exception exc)
            {
                //hopefully this error doesnt contain any quotes... you know?
                context.Response.Write(callback + "({ \"return\": \"" + exc.Message + "\" })");
            }
        }

        private void SaveData(string value)
        {
            //do something with the var1 posted to us
        }

        public bool IsReusable
        {
            get
            {
                return false;
            }
        }
    }
}

So what we effectively have here is the setup for asp.net cross domain ajax calls using jQuery

How to do cross-domain calls to others' servers: YQL

What if you need to do a get but don't have the ability to make the page its posting to return a JSONP response? Using YQL is a great way to achieve this.

YQL is a way to query the internet like it is a database. For example, one could easily run

select * from html where url="http://www.microsoft.com"

and recieve a JSONP return containing all of the site's HTML. You could also, do the following

select * from html where url="http://www.mircorosft.com?var1=var1value"

and HTTP GET values to the server. YQL does not allow this to happen on any area of any site that is blocked by the robots.txt file. Here's full example code:

//feel free to add querystring vars to this
var myurl="http://www.example.com/get-post.ashx?var1=var1value&callback=?";
//make the call to YQL 
$.getJSON("http://query.yahooapis.com/v1/public/yql?"+
                "q=select%20*%20from%20html%20where%20url%3D%22"+
                encodeURIComponent(myurl)+
                "%22&format=xml'&callback=?",
        function(data){
          if(data.results[0]){
            //this data.results[0] is the return object you work with, 
            //if you actually want to do something with the returned json
            alert(data.results[0]);
          } else {
            var errormsg = 'Error: could not load the page.';
            //output to firebug's console
            //use alert() for other browsers/setups
            conole.log(errormsg);
          }
        }
      );

Wednesday, March 31, 2010

Random Order LINQ Query

I get a lot of requests for random ordering at my job... this helps a lot.


(from s in myset orderby Guid.NewGuid() select s).Take(5)

Updated 10/4/2011

Wednesday, February 10, 2010

C# Design Patterns: Factory

What is the factory design pattern? Well to begin with it is something that is referenced frequently in software architecture. Functionally, a factory builds an object. More specifically, factories render instances of an abstract class type.

C# Example

How often have you got caught up in creating objects based upon business logic? Something like this:

 public interface ICar
{
 decimal Price {get;set;}
 int HorsePower{get;set;}
}

//found somewhere randomly in your code...
 ICar result;
 //this code is heavily biased
 if(user.IsRich)
  return new AstonMartin();
 else if(user.IsSteelWorker)
  return new Ford();
 else if(user.IsPractical)
  return new Honda();
 return MomsCar();

I've found code like this in projects many times and have done it myself. What's really wrong with this?

If you need this code again, and copy/paste, you'll be violating DRY (do not repeat yourself)
Not reusable
Not centrally based, so a change in one place will need copied to another
Whatever class this code is in is at this point violating the Single Responsibility Principle

The Factory Design Pattern Fix

Take a look over why this is architectually better:

public enum CarTypes { NotSet, Honda, AstonMartin, Ford };
public sealed class CarFactory
{
    public ICar Create(CarTypes cType)
    {
        if (cType == CarTypes.Honda)
            return new Honda();
        else if (cType == CarTypes.AstonMartin)
            return new AstonMartin();
        else if (cType == CarTypes.Ford)
            return new Ford();
        throw new ArgumentOutOfRangeException("cType");
    }
}

What have we changed in this example?

All of our code is in one area--easy to update and reuse
The factory is sealed, so we need not worry about any other classes changing the factory behavior

Simple stuff!

Monday, February 1, 2010

Dependency Injection in .NET

I'm really on an architecture kick at this point. Honestly, it's probably about time. I find few senior-style job postings that do not ask for some knowledge of Spring.NET or Castle Windsor. Now that i've done a little reading up on the subject, I see that this is for good reason.

Introduction: What is Dependency Injection?

This stuff is tricky--I'm not going to lie. First of all, I would recommend reading up on the basics. Until you start to really see the power behind defining interfaces and working with abstractions dependency injection will be unnecessary for you.

Dependency Injection strives to decouple classes from dependencies. For instance:

public interface IComplaintHearer
{
void RegisterComplaint(string message);
}
public class Manager : IComplaintHearer
{
public Manager() { }
public void RegisterComplaint(string message)
{
//do something with message
}
}
public class Employee
{
//completely dependent upon this exact class
private Manager _itsManager;
public Employee() { }
public void Complain(string complaint)
{
_itsManager.RegisterComplaint(complaint);
}
}

What's the issue? What if we want the employee to report to someone other than its immediate boss? Or, what if we want the employee to complain to a co-worker? We would have to completely change the class. Right now this code completely violates the open closed principle(among others) we discussed when reviewing the S.O.L.I.D. principles. So lets use the dependency inversion principle and make this class dependent upon an abstraction:

public interface IComplaintHearer
{
void RegisterComplaint(string message);
}
public class Manager : IComplaintHearer
{
public Manager() { }
public void RegisterComplaint(string message)
{
//do something with message
}
}
public class Employee:IComplaintHearer
{
//completely dependent upon this exact class
private IComplaintHearer _complaintHearer;
public Employee(IComplaintHearer hearer)
{
_complaintHearer=hearer;
}
public void RegisterComplaint(string message)
{
//do something with the message
}
public void Complain(string complaint)
{
_complaintHearer.RegisterComplaint(complaint);
}
}

And we would need to run this code like so:

//we could pass a manager, or another employee if we wanted to
Employee myEmp = new Employee(new Manager());

Injecting the Class, Instead of Providing It

Imagine a scenario where we had tons of classes that implemented the IComplaintHearer interface. We'd have to recompile the code everytime we want to change who the employee complains to. This is where a dependency injection steps in and allows you to:

Specify a class's dependency at run-time
Dynamically use classes in another assembly
Make changes without recompilation

Let's take a look at an example:

public interface IComplaintHearer
{
void RegisterComplaint(string message);
}
public class Manager : IComplaintHearer
{
public Manager() { }
public void RegisterComplaint(string message)
{
//do something with message
}
}
public class Employee:IComplaintHearer
{
//completely dependent upon this exact class
private IComplaintHearer _complaintHearer;

//use the IComplaintHearer subclass that the Dependency Injection Framework (StructureMap in this case) tells us to
// this depends upon the xml that defines what to use (below)
public Employee(): this ( ObjectFactory.GetInstance()){}
public Employee(IComplaintHearer hearer)
{
_complaintHearer=hearer;
}
public void RegisterComplaint(string message)
{
//do something with the message
}
public void Complain(string complaint)
{
_complaintHearer.RegisterComplaint(complaint);
}
}

What just happened? By using the StructureMap Framework for .NET, we just specified the IComplaintHearer that our employee class will default to--the manager. In the XML above, we mapped the expected type/assembly to the default type/assembly. Further, we could set up defaults for all of our classes that we could change later without ever having to change any code. In some ways, I feel like this is moving a problem from one environment to another, but in other ways i think it is a great architectural tool.

What do you think?

Friday, January 29, 2010

The S.O.L.I.D. Object Oriented Programming(OOP) Principles

Introduction

What does it take to be an Object Oriented Programmer? There was a time where I believed all that meant was that you worked with a language such as C#, C++, or Java. However, the more I get acquainted with newer technologies, the more I realize that there is a set of fundamentals core to the title. And really, these fundamentals are about architecting the best, most update-able, scalable systems. Just yesterday while diving into DataObjects.NET I was greeted by Domain Driven Design(DDD)-- a popular architectural abstraction. It motivated me to think about the basics, which is the purpose of this article.

The S.O.L.I.D. Principles of Class Design

The S.O.L.I.D. principles seem to be the least common denominator of creating great classes; even before Design Patterns. I recommend taking some time to really think about each of them and how you can apply them. Lets dive in, one by one.

The Single Responsibility Principle

There should never be more than one reason for a class to change. Basically, this means that your classes should exist for one purpose only. For example, lets say you are creating a class to represent a SalesOrder. You would not want that class to save to the database, as well as export an XML-based receipt. Why? Well if later on down the road you want to change database type(or if you want to change your XML schema), you're allowing one responsibility's changes to possibly alter another. Responsibility is the heart of this principle, so to rephrase there should never be more than one responsibility per class.

The Open Closed Principle

Software entities(classes,modules,functions,etc.) should be open for extension, but closed for modification. At first this seems to be contradictory: how can you make an object behave differently without modifying it? The answer: by using abstractions, or by placing behavior(responsibility) in derivative classes. In other words, by creating base classes with override-able functions we are able to create new classes that do the same thing differently without changing the base functionality. Further, if properties of the abstracted class need compared or organized together, another abstraction should handle this. This is the basis of the "keep all object variables private" argument.

The Liskov Substitution Principle

Functions that use pointers or references to base classes must be able to use objects of derived classes without knowing it.In other words, if you are calling a method defined at a base class upon an abstracted class, the function must be implemented properly on the subtype class. Or, "when using an object through its base class interface,[ the]derived object must not expect such users to obey preconditions that are stronger than those required by the base class." The ever-popular illustration of this is the square-rectangle example. Turns out a square is not a rectangle, at least behavior-wise.

The Dependency Inversion Principle

Depend on abstractions, not on concretions or High level modules should not depend upon low level modules. Both should depend upon abstractions. Abstractions should not depend upon details. Details should depend upon abstractions. (I like the first explanation the best.) This is very closely related to the open closed principle we discussed earlier. By passing dependencies (such as connectors to devices, storage) to classes as abstractions, you remove the need to program dependency specific. Here's an example: a Employee class that needs to be able to be persisted to xml and a database. If we placed ToXML() and ToDB() functions in the class, we'd be violating the single responsibility principle. If we created a function that took a value that represented whether to print to XML or to DB, we'd be hard-coding a set of devices and thus be violating the open closed principle. The best way to do this would be to: 1) Create an abstract class (named DataWriter, perhaps) that can be inherited from for XML (XMLDataWriter) or DB (DbDataWriter) Saving, and then 2) Create a class (named EmployeeWriter) that would expose an Output(DataWriter saveMethod) that accepts a dependency as an argument. See how the Output method is dependent upon the abstractions just as the output types are? The dependencies have been inverted. Now we can create new types of ways for Employee data to be written, perhaps via HTTP/HTTPS by creating abstractions, and without modifying any of our previous code! No rigidity--the desired outcome.

The Interface Segregation Principle

Clients should not be forced to depend upon interfaces that they do not use. My favorite version of this is written as "when a client depends upon a class that contains inter- faces that the client does not use, but that other clients do use, then that client will be affected by the changes that those other clients force upon the class." Kinda sounds like the inheritance specific single responsibility principle.