• The Worth of Learning a Language Just for Fun

    I really enjoy learning new programming languages. Actually, I love learning new programming languages. I always have something in my mind that I want to learn. Some things I just want to tinker with a bit and others I want to have a deep understanding of.

    Two languages I’ve been trying out recently are Scala and Flex. Both are very enjoyable languages to use. I’m running into a bit of a problem though. How do I find the motivation to really learn these languages and is it worth it? In the past, I’ve usually just learned languages because I’ve had to for my job. But this time it is just for fun. Sure, someday in the future I might want to get a job using one of these languages but not yet. I just find them to be fascinating languages.

    So the question is this: If I am just learning these for fun, should I invest the time to have a deep understanding of it or just a superficial understanding? Will it benefit me that much? If so, how do I motivate myself to learn it? If one day I want to get a job using either of these languages then how do I get the work experience an employer would be looking for?

    Whatever I do, it is critical for myself and every other programmer to continue learning. Whether that comes from thoroughly learning a single new language, superficially learning a multiple languages, reading books, or simply trying new things in your language of choice, it is vital for our continual improvement.

  • The Cost of Quick and Dirty Programming: Technical Debt

    Why do programmers always make such a big deal about best practices, proper object oriented design, unit testing, and the like? Chances are if you ask somebody who champions these ideas they’ll respond “it makes for higher quality software” or something along those lines. They are probably right about saying it makes for better software. Ideas such as these won’t produce “perfect software” but they definitely improve the chances of better software with easier maintenance.

    I am one of those programmers that believes that unit testing, thoroughly thought out OO design, best practices, etc. contribute to better software. But unfortunately, taking these actions does not come for free. They all take time and energy which translates into money.

    Unit testing is a great example. Writing tests for code can definitely add more time to a project. Possibly quite a bit more time. Many people, myself included, often just ignore testing and just wait for bugs to pop up before they are fixed. It is at times very difficult to find the motivation to write the unit tests and do what I know I should be doing.

    But this view is very short sighted.

    Writing unit tests, while time consuming up front, can dramatically save time, money, and energy later in a project. The reason is because they enable you to find bugs earlier when they are easier to fix which means time and money saved later. It is far easier to fix a bug early in a project soon after it is introduced than when it has been in the system for months and covered but by other layers of code.

    A concept that paints a good picture of the consequences for this quick and dirty approach is called technical debt. The idea of technical debt is related to commonly understood financial debt. It is best explained by Martin Fowler in his post on this metaphor.

    In this metaphor, doing things the quick and dirty way sets us up with a technical debt, which is similar to a financial debt. Like a financial debt, the technical debt incurs interest payments, which come in the form of the extra effort that we have to do in future development because of the quick and dirty design choice. We can choose to continue paying the interest, or we can pay down the principal by refactoring the quick and dirty design into the better design. Although it costs to pay down the principal, we gain by reduced interest payments in the future.

    Doing things quick and dirty, as us programmers often do, has consequences that we frequently ignore or thoroughly take the time to understand. Not writing unit tests or ignoring problematic software designs creates debt that can be difficult to pay down later. Allowing excessive technical debt without reason is a quick way to heavily burden a software project. It has happened to me just as I’m sure it has happened to you.

    But technical debt can sometimes be an OK thing. Sometimes software just needs to get done ASAP or sometimes it is a rarely used or temporary piece of software. In circumstances like this I believe it is ok to take on a certain amount of technical debt.

    This is something I need to remember. Software doesn’t always need to be perfect (if it even can be that). Sometimes it is far more important to just get it done. It often is very tempting to analyze and analyze and analyze in an effort to get that perfect piece of software but never actually get anything programmed. By keeping this metaphor in mind I will hopefully think through how I approach software in the future and think through the consequences of the actions I take.

    Additionally, I like the metaphor of technical debt as a way of working with non-technical users to explain the the time is needed to “do it right” the first time or to pay down the debt accumulated in other phases of the project. Most people have some kind of financial debt so this concept seems like an easy concept for people to understand.

    It also seems useful in explaining to other programmers who may be hesitant to embrace new methodologies and practices. In my work in ColdFusion, I’ve often found other programmers wondering why go through all the work some best practices like unit testing. They think it is too much work and don’t bother. But using this metaphor seems like an effective way to demonstrate the fallacy with that thought process.

    In short, before you program something, take the time to consider technical debt and whether you want to avoid it or accept some in the short term. I plan to do so and hope it will lead me to better think through my actions and consequences in the software I write

  • Monitoring a Form for Changes with jQuery

    JQuery constantly amazes me. Its ability to simplify tasks that ordinarily would have been complicated is truly wonderful. Today I found myself amazed at jQuery’s power yet again.

    First a little background. My coworker had been working on a project which required her to notify users if they were leaving a page with unsaved changes in a form. When she asked me if I had any ideas I researched online to see if anybody else had a solution.

    What I found were some pretty great ideas that allowed me to come up with a solution that is pretty darn cool. The solution uses jQuery’s “:input” selector, “each” function, and “data” function to loop through every form field and store whatever the initial value is. Then it uses “window.onbeforeunload” to check the form before user leaves. The code loops back through all the form fields and checks if the current value is different from the initial value. If so, it notifies the user and give them a chance to stay on the page and save.

    Here is the code.

    $(':input').each(function() {
        $(this).data('initialValue', $(this).val());
    });
    
    window.onbeforeunload = function(){
        var msg = 'You haven\'t saved your changes.';
        var isDirty = false;
    
        $(':input').each(function () {
            if($(this).data('initialValue') != $(this).val()){
                isDirty = true;
            }
        });
    
        if(isDirty == true){
            return msg;
        }
    };

    Let me break this down a bit.

    The first part of this code finds all form fields: $(':input'). It finds all inputs, textareas, selects, etc. jQuery returns all found elements as an array which I then loop through with the “each” function.

    The next line takes each element in the array and stores a variable called initialValue containing the current value of the form field (remember, the array is a collection of all found form fields on the page) and associates it with that form field. If the “data” method doesn’t make much sense, the jQuery Docs can probably explain it a bit better.

    Next, the function set to the window.onbeforeunload event is run whenever the user tries to leave the page. When run, it once again finds all form elements and compares all current values to the initial values saved earlier. If some changes have occurred, it returns a message to the user.

    One limitation I know of is if new form elements are added to the page after it loads and the initial values of the fields are stored. That could probably be easily remedied but it is just something I haven’t yet done.

    Overall this is a very useful piece of code and is a beautiful example of what can be done with jQuery. I look forward to using this in future projects. As a reference, here are two websites that provided some very useful information about this: CFSilence and My Dog Boris.

  • Easy Object Oriented JavaScript

    I love object oriented programming. Not because I have the crazy misconception that it is the ultimate programming paradigm but because it fulfills my desire to organize, categorize and modularize my code. Consequently I like to bring OO ideas into whatever language I’m working on. Now being primarily a ColdFusion developer I work a lot with JavaScript. But one thing that really took me a while to get a grasp of is making my JavaScript code object oriented.

    I happen to really like JavaScript…mostly. It has some really weird quirks sometimes and parts of it seem poorly designed but overall it is a very powerful and fun language to work with. Its power and quirkiness is very evident when trying to write object oriented code. There seem to be at least three ways to write OO code with JS with all three being completely different. JavaScript just isn’t clear when it comes to OO and doesn’t have any easy to use mechanisms for doing things like inheritance. This is what really tripped me up for a long time. I’d read tutorials and books that would cover the subject (OO) but they’d all do things differently and not explain why they were doing it the way they were.

    Thankfully though I stumbled upon Douglas Crockford’s JavaScript videos this year and opened my eyes on a few things when writing object oriented code. This is what I’ve learned and how I do OO with JavaScript now.

    First thing first, objects. In JS, creating an object is as easy as doing the following:

    myObject = {};

    Wow, that was easy. No “new” operator or any other mumbo jumbo (is new considered an operator?). Just that one simple line. That simple line creates a singleton object as soon as it is executed. Usually though, you want to have some actually have some functionality in your objects so let’s do that now. There are a couple of ways to do that.

    // OPTION 1
    myObject1 = {
        sayHello: function(){
            alert("hello world");
        },
        sayGoodbye: function(){
            alert("goodbye");
        }
    }
    
    myObject.sayHello(); // alerts hello
    myObject.sayGoodbye(); // alerts goodbye
    
    // OPTION 2
    myObject2 = {};
    
    myObject2.sayHello = function(){
        alert("hello world");
    }
    
    myObject2.sayGoodbye = function(){
        alert("goodbye");
    }
    
    myObject2.sayHello(); // alerts hello
    myObject2.sayGoodbye(); // alerts goodbye

    Both option 1 and option 2 do the same thing but have somewhat different syntaxes. I don’t know if I use either more than the other. It just depends of the circumstances.

