Zack Owens

Fluent NHibernate with System.ComponentModel.DataAnnotations

zowens — Fri, 23 Oct 2009 20:34:00 GMT

So I’m building a product with NHibernate and possibly another ORM. The main focus of the app is a ASP.NET MVC 2.0 web application. So I, obviously, want the built in validation client-side in JavaScript and server side. By default, the MVC stuff uses the System.ComponentModel.DataAnnotations attributes to validate properties on your domain.

What I REALLY wanted was to use the Data Annotations with Fluent NHibernate for my DB mapping files. This presents a problem, right? WRONG! Because Fluent NHibernate is so extensible, I can easily create a convention to use the Attributes. Then all I have to do is add the convention to my Fluent NHibernate AutoMapper instance and the properties with the attributes are just taken care of. Isn’t that neat!!

Required is a common notion with domains. Something HAS to be defined for an entity to be considered a reputable entry. Here’s the convention to use the Required attribute in Fluent NHibernate.

Code Snippet

using System.ComponentModel.DataAnnotations;
using FluentNHibernate.Conventions;
using FluentNHibernate.Conventions.Instances;

namespace Data.NHibernate.Conventions
{
    public class RequiredConvention : AttributePropertyConvention<RequiredAttribute>
    {
        protected override void Apply(RequiredAttribute attribute, IPropertyInstance instance)
        {
            instance.Not.Nullable();
        }
    }
}

As you can see… there’s really only 1 line of code!

I just showed you one out of the handful of attributes in the DataAnnotations library. To use the other attributes, you just basically follow the same basic outline of the example above. Not bad, eh :)

Maybe Monad… my C# version

zowens — Fri, 04 Sep 2009 05:09:00 GMT

Functional programming isn’t a new concept. There is Scheme, Haskell and a bunch of other really cool languages. But new on the block is F#. F# is Microsoft’s stab at creating a functional programming in the .NET world as a first-class language.

Functional programming paradigms are increasingly important to all .NET developers. This post will detail the functional elements (in my mind) of the Maybe Monad, something I’m sure everyone has written in some way or another (ok maybe not everyone, but there are certainly a lot of implementations out there in C#). But let’s discuss the application, functionally, of the maybe monad.

The Basics

There are three areas of the maybe monad that are really important. The first is the concept of “Nothing” and “Just”. Nothing, in this context, would be either a failed operation (say division by 0, for example) or something that just returns null. The “Just” is the opposite of “Nothing”, it’s something. It’s a successful operation that returns a value. Although this concept seems simple, it’s actually quite powerful. We’ll return to this in a bit.

The second of the important concepts with the maybe monad is the “bind” and “return” verbs. “Bind” is the assigning of a value to a Maybe. “Return” the extraction of the value from the maybe. See the Maybe Monad is a container of a value (or lack thereof). The verbs attached to the container relay a façade of the actual value. While my implementation doesn’t deal with bind/return verbs directly, it is used indirectly with the composition and decomposition of the value container.

The third is something we’re ALL familiar with, a type constraint. The formal declaration of the maybe monad declares that there must be a type parameter of type M. This letter is arbitrary, of course, but the concept is important. Without the constraint, the possible values of the true value bound to the maybe container would be limitless. One more reason where generics in C# are invaluable!

Let’s get coding!

Start with Nothing

Let’s start out with some initial xUnit.net tests.

[Fact]
public void Maybe_NothingContructor_SetsNothingFlag()
{
    var maybe = new Maybe<int>();
    bool isNothing = maybe.Nothing;
    Assert.True(isNothing);
}

[Fact]
void Maybe_JustContructor_DoesntSetNothingFlag()
{
    var maybe = new Maybe<int>(5);
    bool isNothing = maybe.Nothing;
    Assert.False(isNothing);
}

[Fact]
public void Maybe_JustContructor_SetsNothingFlag_WhenTheValuePassedIsNull()
{
    var maybe = new Maybe<int>();
    bool isNothing = maybe.Nothing;
    Assert.True(isNothing);
}

As you can see there are 2 constructors. The default is the “nothing” constructor. The constructor with a value is the “nothing” if the value is null and the “just” if there is an actual value. This is an immutable container, so I don’t feel so bad setting these values only in the constructor. Here’s what my “Maybe” looks like now.

/// <summary>
/// Value container for storing 
/// a) a value ("Just")
/// b) lack of a value ("Nothing")
/// </summary>
/// <typeparam name="M">Type M of the Maybe Monad</typeparam>
public sealed class Maybe<M>
{

    /// <summary>
    /// Non-value maybe ("Nothing")
    /// </summary>
    public Maybe()
    {
        // assign the nothing flag
        Nothing = true;
    }

    /// <summary>
    /// Maybe with a value ("Just")
    /// </summary>
    /// <param name="value"></param>
    public Maybe(M value)
    {
        // assign the nothing flag
        // when not null
        Nothing = value == null;
    }

    /// <summary>
    /// Determines the state of the 
    /// value in the container
    /// </summary>
    public bool Nothing { get; private set; }
}

You might be wondering why I made this class sealed. My rationale for this is that, since Maybe is immutable, we don’t want to get rid of some of the guarantees we have with the default implementation. Having a central “maybe” is a nice, simple way to define the maybe behavior.

Now for the return

The notion of a return for the maybe monad essentially means extracting the value from the container. I could define a “Return” function or even a “Value” property. But this is an unnecessary step. We don’t really need added complexity. My theory is that casting would be the best option. It is, after all, just a façade. Plus casting is a bit more expressive anyways. In your implementation, you can do this if you wish :)

[Fact]
public void Maybe_MCast_ReturnsAValue_WhenThereIsAValue()
{
    var maybe = new Maybe<string>("");
    string value = maybe;
    Assert.NotNull(value);
}

[Fact]
public void Maybe_MCast_ReturnsSameValue_WhenThereIsAValue()
{
    var initialValue = "";
    var maybe = new Maybe<string>(initialValue);
    string value = maybe;
    Assert.Same(initialValue, value);
}

[Fact]
public void Maybe_MCast_ReturnsNull_WhenThereIsNotAValue()
{
    var maybe = new Maybe<string>();
    string value = maybe;
    Assert.Null(value);
}

And here is the code that makes the tests pass.

public sealed class Maybe<M>
{
    private M _value;
    
    ...
    // previous code hidden for brevity
    
    public static implicit operator M(Maybe<M> maybe)
    {
        if (maybe == null)
            throw new ArgumentNullException("maybe");

        return maybe._value;
    }
}

I added the value field and the implicit cast operator.

Another way to bind

The constructor to define the “Just” and “Nothing” is useful, but what if we wanted save a little code when we are returning a maybe from a method? To do that, we’ll define another implicit operator. This time the operator will take an M and return a Maybe<M>. Here’s our test:

[Fact]
public void Maybe_MaybeCast_ReturnsNothingMaybe_WhenTheValueIsNull()
{
    string val = null;
    Maybe<string> maybe = val;
    Assert.True(maybe.Nothing);
}

[Fact]
public void Maybe_MaybeCast_ReturnsJustMaybe_WhenTheValueIsNotNull()
{
    string val = "";
    Maybe<string> maybe = val;
    Assert.False(maybe.Nothing);
}

And here’s the code.

public sealed class Maybe<M>
{    
    ...
    // previous code hidden for brevity
    
    public static implicit operator Maybe<M>(M value)
    {
        return new Maybe<M>(value);
    }
}

Apply for Chainability

Being able to chain different functions is really a key part of the maybe monad. The verb I will use is “Apply”. In functional languages, this action just happens due to function composition. We have two options here. We can return the Maybe monad as a Func<Maybe<M>, Maybe<M>> or we can define an Apply method to take a Func<Maybe<M>, Maybe<M2>> where M2 is essentially a new value. The Apply method could serve as a converter. If the initial value of M is Nothing, then the converted value should return a Nothing maybe monad. The great thing about the maybe monad is that there is no need to throw exceptions.Although the throwing of exceptions can provide useful information, most simple operations don’t require such complexity. Just test for Nothing on the maybe. Now you don't have to add a try/catch throughout your code.

I am going to use the Apply method. Although returning a Func<Maybe<M>, Maybe<M>> might add expressive syntax, the Apply method will be a little more flexible for most scenarios. Here’s some tests.

[Fact]
public void Maybe_Apply_ReturnsNothingMonad_WhenTheInitialNothingMaybe()
{
    Func<Maybe<string>, Maybe<string>> fun= (m) => new Maybe<string>("");
    Maybe<string> initialMaybe = new Maybe<string>();
    Maybe<string> appliedMaybe = initialMaybe.Apply(fun);
    Assert.True(appliedMaybe.Nothing);
}

[Fact]
public void Maybe_Apply_ReturnsNewMaybeWithNewValue()
{
    Func<Maybe<string>, Maybe<string>> addBchar = (maybe) => 
                        new Maybe<string>(((string)maybe) + "b");

    string initial = "", final = "b";

    Maybe<string> initialMaybe = new Maybe<string>(initial);
    string afterApply = initialMaybe.Apply(addBchar);
    Assert.Equal(final, afterApply);
}

As you can see, my implementation will automatically return a nothing maybe monad when a nothing maybe monad is the initial value. OK here’s the code.

 /// <summary>
/// Applies the specified function to this object
/// if this maybe is not nothing
/// </summary>
/// <typeparam name="M2">Return Type</typeparam>
/// <param name="applyFunction">Conversion function</param>
/// <returns>A Nothing maybe if the initial maybe is nothing
/// or the value returned by <see cref="applyFunction"/> if
/// this maybe has a value.
/// </returns>
public Maybe<M2> Apply<M2>(Func<Maybe<M>, Maybe<M2>> applyFunction)
{
    if (Nothing)
        return new Maybe<M2>();
    return applyFunction(this);
}

Conclusion

The maybe monad is pretty cool for writing structure C# code. Your methods can now return a Maybe<M> to be perfectly clear if your methods returned actual values or null. This also gives you another definition of null. Perhaps your idea of “null” could be returning a 0. We have added flexibility with this new construct.

I’ve just hit the high marks in this post of my implementation of the maybe monad. Here’s an extended version with tests. Enjoy :)

DOWNLOAD

Cache Abstraction

zowens — Mon, 04 Aug 2008 21:08:04 GMT

I'm in the middle of a project and I wrote my own Cache logic that melds nicely to my project's needs. But how do I test it? I can't mock the System.Web.Caching.Cache class. So I looked in the Abstractions that ship with ASP.NET MVC. What do you know, they're not there!

So I created my own base class and wrapper. Here's the code. Enjoy.

using System;
using System.Collections;
using System.Web.Caching;

namespace WebCommon.Abstractions
{
    public abstract class CacheBase : IEnumerable
    {
        public abstract int Count { get; }
        public abstract long EffectivePercentagePhysicalMemoryLimit { get; }
        public abstract long EffectivePrivateBytesLimit { get; }
        public abstract object this[string key] { get; set; }
        public abstract object Add(string key, object value, CacheDependency dependencies, DateTime absoluteExpiration, TimeSpan slidingExpiration, CacheItemPriority priority, CacheItemRemovedCallback onRemoveCallback);
        public abstract object Get(string key);
        public abstract IDictionaryEnumerator GetEnumerator();
        public abstract void Insert(string key, object value);
        public abstract void Insert(string key, object value, CacheDependency dependencies);
        public abstract void Insert(string key, object value, CacheDependency dependencies, DateTime absoluteExpiration, TimeSpan slidingExpiration);
        public abstract void Insert(string key, object value, CacheDependency dependencies, DateTime absoluteExpiration, TimeSpan slidingExpiration, CacheItemPriority priority, CacheItemRemovedCallback onRemoveCallback);
        public abstract object Remove(string key);

        IEnumerator IEnumerable.GetEnumerator()
        {
            return this.GetEnumerator();
        }
    }
}

using System;
using System.Collections;
using System.Web.Caching;

namespace WebCommon.Abstractions
{
    public class CacheWrapper : CacheBase
    {
        private Cache _cache;

        public CacheWrapper(Cache cache)
        {
            _cache = cache;
        }

        public override int Count
        {
            get { return _cache.Count; }
        }

        public override long EffectivePercentagePhysicalMemoryLimit
        {
            get { return _cache.EffectivePercentagePhysicalMemoryLimit; }
        }

        public override long EffectivePrivateBytesLimit
        {
            get { return _cache.EffectivePrivateBytesLimit; }
        }

        public override object this[string key]
        {
            get
            {
                return _cache[key];
            }
            set
            {
                _cache[key] = value;
            }
        }

        public override object Add(string key, object value, CacheDependency dependencies, DateTime absoluteExpiration, TimeSpan slidingExpiration, CacheItemPriority priority, CacheItemRemovedCallback onRemoveCallback)
        {
            return _cache.Add(key, value, dependencies, absoluteExpiration, slidingExpiration, priority, onRemoveCallback);
        }

        public override object Get(string key)
        {
            return _cache.Get(key);
        }

        public override IDictionaryEnumerator GetEnumerator()
        {
            return _cache.GetEnumerator();
        }

        public override void Insert(string key, object value)
        {
            _cache.Insert(key, value);
        }

        public override void Insert(string key, object value, CacheDependency dependencies)
        {
            _cache.Insert(key, value, dependencies);
        }

        public override void Insert(string key, object value, CacheDependency dependencies, DateTime absoluteExpiration, TimeSpan slidingExpiration)
        {
            _cache.Insert(key, value, dependencies, absoluteExpiration, slidingExpiration);
        }

        public override void Insert(string key, object value, CacheDependency dependencies, DateTime absoluteExpiration, TimeSpan slidingExpiration, CacheItemPriority priority, CacheItemRemovedCallback onRemoveCallback)
        {
            _cache.Insert(key, value, dependencies, absoluteExpiration, slidingExpiration, priority, onRemoveCallback);
        }

        public override object Remove(string key)
        {
            return _cache.Remove(key);
        }
    }
}

ClubStarterKit - where do we go?

zowens — Wed, 30 Jul 2008 01:40:28 GMT

Wow, what a year. Not exactly what I'd expected. Before I dive into where we are now, let me preface this post by saying I have NOT forgotten ClubStarterKit! Although I was not the original author, I feel like it's my baby. I can't just forget about it. Not going to happen.

You might be quick to point out that I have done virtually nothing since the v3 beta 1 release. This, my friends, was a big mistake. Not v3 beta 1, which I will discuss later in this post, but the fact that I was not clear in what I hope to get out of the project. I do take the blame on the lack of development. I'm not going to sit here and tell you I have neglected to look at the code in a while. This is partially due to my work life (I have to make some money.. know that I have 2 jobs) and my personal endeavors (as I am a high school student with more to do than dedicate all my time to ClubStarterKit). No one is perfect. But I do sincerely apologize to the ClubStarterKit community for my lack of communication.

My hopes from the project

When I started the project, I knew that there was a void. ClubStarterKit v1 was an obvious success because of the idea of sample-driven learning. It did a lot of things that people needed. Now, I really wanted to bring CSK to the level of open source. With this, I opened up the door for contributors. This is an area that I found fairly problematic. We did get some amazing contributions, such as the CMS and the RSVP system. Without these core contributions, I would have spent more time building the features rather than tracking down bugs or expanding other CSK features. Trust me, building out a feature for CSK is not easy.

Essentially I had hoped for the ClubStarterKit to be a community effort where there would be multiple core developers. In my idea of "core developer", I seem to be the only one. To me, CSK seems more like a sample that goes through active development.

It is my hope to not be a nag about this, but come on CSK community. You guys do some amazing stuff. Then you demand something out of CSK without contributing? I just don't see how this whole thing can work out like this. I'm not a free contractor. I don't even ask for donations!

I have been asked by one member of the community to forget my role in CSK and pass it along to another developer. Guess what, no one stepped up. Interesting. Guess it looks like I'm still in, win or lose.

What's the deal with the inactivity?

Like I said, there are many factors to this. Mostly, I am starting to regret hard-coding SubSonic into CSK. I now know, after completely hard-coding this stuff in, that I should have allowed external DALs that people wanted. Not everybody like SubSonic and it's query tool. Some people like NHibernate, LINQtoSQL, XML, whatever. I seriously screwed up. My fault, guys. I've learned, let's move on.

A big reason for inactivity was time. I just couldn't devote a lot of time on CSK like I used to. Luckily, this is about to change for me. Just know that I'm going to have more time to work on CSK. Period.

What's next?

I hate to say it, but v3 beta 1 was kind of bad. I really wanted something fresh. That just wasn't what we got with v3 beta 1. The ideas were there, just the delivery was poor. So this is what I plan on doing. What I want to do, and I would LOVE to hear feedback from you on this, is to completely scrap the current codebase and move over to the ASP.NET MVC framework. I am seriously in love with ASP.NET MVC. Honestly, it will move development, deployment, and testability along. I am in the preliminary planning stages. Also, for development, I would like to use C# and convert the C# code over to VB. The fact that the original kit was released in both languages made it all the more appealing. With a refreshed codebase, we can make this process much easier on the person converting (probably me :) ).

Also, we will have an extensive service layer that can be extended to use a custom DAL or any form of data storage.

Lastly, the features listed in my last post will go in.

So, what do you think? I would LOVE to hear your feedback AS LONG AS IT IS PRODUCTIVE AND WILL BENEFIT THE PROJECT.

THANKS FOR YOUR CONTINUED UNDERSTADING CLUBSTARTERKIT COMMUNITY!

One-Line C# #1 : Remove Whitespace from a string

zowens — Fri, 27 Jun 2008 18:37:43 GMT

Thought I'd post a quick extension method that I made today. Hope it helps someone.

public static string RemoveWhitespace(this string str){    try    {        return new Regex(@"\s*").Replace(str, string.Empty);    }    catch (Exception)    {        return str;    }}

Misuse of Delegates???

zowens — Mon, 26 May 2008 23:39:00 GMT

In the evolution of the ideas presented in Java, C# has provided the idea of method delegation and delegates, which I'm sure you are well aware of. In C# 3.0 with the introduction of the LINQ, Lambdas were born. With Lambdas, we are able to provide an easy and elegant way to declare delegates. With great power, however, comes great responsibility. Is there a way to overuse Lambdas and Delegates? Here is a sample of code I've written:

/// <summary>/// If GZIP or DEFLATE compression is accepted/// by the user's browser, we will go ahead/// and compress the file upon the controller action/// </summary>/// <param name="controller">Controller to invoke Compression</param>public static void Compress(this Controller controller){    string gzip = "gzip";    string deflate = "deflate";     //DELEGATES    Func<string, bool> isAccepted = (encoding) => controller.Request.Headers["Accept-encoding"] != null && controller.Request.Headers["Accept-encoding"].Contains(encoding);    Action<string> setEncoding = (encoding) => controller.Response.AppendHeader("Content-encoding", encoding);     if (isAccepted.Invoke(gzip))    {        controller.Response.Filter = new GZipStream(controller.Response.Filter, CompressionMode.Compress);        setEncoding.Invoke(gzip);    }    else if (isAccepted.Invoke(deflate))    {        controller.Response.Filter = new DeflateStream(controller.Response.Filter, CompressionMode.Compress);        setEncoding.Invoke(deflate);    }}

As you can see from the code, I use the .NET built-in delegates of Action and Func to do the setting of the encoding and the determination of the encoding, respectively.

So here is my question: Is this a misuse of delegates? Should I be putting the code into a separate private method? What are your thoughts?

ASP.NET MVC CSS Sprite

zowens — Sun, 27 Apr 2008 04:24:00 GMT

A month ago, I wrote about my technique to output a CSS sprite for ASP.NET. Granted, I made a big mistake. I put WAY too much bloat into the post. Not a K.I.S.S. moment.

I have reworked my CSS sprite technique to be unique to MVC. I've used the most recent build of the MVC framework, which can be downloaded on Codeplex, to take advantage of ActionResult. What the ActionResult does is makes your controllers more testable and makes the execution of the controller more specific. Instead of a void, controllers now return an ActionResult (see Scott Guthrie's post on the interim release).

I have a download, see the bottom of the post, that includes a sample application and a class library that you can distribute with any MVC application. The sample pulls 2 images off the Internet, plugs them into a sprite through the Sprite controller.

What you get when you execute the Sprite controller is a sprite that looks like so:

What you do in your CSS file is reference the sprite URL and specify the background-position, width, and height.

Then with the following HTML:

you get the images separate from the sprite.

What you get is an easy-to-use API for sprite generation and faster-loading website (less HTTP requests per page) using sprites instead of multiple images. If you want all of the CSS editing to happen at runtime, see my original post of CSS Sprites for that code.

DOWNLOAD SAMPLE APPLICATION AND CODE

Source Code conversion as part of a Build Process

zowens — Sat, 12 Apr 2008 04:10:20 GMT

If you're like me, you want a particular piece of code in your particular language. For example, a lot of people want the Club Starter Kit 2.0, 3.0 in C# since it has been written in VB. Me being the lazy open source developer that doesn't get paid :), I keep delaying the CSK conversion to C# because of my personal stuff. Then I was thinking, what if it were a part of the build process? Then I wouldn't have to go through the arbitrary process of converting a whole application to another language for a release that I will most likely be changing.

Before I go ahead and implement this myself, I was wondering if anyone out in the community has done this sort of thing before? I'm sure it wouldn't be that hard, what with the CruiseControl.NET API being what it is. I have already scoped out some ideas of how to implement. But I was wondering if anyone has ever built this plugin.

PLEASE leave a comment on the blog or email me (zowens2009 at gmail dot com)

THANKS COMMUNITY! I will make you proud if this isn't out there ;)

High Interface diet

zowens — Tue, 11 Mar 2008 01:54:25 GMT

As long as I'm on the coding conventions track (as many of users have commented on my last post), I'll talk about a small coding convention I am trying to do.

I feel more compelled to push an interface into a property, method, or any declaration rather than an object type. Why push around List<string> when you can push around IList<string>? It seems like a much better design decision. It also seems that there is more flexibility as far as future extensions are concerned.

Come on, let's hear those thoughts :)

Coding Pet Peeves

zowens — Sun, 09 Mar 2008 19:32:40 GMT

Any experienced coder has pet peeves when it comes to reading other people's code or writing code. It might be that you don't like regions or that every method should have comments. But here are my two biggest pet peeves when it comes to C#.

Var abuse

WAY too many people are using var instead of a type. This might save your hand, but it doesn't do much for readability of your code. Don't get me wrong, it is VERY usefully at times. But only about 10% of var use is actually productive. The rest is just abuse!

Angle Bracket Abuse

Yes, we all know that C# has angle brackets, get over it :). If you have one statement inside of an if, while, do, for, foreach, or using, you DON'T NEED ANGLE BRACKETS! Take the following code sample I found in a starter kit for example:

lock (SyncLock){    if (_cancelRequested == false)    {        _cancelRequested = true;        if (_uploads != null)        {            foreach (FileUpload upload in _uploads)            {                upload.UploadAsyncCancel();            }        }    }                }

That same thing can be accomplished through this:

lock (SyncLock){    if (_cancelRequested == false)    {        _cancelRequested = true;        if (_uploads != null)            foreach (FileUpload upload in _uploads)                upload.UploadAsyncCancel();    }                }

Isn't the second better to read? Plus this saves lines. For those of you who get paid by line, go ahead and put those in there, fine, but for those of you who don't, make your code more readable!

CSS Sprite for ASP.NET

zowens — Wed, 05 Mar 2008 12:32:00 GMT

CSS sprites are becoming popular as a way to increase application performance by eliminating HTTP requests by the client to the web server. It also serves as a path for better cache management. I will need to go through a bit of background before we start (sorry... we'll get to the generation in a second!).

What is a CSS Sprite?

A CSS sprite is a fancy name for an image that is composed other images. In your CSS you have this one base image, but you pick the image inside of the image. It does what you expect, to just show that one single image. The following is an example of a CSS sprite:

In your CSS file, you would pick this base image as your background-image and then to pick a particular image, you could need to specify the height, width, and also the position of where the image starts... background-position.

Long story short, it cuts down on the number of images that need to be passed from your server to the client that is visiting the site. This adds up over time, especially is you have A LOT of images.

A lot of people go to a sprite generator online. Those are nice, but what happens when you need to change an image? You have to go back and re-upload all your images and recalculate your background-positions. That's not a good thing if you are following DRY (don't repeat yourself). Plus it's hard work NO ONE should have to do.

How I handle output caching

What I do is this: all my CSS files, JavaScript files, etc have a common ID which is basically what I call the "Distro ID" which is generated at the beginning of my ASP.NET applications so every time you do a publish, the number will change and your files will also change to their updated content. The heavy lifting is all done in the Global.asax file.

   1: //C#   2: public class Global : System.Web.HttpApplication {   3:     public static string DistributionNumber { get; set; }   4:     5:     protected void Application_Start(object sender, EventArgs e){   6:         SetDistro();   7:         ..   8:     }   9:     private void SetDistro() {  10:         DistributionNumber = DateTime.Now.Year.ToString() + DateTime.Now.Month.ToString() + DateTime.Now.Day.ToString() + DateTime.Now.Hour.ToString() + DateTime.Now.Minute.ToString() + DateTime.Now.Second.ToString();  11:     }  12: }

   1: 'VB   2: Inherits System.Web.HttpApplication   3:     4: Private m_distronum as String = String.Empty   5: Public Shared Property DistributionNumber() As String   6:     Get   7:         Return m_distronum   8:     End Get   9:     Set(ByVal value as String)  10:         m_distronum = value  11:     End Set  12: End Property  13:    14: Sub Application_Start(ByVal sender As Object, ByVal e As EventArgs)  15:     SetDistro()  16:     ..  17: End Sub  18: Private Sub SetDistro()  19:     DistributionNumber = DateTime.Now.Year.ToString() + DateTime.Now.Month.ToString() + DateTime.Now.Day.ToString() + DateTime.Now.Hour.ToString() + DateTime.Now.Minute.ToString() + DateTime.Now.Second.ToString()  20: End Sub

Now since I am mostly using ASP.NET MVC, I have a controller for the Distribution. So my URLs look like this: [BASE_URL]/20080303010101/CSS. I ago ahead and create separate Routes for this, but the default would also work. The ID that is passed to the CSS, JavaScript, Sprite is the distro number. To do the output caching, you need a separate ID so that the browser can differentiate one distribution's CSS, JS, etc from another. This way, we can client side cache with no worries.

If you were using Web Forms, I would suggest using separate Generic handlers (one for CSS, JS, and the Sprite). They would accomplish the same thing. Just remember that you need to take in an ID as a Query String for the Client-side cache to be safe for your users.

If the ID is not passed to the controller OR the generic handler (whichever you are using), don't client-side cache. But this is something you will want to offer as it, too, saves you in the performance department.

Back to the Sprite generator... get me some images!

You have 2 options when it comes to getting images. You have the option of enumerating through a folder(s) or putting the URLs in by hand. I suggest enumeration if you have the same file type. If you put in the file names by hand, that can get tricky... which I'm sure you can tell why. In my Sprite generator, I process the CSS and replace the images with the sprite reference... so we don't need the URL of the original images anyway's. All that we need to reference the images is the name (lowercase).

As I showed you in my last post about CSS Minification (which we will refer back to), I have a file enumerator method:

   1: //C#   2: public static IList<System.IO.FileInfo> GetFiles(string serverPath, string extention)   3: {   4:   if (!serverPath.StartsWith("~/"))   5:   {      6:     if (serverPath.StartsWith("/"))      7:       serverPath = "~" + serverPath;   8:     else   9:       serverPath = "~/" + serverPath;  10:   }  11:   string path = HttpContext.Current.Server.MapPath(serverPath);  12:     13:   if (!path.EndsWith("/"))  14:     path = path + "/";  15:     16:   if (!Directory.Exists(path))  17:     throw new System.IO.DirectoryNotFoundException();  18:     19:   IList<FileInfo> files = new List<FileInfo>();  20:     21:   string[] fileNames = Directory.GetFiles(path, "*." + extention, System.IO.SearchOption.AllDirectories);  22:     23:   foreach (string name in fileNames)  24:     files.Add(new FileInfo(name));  25:     26:   return files;  27: }

   1: 'VB   2: Public Shared Function GetFiles(ByVal serverPath As String, ByVal extention As String) As IList(Of System.IO.FileInfo)   3:   If Not serverPath.StartsWith("~/") Then   4:     If serverPath.StartsWith("/") Then   5:       serverPath = "~" + serverPath   6:     Else   7:       serverPath = "~/" + serverPath   8:     End If   9:   End If  10:   Dim path As String = HttpContext.Current.Server.MapPath(serverPath)  11:    12:   If Not path.EndsWith("/") Then  13:     path = path + "/"  14:   End If  15:    16:   If Not Directory.Exists(path) Then  17:     Throw New System.IO.DirectoryNotFoundException()  18:   End If  19:    20:   Dim files As IList(Of FileInfo) = New List(Of FileInfo)()  21:    22:   Dim fileNames As String() = Directory.GetFiles(path, "*." + extention, System.IO.SearchOption.AllDirectories)  23:    24:   For Each name As String In fileNames  25:     files.Add(New FileInfo(name))  26:   Next  27:    28:   Return files  29: End Function

If you were going the hand-coded image URLs, you would just need to have a method that return a List of FileInfo, just like the method from above.

Get the SPRITE already!

We need to specify a few data models so that they can contain the individual images and the sprites. They aren't anything special... just basically containers for the bitmap object (SpriteImage) and a Collection of the images for the Sprite generation (SpriteImageCollection). The collection is actually the generator of the sprite. The GetSprite() method is the method that will do the image concatenating for you.

   1: //C#   2: public class SiteImage   3: {   4:   public SiteImage(FileInfo fileinfo)   5:   {   6:     Stream stream = new FileStream(fileinfo.FullName, FileMode.Open, FileAccess.Read);   7:     Img = (Bitmap)Bitmap.FromStream(stream);   8:     File = fileinfo;   9:   }  10:    11:   public int Width {  12:     get { return Img.Width; }  13:   }  14:   public int Height {  15:     get { return Img.Height; }  16:   }  17:   public int YValue { get; set; }  18:   public Bitmap Img  { get; set; }  19:   public FileInfo File { get; set; }  20: }  21:    22: public class SiteImageCollection : List<SiteImage>  23: {  24:     public int MaxWidth   25:     {  26:         get   27:         {  28:             int largest = 0;  29:    30:             foreach (SiteImage s in this)  31:                 if (s.Width > largest)  32:                     largest = s.Width;  33:    34:             return largest;  35:         }   36:     }  37:    38:     public int TotalHeight   39:     {  40:         get   41:         {  42:             int ttl = 0;  43:             foreach (SiteImage s in this)  44:                 ttl += s.Height;  45:             return ttl;  46:         }  47:     }  48:    49:     public Bitmap GetSprite()   50:     {  51:         Bitmap spriteImg = new Bitmap(MaxWidth, TotalHeight);  52:    53:         Graphics sprite = Graphics.FromImage(spriteImg);  54:    55:         int curY = 0;  56:         foreach (SiteImage si in this)  57:         {  58:             sprite.DrawImage(si.Img, 0, curY);  59:             si.YValue = curY;  60:             curY += si.Height;  61:         }  62:    63:         return spriteImg;  64:     }  65: }

   1: 'VB   2: Public Class SiteImage   3:     Public Sub New(ByVal fileinfo As FileInfo)   4:         Dim stream As Stream = New FileStream(fileinfo.FullName, FileMode.Open, FileAccess.Read)   5:         Img = DirectCast(Bitmap.FromStream(stream), Bitmap)   6:         File = fileinfo   7:     End Sub   8:     9:     Public ReadOnly Property Width() As Integer  10:         Get  11:           Return Img.Width  12:         End Get  13:     End Property  14:    15:     Public ReadOnly Property Height() As Integer  16:     Get  17:       Return Img.Height  18:     End Get  19:     End Property  20:     21:     Private m_yval As Integer  22:     Public Property YValue() As Integer  23:     Get  24:         Return m_yval  25:     End Get  26:     Set(ByVal value As Integer)  27:         m_yval = value  28:     End Set  29:     End Property  30:    31:     Private m_Img As Bitmap  32:     Public Property Img() As Bitmap  33:         Get  34:             Return m_Img  35:         End Get  36:         Set(ByVal value As Bitmap)  37:             m_Img = value  38:         End Set  39:     End Property  40:     41:     Private m_file As FileInfo  42:     Public Property File() As FileInfo  43:         Get  44:             Return m_file  45:         End Get  46:         Set(ByVal value As FileInfo)  47:             m_file = value  48:         End Set  49:     End Property  50: End Class  51:    52: Public Class SiteImageCollection  53:   Inherits List(Of SiteImage)  54:   Public ReadOnly Property MaxWidth() As Integer  55:     Get  56:       Dim largest As Integer = 0  57:    58:       For Each s As SiteImage In Me  59:         If s.Width > largest Then  60:           largest = s.Width  61:         End If  62:       Next  63:    64:       Return largest  65:     End Get  66:   End Property  67:    68:   Public ReadOnly Property TotalHeight() As Integer  69:     Get  70:       Dim ttl As Integer = 0  71:       For Each s As SiteImage In Me  72:         ttl += s.Height  73:       Next  74:       Return ttl  75:     End Get  76:   End Property  77:    78:   Public Function GetSprite() As Bitmap  79:     Dim spriteImg As New Bitmap(MaxWidth, TotalHeight)  80:    81:     Dim sprite As Graphics = Graphics.FromImage(spriteImg)  82:    83:     Dim curY As Integer = 0  84:     For Each si As SiteImage In Me  85:       sprite.DrawImage(si.Img, 0, curY)  86:       si.YValue = curY  87:       curY += si.Height  88:     Next  89:    90:     Return spriteImg  91:   End Function  92: End Class

There you have it. A nice sprite generated from existing images. The cool thing is that if you are working with PNGs, your transparencies are preserved, which is what I really needed. The only real down side to this is that all your images are lined up vertically. I'm not quite sure that this matters, but most other generators have the ability to somehow line the images up vertically and horizontally.

But you're not exactly done...

You still have the CSS to deal with...

CSS sprites don't have CSS in the name for nothing. They are best used in a CSS files. Lucky for me, I already have a process for my CSS files to go through (which includes minification, cache, etc). All I need to do is augment it to replace the image reference to the actual URL of the sprite and to calculate the background-position.

Logic Class

I really like having Logic classes (which you have seen in my past blog posts) because they are a pivotal point between the server-side cache mechanisms that I put into place. Here's what my Logic controller looks like:

   1: //C#   2: public class CssSprite   3: {   4:     public static SiteImageCollection GetImagesForSprite()    5:     {   6:         SiteImageCollection col = Helpers.UrlMapping.UrlHandler.Images.AllImages();   7:         //we need to grab the sprite so that the Y-values are usable   8:         Bitmap sprite = col.GetSprite();   9:         return col;  10:     }  11:    12:     public static Bitmap GetCachedSprite()   13:     {   14:         ICache<SiteImageCollection> cache = new Helpers.Cache.CssSpritCache();  15:    16:         return cache.Get().GetSprite();  17:     }  18:    19:     public static string CleanCSS(string currentCSS)   20:     {  21:         string newCSS = currentCSS;  22:         foreach(SiteImage si in CssSprite.GetImagesForSprite())  23:         {  24:    25:             string orig = "background-image: url(" + si.File.Name.ToLower() + ");";  26:    27:             newCSS = newCSS.Replace(orig + "/**/", ReplacementCss(si, false));  28:             newCSS = newCSS.Replace(orig, ReplacementCss(si, true));  29:         }  30:         return newCSS;  31:     }  32:    33:     private static string ReplacementCss(SiteImage si, bool widthheight)   34:     {  35:         string rep = "background-image: url(" + Helpers.UrlMapping.UrlHandler.Images.Sprite().ToAbsoluteURL() + ");";  36:         rep += "background-position: 0px -" + si.YValue + "px;";  37:         if (widthheight)  38:         {  39:             rep += "width: " + si.Width + ";";  40:             rep += "height: " + si.Height + ";";  41:         }  42:         return rep;  43:     }  44: }

   1: 'VB   2: Public Class CssSprite   3:   Public Shared Function GetImagesForSprite() As SiteImageCollection   4:     Dim col As SiteImageCollection = Helpers.UrlMapping.UrlHandler.Images.AllImages()   5:     'we need to grab the sprite so that the Y-values are usable   6:     Dim sprite As Bitmap = col.GetSprite()   7:     Return col   8:   End Function   9:    10:   Public Shared Function GetCachedSprite() As Bitmap  11:     Dim cache As ICache(Of SiteImageCollection) = New Helpers.Cache.CssSpritCache()  12:    13:     Return cache.[Get]().GetSprite()  14:   End Function  15:    16:   Public Shared Function CleanCSS(ByVal currentCSS As String) As String  17:     Dim newCSS As String = currentCSS  18:     For Each si As SiteImage In CssSprite.GetImagesForSprite()  19:    20:       Dim orig As String = "background-image: url(" + si.File.Name.ToLower() + ");"  21:    22:       newCSS = newCSS.Replace(orig + "/**/", ReplacementCss(si, False))  23:       newCSS = newCSS.Replace(orig, ReplacementCss(si, True))  24:     Next  25:     Return newCSS  26:   End Function  27:    28:   Private Shared Function ReplacementCss(ByVal si As SiteImage, ByVal widthheight As Boolean) As String  29:     Dim rep As String = "background-image: url(" + Helpers.UrlMapping.UrlHandler.Images.Sprite().ToAbsoluteURL() + ");"  30:     rep &= "background-position: 0px -" + si.YValue + "px;"  31:     If widthheight Then  32:       rep &= "width: " + si.Width + ";"  33:       rep &= "height: " + si.Height + ";"  34:     End If  35:     Return rep  36:   End Function  37: End Class

So what this does is replace the values where the background-image is equal to an image in the sprite's name. It then appends the style definition to add the background-position (which we found during the Sprite generation) and the height and width and changes the background-image URL. If you add a /**/ right before the background-image definition, the height and width are not added in.

The only thing that I didn't really show you was the AllImages() method (which returns a SiteImageCollection on the images I need sprited) and the Sprite() method in my UrlHandler. That's not that important... I'm sure you can figure out what those do :)

Cache mechanism

When we last talked about my CacheBase class, it was sort of all over the place. I have revamped it to use delegates so that we don't get any weird methods that don't make sense. It's still a little messy since I needed to write 2 classes, one that took 1 generic and another that took 2. Here it is:

   1: //C#   2: public abstract class CacheBase<T> : ICache<T>   3: {   4:     public abstract Func<T> Method { get; }   5:     public abstract string CacheKey { get; }   6:     public abstract CacheItemPriority Priority { get; }   7:     public abstract TimeSpan CacheDuration { get; }   8:     9:     public T Get() {   10:         T CurValue = ((T)HttpContext.Current.Cache[CacheKey]);  11:    12:         if (CurValue == null)  13:             CurValue = Invoke();  14:    15:         return CurValue;  16:     }  17:    18:     /// <summary>  19:     /// Removes the Cache Object from the   20:     /// current cache.  21:     /// </summary>  22:     public void Delete(){  23:         HttpContext.Current.Cache.Remove(CacheKey);  24:     }  25:    26:     private T Invoke(){  27:         return Method.Invoke();  28:     }  29:    30:     /// <summary>  31:     /// Adds the value into the Cache  32:     /// </summary>  33:     /// <param name="Value">Value of T</param>  34:     internal T Insert(T Value)  35:     {  36:         HttpContext.Current.Cache.Add(CacheKey, Value, null, DateTime.Now.Add(CacheDuration), TimeSpan.Zero, Priority, null);  37:         return Value;  38:     }  39: }  40:    41: public abstract class CacheBase<T, P1> : ICache<T>  42: {  43:     public abstract Func<P1, T> Method { get; }  44:     public abstract P1 ObjectDescripter { get; }  45:     public abstract string CacheKey { get; }  46:     public abstract CacheItemPriority Priority { get; }  47:     public abstract TimeSpan CacheDuration { get; }  48:    49:     public T Get(){  50:         T CurValue = ((T)HttpContext.Current.Cache[CacheKey]);  51:    52:         if (CurValue == null)  53:             CurValue = Invoke();  54:    55:         return CurValue;  56:     }  57:    58:     /// <summary>  59:     /// Removes the Cache Object from the   60:     /// current cache.  61:     /// </summary>  62:     public void Delete(){  63:         HttpContext.Current.Cache.Remove(CacheKey);  64:     }  65:    66:     private T Invoke(){  67:         return Method.Invoke(ObjectDescripter);  68:     }  69:    70:     /// <summary>  71:     /// Adds the value into the Cache  72:     /// </summary>  73:     /// <param name="Value">Value of T</param>  74:     internal T Insert(T Value){  75:         HttpContext.Current.Cache.Add(CacheKey, Value, null, DateTime.Now.Add(CacheDuration), TimeSpan.Zero, Priority, null);  76:         return Value;  77:     }  78:    79: }

   1: 'VB   2: Public MustInherit Class CacheBase(Of T)   3:   Implements ICache(Of T)   4:   Public MustOverride ReadOnly Property Method() As Func(Of T)   5:   Public MustOverride ReadOnly Property CacheKey() As String   6:   Public MustOverride ReadOnly Property Priority() As CacheItemPriority   7:   Public MustOverride ReadOnly Property CacheDuration() As TimeSpan   8:     9:   Public Function [Get]() As T  10:     Dim CurValue As T = (DirectCast(HttpContext.Current.Cache(CacheKey), T))  11:    12:     If CurValue Is Nothing Then  13:       CurValue = Invoke()  14:     End If  15:    16:     Return CurValue  17:   End Function  18:    19:   ''' <summary>  20:   ''' Removes the Cache Object from the   21:   ''' current cache.  22:   ''' </summary>  23:   Public Sub Delete()  24:     HttpContext.Current.Cache.Remove(CacheKey)  25:   End Sub  26:    27:   Private Function Invoke() As T  28:     Return Method.Invoke()  29:   End Function  30:    31:   ''' <summary>  32:   ''' Adds the value into the Cache  33:   ''' </summary>  34:   ''' <param name="Value">Value of T</param>  35:   Friend Function Insert(ByVal Value As T) As T  36:     HttpContext.Current.Cache.Add(CacheKey, Value, Nothing, DateTime.Now.Add(CacheDuration), TimeSpan.Zero, Priority, _  37:       Nothing)  38:     Return Value  39:   End Function  40:    41: End Class  42:    43: Public MustInherit Class CacheBase(Of T, P1)  44:   Implements ICache(Of T)  45:   Public MustOverride ReadOnly Property Method() As Func(Of P1, T)  46:   Public MustOverride ReadOnly Property ObjectDescripter() As P1  47:   Public MustOverride ReadOnly Property CacheKey() As String  48:   Public MustOverride ReadOnly Property Priority() As CacheItemPriority  49:   Public MustOverride ReadOnly Property CacheDuration() As TimeSpan  50:    51:   Public Function [Get]() As T  52:     Dim CurValue As T = (DirectCast(HttpContext.Current.Cache(CacheKey), T))  53:    54:     If CurValue Is Nothing Then  55:       CurValue = Invoke()  56:     End If  57:    58:     Return CurValue  59:   End Function  60:    61:   ''' <summary>  62:   ''' Removes the Cache Object from the   63:   ''' current cache.  64:   ''' </summary>  65:   Public Sub Delete()  66:     HttpContext.Current.Cache.Remove(CacheKey)  67:   End Sub  68:    69:   Private Function Invoke() As T  70:     Return Method.Invoke(ObjectDescripter)  71:   End Function  72:    73:   ''' <summary>  74:   ''' Adds the value into the Cache  75:   ''' </summary>  76:   ''' <param name="Value">Value of T</param>  77:   Friend Function Insert(ByVal Value As T) As T  78:     HttpContext.Current.Cache.Add(CacheKey, Value, Nothing, DateTime.Now.Add(CacheDuration), TimeSpan.Zero, Priority, _  79:       Nothing)  80:     Return Value  81:   End Function  82:    83: End Class

So the Sprite cache looks like this:

   1: //C#   2: public class CssSpritCache : CacheBase<SiteImageCollection>   3: {   4:     public CssSpritCache()    5:     {   6:     }   7:     8:     public override Func<SiteImageCollection> Method   9:     {  10:         get { return new Func<SiteImageCollection>(Logic.CssSprite.GetImagesForSprite); }  11:     }  12:    13:     public override string CacheKey  14:     {  15:         get { return "CssSprite_" + Global.DistributionNumber; }  16:     }  17:    18:     public override System.Web.Caching.CacheItemPriority Priority  19:     {  20:         get { return System.Web.Caching.CacheItemPriority.Default; }  21:     }  22:    23:     public override TimeSpan CacheDuration  24:     {  25:         get { return new TimeSpan(1, 0, 0, 0); }  26:     }  27: }

   1: 'VB   2: Public Class CssSpritCache   3:   Inherits CacheBase(Of SiteImageCollection)   4:   Public Sub New()   5:   End Sub   6:     7:   Public Overloads Overrides ReadOnly Property Method() As Func(Of SiteImageCollection)   8:     Get   9:       Return New Func(Of SiteImageCollection)(Logic.CssSprite.GetImagesForSprite)  10:     End Get  11:   End Property  12:    13:   Public Overloads Overrides ReadOnly Property CacheKey() As String  14:     Get  15:       Return "CssSprite_" + [Global].DistributionNumber  16:     End Get  17:   End Property  18:    19:   Public Overloads Overrides ReadOnly Property Priority() As System.Web.Caching.CacheItemPriority  20:     Get  21:       Return System.Web.Caching.CacheItemPriority.[Default]  22:     End Get  23:   End Property  24:    25:   Public Overloads Overrides ReadOnly Property CacheDuration() As TimeSpan  26:     Get  27:       Return New TimeSpan(1, 0, 0, 0)  28:     End Get  29:   End Property  30: End Class

Back in the CSS logic handler, you need to change the CombineCSS method to also do the replacement for the CSS Sprite.

   1: //C#   2: public static string CombineCSS()   3: {   4:     string allCSS = string.Empty;   5:     6:     foreach (FileInfo fi in Logic.Files.GetFiles("~/Content/CSS/", "css"))   7:     {   8:         using (StreamReader sr = new StreamReader(fi.FullName))   9:             allCSS += sr.ReadToEnd();  10:     }  11:    12:     allCSS = CssSprite.CleanCSS(allCSS);  13:    14:     allCSS = Compress(allCSS);  15:    16:     return allCSS;  17: }

   1: 'VB   2: Public Shared Function CombineCSS() As String   3:   Dim allCSS As String = String.Empty   4:     5:   For Each fi As FileInfo In Logic.Files.GetFiles("~/Content/CSS/", "css")   6:     Using sr As New StreamReader(fi.FullName)   7:       allCSS += sr.ReadToEnd()   8:     End Using   9:   Next  10:    11:   allCSS = CssSprite.CleanCSS(allCSS)  12:    13:   allCSS = Compress(allCSS)  14:    15:   Return allCSS  16: End Function

Are you done?

You're sort of done. I've only given you the loose bits of the solution. You will have to display the CSS Sprite onto and image. I didn't show you my code for the actual CSS Sprite MVC Controller Action/Generic Handler. This is needed for you to use the CSS Sprite, after all.

   1: //C#   2:     3: //--------MVC--------   4: [ControllerAction]   5: public void Sprite(string id)   6: {   7:     Response.ContentType = "image/png";   8:     9:     Bitmap bmp = Logic.CssSprite.GetCachedSprite();  10:    11:     MemoryStream stream = new MemoryStream();  12:     bmp.Save(stream, System.Drawing.Imaging.ImageFormat.Png);  13:    14:     Response.BinaryWrite(stream.ToArray());  15:    16:     if(id != null)  17:         Response.Cache.SetExpires(DateTime.Now.AddYears(3));  18: }  19:    20: //-------WebForms-----  21: public void ProcessRequest(HttpContext context)  22: {  23:     context.Response.ContentType = "image/png";  24:    25:     Bitmap bmp = Logic.CssSprite.GetCachedSprite();  26:    27:     MemoryStream stream = new MemoryStream();  28:     bmp.Save(stream, System.Drawing.Imaging.ImageFormat.Png);  29:    30:     context.Response.BinaryWrite(stream.ToArray());  31:    32:     if (context.Request.QueryString["id"] != null)  33:         context.Response.Cache.SetExpires(DateTime.Now.AddYears(3));  34: }

   1: 'VB   2: '--------MVC--------   3: <ControllerAction()> _   4: Public Sub Sprite(ByVal id As String)   5:   Response.ContentType = "image/png"   6:     7:   Dim bmp As Bitmap = Logic.CssSprite.GetCachedSprite()   8:     9:   Dim stream As New MemoryStream()  10:   bmp.Save(stream, System.Drawing.Imaging.ImageFormat.Png)  11:    12:   Response.BinaryWrite(stream.ToArray())  13:    14:   If Not id Is Nothing Then  15:     Response.Cache.SetExpires(DateTime.Now.AddYears(3))  16:   End If  17: End Sub  18:    19: '-------WebForms-----  20: Public Sub ProcessRequest(ByVal context As HttpContext)  21:   context.Response.ContentType = "image/png"  22:    23:   Dim bmp As Bitmap = Logic.CssSprite.GetCachedSprite()  24:    25:   Dim stream As New MemoryStream()  26:   bmp.Save(stream, System.Drawing.Imaging.ImageFormat.Png)  27:    28:   context.Response.BinaryWrite(stream.ToArray())  29:    30:   If Not context.Request.QueryString("id") Is Nothing Then  31:     context.Response.Cache.SetExpires(DateTime.Now.AddYears(3))  32:   End If  33: End Sub

OK, NOW you're done

There you have it. A nice, clean CSS Sprite generated for you. This should save you some time in the long run and will certainly help your users avoid waiting too long for your website to load. If, say, you were to put 10 images into the sprite, for instance, you would save your users 9 requests from the server. That is HUGE for performance. And if your user has a primed cache, it doesn't need to have any request to the server until your distro number changes.

So in the end, we all win with CSS sprites. You get them generated for you at runtime, your users don't have to wait, and your website will be leaner and meaner!