This week, I am going to discuss some of the benefits of the output-oriented software development methodologies we have been studying in class, like Scrum and Agile. I have built many projects both on my own and for my career from the ground up, which is where the value of these methodologies is most apparent to me. The blog post I will be referencing throughout this discussion can be found here. This post provides an in-depth overview of the Agile methodology, including lower-level analysis of different stages of the process, benefits and purposes of each stage, and various best practices for different stages. I chose this post because it correctly emphasizes the importance the Agile methodology places on continuous delivery, which is something that has been a cornerstone for me in various projects.

A project that comes to mind is actually the same project I discussed in this week’s post for CS-343. A friend came to me and asked for something way above my skill level, and it was daunting to imagine planning out the whole project and implementing/deploying some large framework in any reasonable amount of time. If I hadn’t remembered something an old professional connection had said to me years prior, I don’t think I ever could have known where to start: “your code has to exist before it can be made better.” I think that is what the heart of these continuous delivery methodologies is really about, and it really is a true statement for many reasons. The first, which this blog touches on and which we have discussed briefly in class, is that your perspective will always be better or wider after the work is started. Another reason that statement holds true, and perhaps the most practical reason (at least in my case), is that half the time, I don’t even know what I am doing when I start. It would be impossible for me to design anything, because I don’t know anything. For me, learning my way through a project is almost all the fun, and that is not really an option without methodologies like Scrum and Agile, even if I don’t know I’m using them. This blog post discusses release management a lot, which is not so applicable to personal projects, but I think it connects nicely if you compare a release to a checkpoint or successful implementation of a new feature. In my project, as I built it, I started with: what. Then, I said, I guess I could try that. Then, I implemented a single shoddy feature. I saw my feature and saw a way to make it better. Compound that a hundred times, and I had a full framework that I had approached incrementally, following Agile principles and release management without even knowing it. I think that is a testament to the elegance of these continuous delivery oriented methodologies, and I will more consciously follow these principles moving forward.

From the blog CS@Worcester – Mr. Lancer 987's Blog by Mr. Lancer 987 and used with permission of the author. All other rights reserved by the author.

In this post, I will be discussing some basic concepts about abstraction in object-oriented programming. The resource I reference can be found here. I chose this topic because in the first few group activities, I was made aware that I had forgotten some of the key differences between kinds of abstraction (at least in Java), such as the difference between an abstract class and an interface. Though we reviewed these differences in class, I wanted to continue to find specific examples that would allow me to better understand the concept. Abu’s blog post briefly explains what abstraction in object-oriented programming is, some of its benefits/purposes, and some elementary examples of how to implement (no pun intended) abstract classes and interfaces.

The largest point of note I want to discuss is the benefit of modularity when abstraction is implemented. I recently finished up a personal coding project, in which modular development was a necessity. The project involved a central framework capable of ingesting a very specific kind of data from different websites, but each of the websites had a different way of providing that data. The most efficient answer I could think of was to make the central framework independent of the websites themselves, and to make it deal only with the data after it was obtained. I then created a module for each website that held the data, and implemented whatever logic I needed to in order to get the data and pass it to the central framework. I used Python for this project, but in hindsight, I may have been better suited to implement an object-oriented programming language like Java. Each module could have been a concrete class that extends some abstract class Parser, as each of my parser modules in my Python project used the same methods anyways. Each module had a get_url_and_interactions() function, a get_parse_function() function, and an stm() function (we can leave that abstract for now). It would have been very simple to use an object-oriented programming language, as I could have abstracted what it means to be a parser.

Now, this worked fine; since I was the only developer on the project, and had full visibility over all of the code, I didn’t need to check if a module implemented, for example, the get_url_and_interactions() function. I knew that it did. Moving forward, if I were coding more collaboratively with anyone else, or even if I just wanted to code according to some better practices moving forward, I could have used abstraction to standardize the parsers to ensure they all implemented certain methods or functions.

Though a little cliche, I do think it is very nice and elegant that the same thing can be accomplished in so many different ways. I think it is a good metaphor for most problems worth solving.

From the blog CS@Worcester – Mr. Lancer 987's Blog by Mr. Lancer 987 and used with permission of the author. All other rights reserved by the author.