Chris! I told you
Once considered th
Once considered th
Once considered th
Stop dancing like
Chapter 1. Once
Chapter 1. Our st
Chapter 1. Our st
Chapter 1. Our st
Ships were lost du

Tiffany, you reall
Tiffany, you reall
Tiffany, you reall
Joe's Bar and Gril
Quietly, Quiggly s
Quietly, Quiggly s
Quietly, Quiggly s
Quitetly, Quiggly
Concrete may have
Concrete may have
<<      >>
We've recently discovered a new method to create and maintain a scalable, performant, and extensible web application. This is the presentation of that method: Creating a dynamic web application from a database - A quick introduction to the technique and example code. I had used this approach in another project. This project was much more complicated: there were a lot of business logic classes, more than 400 SQL queries, about 800 lines of code for each one, and, because it was a big project, there was a lot of change management; it was a web app that took over the main page of a newspaper (about a hundred and fifty pages), and it had to be released periodically. The system did a lot of tasks on the server side, was very fast, and provided many features that make a lot of sense for such application. But the development and change management processes in my previous application were very demanding and time consuming. So, I tried to make this faster, and, in addition to the web server, I created a PHP script that created a SQL database, used PHP code to read the database, and created the necessary pages for the application. I was able to make it work. The only problem was that the application could only work on a single server, and that the changes were made every time from the main page. For example: you can only go to the home page; when you want to go to a page from the menu, you must first go back to the home page and then to the requested page. The logic is simple: if you are at the home page, you cannot go to any other page. That means that a lot of users will not be able to see all the features that the application provides. In my previous approach, this did not matter. The users were only a few and, for my purposes, it was not important if some of the pages did not show up. Using a single server for a dynamic web application means that you only have one server to serve your application. A single server will never have enough resources to serve the amount of requests that would be made. If, for example, your web application needs to make about 50 requests, when there are only 3 servers that are available to answer requests, it is useless. It is just as useless as making 50 requests to your e-mail program and not being able to open the e-mails because there are not enough servers available to receive requests. To address that problem, I solved it by using several PHP scripts to connect to the database. The server receives the requests and, for each one, creates a new connection to the database. Every new connection is an independent transaction. I used a file called server.php that did the initialization: read the configuration file and load the data from the database. For every new connection that needed to be opened, the server.php script uses the PDO extension to create a new PDO object. It uses the name that the user typed, and then it uses the credentials that the user typed, and the database. The script calls the method createNewConnection, which creates the connection and commits it. This means that there is only one script that connects to the database. A single script is a problem, because, as stated before, only one script can connect to the database. In order for that script to accept new requests, it must wait until it is closed, which is only possible when the script exits. That means that if I have a script that is used to create a new connection, it cannot open a connection to the database when a new request is made. The same happens to the other scripts that use a new connection. To solve this, I created a file called createconnection.php, which does the initialization for the script that creates new connections: creates a new PDO object, sends a query that loads the initial page (if it exists), accepts the user's name, user's password, and the database and username, and then, stores it in the configuration file. The script is called every time the server receives a request. If the initial page does not exist, the script adds it to the database and then sends the file that created the initial page back to the user. The same script checks every two minutes if the pages have changed, if a page was changed, it will be displayed to the user, and, if a user wants, he can change that page. There is a new connection for every request. I created another script, called getconnection.php, which is called every time a new page is requested by the user. When that script receives the request, it connects to the database and sends back the data to the user. Now that I have several scripts to create connections and open and close connections, I don't need to wait until a server is available to open connections to the database. I created a few of these scripts, and I have to server that, one for each function or service that the web application needs to be able to provide. For example, in the home page, I create a new connection to the database, send the initial page to the user, then, when the user requests a page from the menu, I will use that page as a starting point, then I will create a new connection, go to that page, and then send the final page to the user. One of the problems that I did not manage to solve is that I am only able to use one user and one password to access the database. So, the user name and password that I am using in the scripts that I created are the ones that the user put in the configuration file. This is not a problem for me, but this will be a problem for other users that do not know what the credentials are and have not configured the web application. They will not be able to access the application. When the user opens the application for the first time, the server.php script will open a connection to the database. When the application is launched, it will ask the user for the credentials. The data will be saved in the configuration file. The user will see a message and he can choose the initial page, which, by the way, must exist before the application can be launched for the first time. In addition, the user must choose a user name and a password that will be used by all scripts that have to connect to the database. It is a simple text file with the user's name and the password. Then the server.php script will create a file called config.php. It will load the data from the database. It will check if the user is connected, then it will check if there is a initial page that must be displayed. If there is no initial page, then the data that it has loaded and the initial page that it wants to display to the user will be sent to the browser. The config.php script is executed when the server.php script is loaded. The scripts are executed in the order that I wrote them. If, for example, the application needs to display two pages, the server.php script will wait until it is finished to execute the createconnection.php script, then the getconnection.php will be executed, and then it will start the server.php script. The server.php script loads the data from the