Recently I got interested in allowing people to add feedback, but from previous experience I noticed that such feedback mechanisms will allow spammers to compromise the site contents. Now I decided to implement an OpenID mechanism, allowing people to log in with their already-existing credentials from Google, Yahoo, and other providers.
For this I found on the internet the http://gitorious.org/lightopenid API which will help me authenticate the consumers with their OpenID account. So, inside the LightOpenID class I defined few blocks.
This statement is executed every time the page is loaded, and it simply will set in my session the contact name and email, according to the authentication protocol.
But getting the name and email are not trivial task, for which I defined a getContactDetails() method inside that class.
This function, getContactDetails will return an array containing the email and the name (friendly, full name, or email, whichever is available, preferably the first).
How can be seen in the fist statement, I used a function getAllProviders that will return me all the identity providers recognized by my site. The code for this function is getAllProviders.
Finally, all this functionality is used in the PHP page, as in Usage.php.
The code is functioning for the moment only on http://my.sorescu.eu.
Have fun coding!

<?if(@$_SESSION['contact']){?>
	<a class='dms-button' href='#' onclick='$(this).siblings().toggle()'>
		<%icon:loggedin%><?&@$_SESSION['contact']['email']?>
	</a>
	<a style='display:none' class='dms-button' href='?OpenIDAuthenticationProvider='><%icon:logout%>Logout</a>
<?}else{?>
	<a class='dms-button' href='#' onclick='$(this).siblings().toggle()'><%icon:login%>Login</a>
	<?foreach(LightOpenID::getAllProviders() as $providerName => $provider){?>
		<a style='display:none' class='dms-button' href='?OpenIDAuthenticationProvider=<?=$providerName?>'><img src='<?=$provider['icon']?>'/><?=$provider['label']?></a>
	<?}?>
<?}?>

Sample of a small C# SMTP server with the following features:

• 17KB source code;
• five files of code only;
• able to receive mails;
• relay (forward) mails to external servers;
• able to resolve the MX machine IPs for external mails;
• mails to external recipients (external elaying) requires authentication.

1. "Listener" daemon that takes the incoming SMTP requests;
2. "MXDiscovery" module able to identify the IP of the external recipient servers;
3. "Sender" component that is responsible for forwarding the message to external recipients;
4. "SMTP" library with useful functions;
5. "SMTPServer" the main application that starts the listener.

I hope this information might be useful for others, too.

When I decided to publish my CV on a web page (http://dragos-matei.sorescu.eu) I wanted to publish also my email, but without letting the bots read it.
So, I came with the following solution:

1. encode the email in some kind;
2. let the browser decode it after the page gets loaded.

For this I used the following encoded text:

	
		<%icon:email%>[b36]831898108[b36][b36]1348423876[b36]@[b36]27913917[b36].[b36]16438[b36]
	

The JavaScript to decode it is attached.
A small demonstration of the algorithm is put also below:

Source	Encoding	Decoding

Now the site http://www.sorescu.eu/ offers a search functionality in the left bar.
The code is relatively simple, with two components, client-side and server-side.
Notes:

• the PSP is a wrapper syntax on top of PHP;
• after loading data, the client is responsible with formatting with JQuery the result;
• in case of fast typing it is possible that multiple distinct requests to come in a different order (I shall fix it by adding a counter logic in few minutes).

My inspiration was http://www.mkyong.com/java/how-to-resize-an-image-in-java/ with the following small and compact code:

BufferedImage resizedImage = new BufferedImage(IMG_WIDTH, IMG_HEIGHT, type);
Graphics2D g = resizedImage.createGraphics();
g.drawImage(originalImage, 0, 0, IMG_WIDTH, IMG_HEIGHT, null);
g.dispose();

.

My wife, the user of the Screen Capture application as described in  http://www.sorescu.eu/article/120220221500/Free-Screen-Capture-application explained me that she needs to run it also while in Full Screen mode.

Somebody requested me to generate a web page with a wall background that should change the wall colour based on the user's desire.
I thought a transparency-enabled background will suffice, with a dynamic background of the image.
Steps:

1. Take the background image from my customer;
2. Generate a mask image as a copy of the background image, and remove the area (set it to black) where I want to have transparency;
3. Run the Java program that is written below;
4. Use the final result image and put it in a web page.

Seldom I was in need to perform client filtering on tables and other data, and eventually save the filter. One solution was sending the filter to the server and generate the filtered page.
The functions below (I am using them few years already in Internet Explorer) are used to store and get the properties from fragments.
Some useful clues:

• When the page is loaded, check all the fragment parameters and update the controls and perform filtering, etc.;
• Whenever you update a filter in a text field, etc., you can save it in the fragment;
• You can send the link with fragment to anybody else, and if the first bullet (filter executed on page loading) is accomplished, the URL receiver will see the same page as you.

As I started to develop small applications in C#, I thought why shouldn't I use the same paradigms that I used in Java... dynamic code execution, execution without compile/build/deploy process, hot-swap, etc.

• the list of CS files that should be compiled;
• the list of external binaries, if required;
• the title of the application;
• the entry point (the class name that contains the Main function).

One of my last tasks is to build clear requirements based on customer WinWord documents; I also do have a strong feeling that the documentation will change in time, and it will be a dirty work to trace it.
As a consequence, I wrote a small utility to render to the browser the contents (not rendition) of a diven DOCX file. And the transformation (in incipient phase) is below.