    Now one thing you might be wondering is how to have multiple instances of this object? JavaScript doesn’t really have a concept of classes, only objects, which is what this is. Essentially, you’d need to make a new copy of the object every time you wanted a new instance. This is actually very, very easy to do. Simply wrap the object in a function that returns a new instance. Here is how it works.

    function myObject(){
        var obj = {};
    
        obj.sayHello = function(){
            alert("hello world");
        }
    
        obj.sayGoodbye = function(){
            alert("goodbye");
        }
    
        return obj;
    }
    
    myObject = myObject();
    myObject.sayhello();
    myObject.sayGoodbye();

    The wrapper function acts as a class and constructor for the object. Every time the myObject function is called it returns a new instance of your object. You can even have private methods and can pass in constructor parameters.

    function greeter(name){
        var obj = {};
        obj.name = name;
    
        // private function
        function greet(){
            alert("hello " + obj.name);
        }
    
        obj.sayHello = function(){
            greet();
        }
    
        return obj;
    }
    
    myObject = greeter("Dustin");
    myObject.sayhello(); // alerts "hello Dustin"

    Simple and concise. Another nice thing about this method is that inheritance is a breeze.

    function parentObject(){
        var obj = {};
    
        obj.sayGoodbye = function(){
            alert("goodbye");
        }
    
        return obj;
    }
    
    function greeter(name){
        // This causes the object to "inherit" from parentObject
        var obj = parentObject();
    
        obj.name = name;
    
        // private function
        function greet(){
            alert("hello " + obj.name);
        }
    
        obj.sayHello = function(){
            greet();
        }
    
        return obj;
    }
    
    myObject = greeter("Dustin");
    myObject.sayhello(); // alerts "hello Dustin"
    myObject.sayGoodbye(); // alerts "goodbye" - sayGoodbye is inherited from parentObject

    What is happening is that greeter calls the parentObject “constructor” to get an object, then modifies it with its own methods, essentially inheriting the functionality from parentObject. Very powerful but very simple. This is only possible because JavaScript objects are modifiable. Many language do not allow objects to be modified with new methods but JavaScript is an exception. Scala, Ruby, and other languages allow this and, in my opinion, this makes for some very concise yet powerful features.

    Unfortunately, this method of doing OO in JS is not something I’ve seen even though it is very straight forward. I suspect that is because object mutability seem to be a rare thing among most mainstream languages like Java. As languages like Ruby gain in popularity I imagine that might change.

  • Old School Batch Programming and Code Generation

    Recently I had the opportunity to do some old school Windows batch file programming at work. At first I was not too interested in using a language (is that even the correct word for it?) as old as DOS, but after I got into the project it was actually rather fun.

    First, a little back story about the project. The company I work for uses many mainframe systems today that it has used for decades. But, like most big companies, it has many modern server systems as well. To use these systems together we often have to move files containing large amounts of data between servers and mainframes. To accomplish this task, a piece of software called Connect:Direct is used.

    Because of the way we used this software, the setup and maintenance was really a pain. First, to setup a file to be moved to/from a mainframe and server, a Connect:Direct “process” has to be setup which consists of a file containing some propriety code to instruct Connect:Direct on what the process should do and how it should run. The file was just text containing some code with a CDP extension. In addition to the CDP file, we had two more simple files per process. One simple batch file so the users of the process can run it simply by double clicking on an icon and a simple text file to be used by the batch file which contains some instructions on which CDP file to use and some variables to use. What that meant was three files per process and a maintenance nightmare. Why a maintenance nightmare you ask? Because a server was recently done away with so every single of our 300 files had to have some details changed.

    Since my coworkers and I knew we would have to change many files we decided to streamline the process a bit. Initially we decided to just combine the batch files into one and have a simple menu so the users could choose the process they wanted from there. Since I was the one writing the batch file I got to work.

    After a bit of work though, I realized that more could be done than merely combining the batch files – we could actually do some code generation. That’s right, code generation with a batch program. Now obviously it couldn’t be too complex seeing how this would be done in a batch program, but a rudimentary form of code generation is possible.

    What I ended up doing was generating the text and CDP files on the fly, executing the Connect:Direct process with them, and then deleting the files afterwards. All the configuring is done in the batch file by setting few variables containing the changing data. What this resulted in was the elimination of 300 files and and extremely easy way to modify processes as well as an easy way to add new processes in the future.

    Before this project I had done a very limited amount of batch file work before. Perhaps what we did could have been done in some other “better” way, but I doubt in the time it took to get this up and running. Batch files might not be very sophisticated, but don’t underestimate there power. Under the right circumstances, they can really prove to be invaluable.