database, creates a new object for the PDO extension, and then creates the initial page that it will display to the user. This includes a welcome message, the user's name and password. Now, if I want to change something on the website, I change the page that I have created and that will be executed to create a new connection to the database. This way, I have complete control over the pages that will be displayed and I do not have to wait for another server to open a new connection. The script will create a new connection to the database, it will connect to the database and then check the database, if there is an initial page that must be displayed, it will create the page and then the new connection will be closed. The connection will be closed because the method that created the connection will be called, and this means that the connection will be closed when the user closes the browser. This also means that there will be no more connections and I can use my server for other purposes. Now, if I want to change anything on the website, I just change the file that I created, execute the script again, and the server.php script will do everything that it did the first time and will also check if there is a file that it should display to the user. It will check if the file exists and will return the initial page if it is different from the page that it has already created. It will return the page that has been changed to the user. The getconnection.php script will create a connection to the database, load the data from the database, and then send back the data to the user. If there is a new connection that needs to be created, the createconnection.php script will create it. When the script closes the connection, it will wait for two minutes, and then it will start it again and then close the connection. The getconnection.php script is executed when the server receives a request. It will wait until it is finished to execute the createconnection.php script, and then the script will execute the getconnection.php script. When the script is finished, it will send back the data to the user. With this setup, I am able to have multiple pages in the same web application. If I have a page that opens the configuration file, and then I have another page that opens the data from the database, and a third page that allows the user to change the database data, all of them will be executed, and none of them will block the others. I will not need to wait until a server becomes available. All the requests will be processed, and I will be able to change the pages whenever I want. This is very important, because when I created this application, I did not know how many pages would I have in my application, and it is not a good idea to make the server wait for a server to become available. The database connection must be closed every time it is finished using it. If a request is made, it will wait for a connection to be created, then, when the request is closed, it will wait for a couple of minutes, then it will close that connection. If I want to open another connection, the script that I use to create a connection will create a connection and then, when the script is finished, it will close the connection. When I want to change the database, I need to create a new script and to call the script in a new way. If I want to update the name of the user, I need to call the createconnection.php script, because the user is already connected. I then need to call getconnection.php, which will create a new connection. I must pass the user name and the password. The connection will be closed when it is finished. If I want to update the page, I will create a new script and I will call it with the page number, like getconnection.php?page=1. This will send the data to the user, which must be a page that does not exist yet, and then, I will call the getconnection.php script, with the page that I want the user to see. The data will be returned to the user, and the new page that I created will be displayed to the user. As you can see, the scripts that I created are all different. They work independently. I need to create several scripts to support all the requirements that my application needs. Now, there are other aspects that I did not manage to solve, but I did not try to solve them because I was satisfied with the setup that I was able to accomplish, or because I did not know how to make it work. The first thing that I did not know how to accomplish is to be able to change the pages that will be displayed to the user. I will use the page, that I wrote in step 17, as an example. If a user wants to change the first page, I need to call the getconnection.php script. I need to pass it the page that I want it to display to the user. The page that it will display to the user will be returned to the user. The getconnection.php script must find the page that I want the user to see, and then I will create the page, send the user back to the home page, and then, when the user is back, the data will be sent to the page that I want. The getconnection.php script does not have to be changed, but if I want to change the script that changes the pages, I will have to make a new script. I will not have to add any new libraries to the system. All of my pages are independent. I will also need to change the way in which I send the data to the page. I need to send the data to the page before I close the connection. In addition, I will need to handle the situation that happens if the user does not have access to the page. I will need to call a different page. The other thing that I did not know how to accomplish is to check the username and the password. If I want to get the user's name from the user, I need to make a connection to the database, and then, when the connection is closed, I need to check the username. This is not a problem. I need to check if the user is a normal user, or if it is a super user. There is a page in my database, that contains all the information that I need to check. The page contains a table that contains the user name and the level of access that the user has. In addition, if the user has access, I will send the page where the user's data is stored to the user. If the user does not have access, then I will show the login form to the user, and then, when the login form is submitted, I will check the username and the password. If the username and the password are correct, then I will connect to the database, load the data from the database, and then send it to the user. If the data was not updated, the script will return the login form to the user and will display it. I need to check this information every time that the page that changes the data is called. As you can see, this was not a big problem. I think that I created a solution that works with my scenario, but if you want to create your own application, you will have to make changes. So, I hope that I have not scared you. I hope that you were able to see what the problem is. The steps that I wrote may seem long, but they are not, because all of this has been written on the computer and the monitor, and I do not need to copy the codes of each step in this book. This is a working example of the solution that I created. I will also post this example in GitHub, so that it is easier for you to check it out. It will be called step12Example. # Chapter 9: