Nuno Silva

reCaptcha - Part 2

In this second part, i will show how to make an ASP.NET Validation control that uses the reCaptcha framework.

I named this control ReCaptchaValidator. This validator extends the abstract class BaseValidator. And so soon we arrived to the first point.

ReCaptchaValidator could implement the interface IValidator, but he would lose a functionality that is very useful. the property ValidationGroup. This property allows you to tell the validator that he will only make the validation when an input with the same validation group makes a post back.

Example

<asp:Button ValidationGroup="test" runat="server"/>

When the above control makes a post back to the server, only the controls in the validation group named “test” will be called.

So, the ReCaptchaValidater extends BaseValidator. Starting explanation, all the action will occur in the EvaluateIsValid, a method inherited from BaseValidator:

protected override bool EvaluateIsValid() {
     string parameters = ReCaptchaParameters();

     string result = ValidateCaptcha(parameters);
     bool isValid = result.Equals("true\nsuccess");

     if( !isValid ) {
          ErrorMessage = "Your are not a human!";
          Controls.Add( new LiteralControl(ErrorMessage));
     }
     return isValid;
}

This method will call the private method ReCaptchaParameters that will return all the parameters needed to make the Web Request to the reCaptcha url. This Web Request will be made by the method ValidateCaptcha to which the parameters are passed. (ReCaptchaParameters and ValidateCaptcha are defined below)

Then the result of the Web Request is validated. If the result is different from the string “true\nsuccess”, then the user didn’t inserted the captcha correctly (or a BOT is trying to register in your site).

/// <summary>
/// Gathers the parameters for reCaptchaValidator
/// </summary>
/// <returns></returns>
private static string ReCaptchaParameters() {
     string response = GetFormValue("recaptcha_response_field");
     string challenge = GetFormValue("recaptcha_challenge_field");
     string IP = HttpContext.Current.Request.UserHostName;

     StringBuilder builder = new StringBuilder();

     builder.AppendFormat( "privatekey={0}&", PrivateKey );
     builder.AppendFormat( "remoteip={0}&", IP );
     builder.AppendFormat( "challenge={0}&", challenge );
     builder.AppendFormat( "response={0}", response );

     return builder.ToString();
}

/// <summary>
/// Validates the capcha inserted by the user
/// </summary>
/// <param name="parameters">parameters to verify</param>
/// <returns>response of the capcth</returns>
private static string ValidateCaptcha( string parameters ) {
     try {
          WebRequest request =
          WebRequest.Create( "http://api-verify.recaptcha.net/verify" );

          request.ContentType = "application/x-www-form-urlenpred";
          request.Method = "POST";

          StreamWriter writer =
          new StreamWriter( request.GetRequestStream() );
          writer.Write( parameters );
          writer.Close();

          //Response
          HttpWebResponse webResponse =
          (HttpWebResponse)request.GetResponse();
          StreamReader myReader =
          new StreamReader( webResponse.GetResponseStream() );

          string response = myReader.ReadToEnd();

          myReader.Close();
          webResponse.Close();

          return response;
     }catch( WebException ) {
          //Do Some Exception threatment
          return string.Empty;
     }
}

In the Render method we will render all the code necessary to show the reCaptcha validator:

/// <summary>
/// Writes the javascript to show the validation Captcha
/// </summary>
/// <param name="writer"></param>
protected override void Render(HtmlTextWriter writer) {
     writer.Write(
     @"
     <script>
     var RecaptchaOptions = {{theme : 'blackglass', tabindex : 2}};
     </script>
     <script type='text/javascript'
	 src='http://api.recaptcha.net/challenge?k={0}'&
	 gt;</script>
     <noscript>
     <iframe src='http://api.recaptcha.net/noscript?k={0}'
	 height='300' width='500'
	 frameborder='0'></iframe><br/>
     <textarea name='recaptcha_challenge_field' rows='3'
	 cols='40'></textarea>
     <input type='hidden' name='recaptcha_response_field'
	 value='manual_challenge'>
     </noscript>", PublicKey );

     dummy.Visible = false;

     base.Render(writer);
}

Noticed the dummy.Visible = false in the render? This is the second point i wanted to talk about. The validators that inherit from BaseValidator were made to validate other controls. In this case, we don’t want to validate any asp.net control. The control we want to validate is provided by a third party code (all the javascript in the render method).

Because of this fact, the validator has an hack. A TextBox with the name “dummy” is created and it’s ID is passed to the property of BaseValidator, ControlToValidate. This way we can surpass the limitation of BaseValidator. Notice that, if this validator didn’t extend BaseValidator, but implemented IValidator, this hack would not be needed.

I made this hack on the event OnInit:

protected override void OnInit(EventArgs e) {
     dummy.ID = "aaa";
     Controls.Add(dummy);
     ControlToValidate = "aaa";

     base.OnInit(e);
}

An that’s it. A free captcha control that you can use in all yout web applications.

For your commodity, the ReCaptchaValidator code is available here (in C#) and here (in VB) (Note: the VB code was created using a converter so it might not be exactly like the C# original file).

reCaptcha - Part 1

What is reCaptcha?

reCaptcha is a Captcha program that can tell the difference between a human and a machine. And Plus, it helps to digitalize books. And how does it do it?

Recapcha uses a 2 word challenge. A first word, that is an already known word, and a second word that comes from a digitalized book. If the user inserts the known word correctly, then the second word is assumed as correct and you helped to translate a digitalize one word.

A more complete explanation can be found here, at reCaptcha official page.

How to use reCaptcha

1. Register in the website. With your registration you will have to give the domain where you intend to use reCaptcha.
2. Then, a private and a public key will be generated (for localhost test proposes, you can use the same keys as in production).
3. Insert the following piece of javascript into your page:

   <script type="text/javascript"
   src="http://api.recaptcha.net/challenge?k=your_public_key">
   </script>
   
   <noscript>
   <iframe src="http://api.recaptcha.net/noscript?k=your_public_key"
   height="300" width="500" frameborder="0"></iframe><br>
   <textarea name="recaptcha_challenge_field" rows="3" cols="40">
   </textarea>
   <input type="hidden" name="recaptcha_response_field"
   value="manual_challenge">
   </noscript>

And a box, similar to the one below, will appear.

Then, for this process to work, a request must be made to http://api-verify.recaptcha.net/verify with the right parameters in order to validate the captcha.

The steps to do this are:

1. Insert a submit button and place the above script and that button inside a form.
2. Click on the submit button so all the input fields can be posted into the server.
3. Then make a Web Request to “http://api-verify.recaptcha.net/verify” with the right parameters. These parameters are:
   1. privatekey - the private you already generated
   2. remoteip - the DNS name of the remote client
   3. challenge - the text you wrote in the reCaptcha control. The value of “recaptcha_challenge_field” sent via the form
   4. response - The value of “recaptcha_response_field” sent via the form
4. Verify if the response is correct.

The Web request you make to the above url will retrieve a response separated by a \n. In the first part of the response, the possible values are true or false. If the first part has the value false, then the second part is the error code. You can find the list of error codes here in the section “Error Code Reference”. This page also has all the information you need make your implementation.

In Part 2 of this article i will show how to implement an ASP.NET Validator control that uses this API.

Divs instead of tables

One of the complicated things when we are doing a design is to doing it only using divs and not tables. Some of us always use tables. And why? because it’s simple? Maybe using div’s is not so dificult after all.

First, there are several reasons to use a div layout instead of a table based layout:

• Flexibility: The main reason not to use tables in the main layout of your website is, in my opinion, flexibility. Has you can see in csszengarden, using divs to divide the content, turns the page layout into a flexible layout where you can make almost everything you want without change any html. If you want change the style of your page, you change the css. This is the way it should be made.
• Separation of Concepts: Using a table based layout you are mixing two concepts: Style and Content. Why? Because using a table based layout you only need to write some xhtml (has you will see below).
• Accessibility: Div layouts are more accessibility friendly.
• SEO Optimization: A Div layout is more optimized for search engines crawls.
• Bandwidth: using divs saves bandwidth because you need less xhtml to reach your objective.

I will take has example webgames.zi-yu.com, a website whose design is integrally constituted by divs but, in it’s original form, it was made using a table.

Let´s start with the layout of the page:

As we can see in the image above, there are 4 main areas, each one divided in subareas:

• A: the main title
• B: the left menu
• C: the main content
• D: the footer

Using tables this layout would be very simple to make:

	<table>
		<tr>
			<td rowspan="2">A1</td>
			<td colspan="2">A2</td>
		</tr>
		<tr>
			<td colspan="2">A3</td>
		</tr>
		<tr>
			<td rowspan="2">B1</td>
			<td colspan="2">C1</td>
		</tr>
		<tr>
			<td>C2</td>
			<td>C3</td>
		</tr>
		<tr>
			<td colspan="3">D1</td>
		</tr>
	</table>

With div’s the layout would be even simpler (and more easy to read - in my opinion, semantically correct). Let’s start by the beginning. We have 4 different areas, so we will have 4 different divs:

<div id="allContent">
	<div id="topMenu"></div>
	<div id="leftMenu"></div>
	<div id="mainContent"></div>
	<div id="footer"></div>
</div>

Now that we have the sections defined, let’s take care of the sub sections:

<div id="allContent">
	<div id="topMenu">
		<div id="logo"></div>
		<div id="topMenuUp"></div>
		<div id="topMenuDown"></div>
	</div>
	<div id="leftMenu"></div>
	<div id="mainContent">
		<div id="mainMenu"></div>
		<div id="mainContentLeft"></div>
		<div id="mainContentRight"></div>
	</div>
	<div id="footer"></div>
</div>

At this point all the sections of the page are defined. We only need to apply the css style to adjust it’s position and size. In this style (and in all the styles i make) i always consider a resolution of 1024×768. For this resolution, the maximum size i use is 1000, because of the vertical scroll of the browser. Normally this scroll is hidden but when it appears, the size of the layout is reduced and the horizontal scroll appears (thing that we don’t want).

So, considering the width of the layout 1000 pixels, let’s start with the logo:

#logo {
	width:252px;
	height:110px;
	float:left;
}

#topMenuUp {
	height:20px;
	border:0px;
}

#topMenuDown {
	height:90px;
	padding-left:0px;
	text-align:center;
	border-bottom:solid 1px black;
}

What the css above will do? Normally the 2 divs are positioned by default one below the other. So we will maintain that behavior for the div’s with the id topMenuUp and topMenuDown (only the height of them will be adjusted). Regarding the div with the id logo, we need to align it to the left, so we use float:left to obtain that result.

Next, the main content and the left menu. Generally i use the following stategy: if i have several containers in the same level, i separate them inside larger divs. So, i separated the leftMenu (area B of the picture above) and the div’s in the area C. The div’s in the area C i putted them inside a larger div in order to aligned them all at the same time into the right of the leftMenu div:

#leftMenu {
	width:252px;
	float:left;
	display:block;
}

#mainContent {
	width:745px;
	float:right;
	display:block;
}

Regarding the different sections inside the div mainContent, there is a top section (area C1), that will fill mainContent top, and two other div (areas C2 and C3) that will fill the rest of the div.

So regarding the div mainMenu (area C1), we will apply the following style:

#mainMenu {
	height:48px;
	padding:0;
	margin:0;
	clear:both;
}

The clear:both style is very important. It allows the div mainMenu not to have any content at it’s left or at it’s right. So, if other div (like the ones below) have a float style applied to it, that div will appear below mainMenu div.

Because the div mainMenu has the clear:both applied, the divs mainContentLeft and mainContentRight can be aligned with floats without a problem:

#mainContentLeft {
	float:left;
	width:584px;
}

#mainContentRight {
	float:right;
	width:160px;
}

Finally the footer. To this div i’ll just apply a clear:both style just to guaranty that no other div will overlap it:

#footer {
	height:48px;
	clear:both;
}

And that’s it, a layout that, at first sight, would be difficult to be made with divs, but, with simple css, it becomes very simple.

In conclusion, using divs, floats, clears and widths you can make almost any layout you want.

Master Pages

In the end of 2007, i worked for a big company that still used in their sites asp.net 1.1. In fact, in their main site, the entire middleware is made in C# 1.1 (their site was launch at the beginning of 2008).

For this reason, there must still exist several applications made in ASP.NET 1.1 that will need some maintenance and new developments.

The challenge here is to develop a control that can simulate master pages in asp.net 1.1 similar to the one that exists in asp.net 2.0+. You make the master page using a .ascx control and make a .aspx webform has you would make in asp.net 2.0, meaning:

• In the ascx of the master page you will have something like this:

<asp:placeholder id='placeholder' runat="server"/>

• And in the aspx you will have something like this:

<ctrl:contentplaceholder master="Master.ascx" runat="server">
	YOUR CONTENT
</ctrl:contentplaceholder>

where contentplaceholder is the control described below and master the name of the file of your master page. Has you will see below in the code, you can also omit this field and configure it’s value in the app settings of the web.config:

The code control is simple has below:

public class ContentPlaceholder : Control {	

	private string masterPage;

	public string MasterPage {
            get {
		if( masterPage == null ||
		masterPage == string.Empty ) {
		     return "~/" +
                     ConfigurationSettings.AppSettings["MasterPage"];
		}
		return masterPage;
	     }
	     set { masterPage = value; }
	}

	private void ChangeControls( Control src, Control dst ) {
		while( src.Controls.Count > 0 ){
			Control c = src.Controls[0];
			src.Controls.Remove( c );
			dst.Controls.Add( c );
		}
	}

	private void LoadMaster( Control master ) {
		PlaceHolder contentPlaceholder = (PlaceHolder)
                        master.FindControl( “placeholder” );
		ChangeControls( this, contentPlaceholder );
		ChangeControls( master, this );
	}

	protected override void OnInit(EventArgs e) {
		Control master = Page.LoadControl( MasterPage );
		LoadMaster( master );
	}
}

In Resume, in the OnInit event the controls of the web page are removed from the aspx page, the content of the master page is inserted in the aspx, and the content previously removed is inserted in the placeholder of the master page:

• 1. The OnInit is called. At this point the aspx only has the content inside (the control contentPlaceholder with the content of the page inside).
• 2. The content is removed from the aspx.
• 3. The content of the master page is added into the aspx control tree.
• 4. The original content is inserted inside the placeholder of the master page. This placeholder must have the id = ‘placeholder’ (name hardcoded for the example inside the above control )

I hope this example helps someone that still work with the old framework 1.1. This can be handy in big projects where you have hundreds of aspx to change.

Wrapp a div around other div’s with float

Today i discovered something interesting with div’s.

Imagine that you have the following code:

<div id='mainDiv'>
   <div style='float:left'>Div #1</div>
   <div style='float:right'>Div #2</div>
</div>

And you wanted that the div named “mainDiv” to have a background color. Well in IE7 everything works fine with the xhtml code

above. But in FireFox the “mainDiv” there is an unusual behaviour.

Below is a image that ilustrates what you get in Firefox (on the left) and what you get using IE7 (on the right) with the xhtml code above:

Notice that the main div, in IE7, warps all content, but in Firefox the div remains with a minimum height.

Io solve the problem you could try (has i did) to put a height in the “mainDiv”. Although it worked (for the content that i

had inside the Div), you never know what the the height of the mainDiv is going to be.

The correct solution for the problem must be one that adapts the height the mainDiv to the maximum height of it’s content.

The solution is simple:

<div id='mainDiv'>
   <div style='float:left'>Div #1</div>
   <div style='float:right'>Div #2</div>
   <div style='clear:both'></div>
</div>

Adding the empty div with the style property “clear” set to “both” makes the mainDiv’s height extend to the full height of

it’s childs, making the xhtml generic for all browsers.

(UPDATE)
Due to the several solutions in the comments i will comment there here one by one.

Alternate Solution 1: the clearFix : thank to Joao A.

The solution in in fact very elegant and we only need to add something to your style sheet. The solution objective is to add

the pseudo element :after into that

“mainDiv”:

#mainDiv:after {
	content: ".";
	line-height: 0;
	height: 0;
	visibility: hidden;
	display: block;
	clear: both;
}

The pseudo element objective is to add some content after the element “mainDiv”. So the soluction passes by adding the

content “.”, removing the height of the content, hide that content and finnally clear both side of the div. in fact we are

doing the same has the original soluction without adding the extra div to the markup.

Alternate Solution 2: adding the float:left style to the “mainDiv” element : thanks to Vítor Pires.

This solution is actually very simple:

The solution is simple:

<div id='mainDiv' style='float:left' >
   <div style='float:left'>Div #1</div>
   <div style='float:right'>Div #2</div>
   <div style='clear:both'></div>
</div>

Alternate Solution 3: adding the overflow style to the “mainDiv” element : thanks to Jeff and href=’http://btm.anfo.pl/’ target=’_blank’>BTM.

very simple just like the previous one. Add the overflow style property with the value ‘hidden’ or ‘auto’.

The solution is simple:

<div id='mainDiv' style='overflow:hidden;' >
   <div style='float:left'>Div #1</div>
   <div style='float:right'>Div #2</div>
   <div style='clear:both'></div>
</div>

The only difference bettween the usage of overflow:hidden and overflow:auto is just that if you have a fixed height in the mainDiv, with overflow:hidden, the content will the content that surpasses the size of the div will be hidden, but with overflow:auto a sroll will appear.

The Solution 2 and the solution 3 are actually very good (and simple) to use in specific situation. if you want the content of the mainDiv to appear below it, than use solution 3 (or 1). Otherside, if you want the content to appear side by side with the div, use the float property in the main element.