As I become a more knowledgeable developer, I hear the term ‘software framework’ being used more frequently. In my research, I’ve seen an increased interest in frameworks such as Angular and React. To understand what these are, I must first define what is a software framework.

In the article What is a Framework? Software Frameworks Definition, written by Joel Olawanle, walks through a general definition of software frameworks, comments on how they may be applied, and how they differ. Olawanle defines a framework as such, “…is a structure that you can use to build software. It acts as a foundation so you don’t have to deal with creating unnecessary extra logic from scratch.”(Olawanle). Supposing a framework is implemented correctly, developers will save time allowing them to start the project earlier. Additionally, the foundation it provides will not be prone to human error. If the components of a framework are implemented from scratch, then there could be errors that would be much more difficult to fix further in development. Since frameworks can be modified, there is less of a reason to implement their functionality from scratch.

Olawanle expands his definition to other aspects of software development. Before reading this article, I understood that there were both frontend and backend frameworks, but I was surprised to learn that mobile applications and data sciences have their respective frameworks. While reading Olawanle’s article, I noticed a framework I used a few years ago, that being Bootstrap. This qualifies as a software framework as it gathers files needed for a functioning website (.html, . css, and .js) into one structure. This allows the developer to easily make their website without having to build this structure. Angular and React both are classified as Frontend frameworks. Both of these are used for creating interfaces for websites, but each has its unique components. React can use JavaScript to create HTML and CSS files, meanwhile Angular has dependency injection allowing it to more freely communicate with other applications. Each of these creates a specialized purpose for the framework. 

Depending on where a developer is working (front end, back end, etc), there will be a framework that can create a strong foundation for their code. Before selecting their framework, the developer must weigh the tradeoffs between the various frameworks and consider which would support their project principles the most. Using a framework in this scenario would save time and reduce any human error manually implementing these components introduce. In my experience with Theas’ Pantry, the backend does not use any framework listed within Olawanle’s article. After reviewing the documentation, it seems as if the back end would not support these listed frameworks as they are not directly language-compatible. This means more time and resources would be invested into implementing the framework, than potentially implementing components of such individually. Frameworks provide a strong starting point for development teams, depending on the design choices frameworks may be incompatible with certain projects.

Link to Article:

https://www.freecodecamp.org/news/what-is-a-framework-software-frameworks-definition/

-AG

From the blog CS@Worcester – Computer Science Progression by ageorge4756 and used with permission of the author. All other rights reserved by the author.

Communication with those who help construct it or those who will use it is an essential component of any project. Software documentation serves both these ends, as it informs anyone approaching the project of topics such as the requirements used to run said software, instructions, and myriad other facets of the project. Documentation is created for those interested in the specified software/project. The Swimm Team, in their article What is Software Documentation? Types, tools, and best practices, list two types of documentation. These are external and internal documentation.

External documentation seeks to explain all aspects of the software that are not locally accessed. Most of this documentation involves user interaction. This is seen through End-User documentation, which provides the user with instructions on using the software, or through Just-In-Time Documentation, which guides the user while the program executes. A more technical piece of documentation, API documentation, is also considered external. Although most users may not directly interact with the API, this type of documentation is targeted towards developers who may want to expand the API functionality or use it in their projects. Due to the accessibility of external functions, its documentation must be equally accessible to its audience. A level of abstraction must be provided to reach this goal, as uninformed users do not need to understand how the internal systems of the software function, but rather they must learn how to interact with it.

Internal documentation refers to everything ‘behind the scenes’ so to speak. Contrary to external documentation, this documentation is less accessible, meaning it can go into detail on how the systems of the specified software work. Due to this complexity, it serves as a great introduction to the project and can be used to onboard new developers. Internal documentation ranges from information regarding the development cycle of the software, as seen through scheduling documentation,  to design choices seen throughout, which can be found in Software design documents. During the development process, software engineers can refer to internal documentation to ensure their contributions follow the team’s vision for the software.

My experience with software documentation has been limited up until this year. My most recent experience with documentation was through the HFOSS project Theas’ Pantry. Within this project, there was API documentation, available through a .yaml file, which listed all functions and associated code with all API calls. Additionally, there was indirect documentation offered through activities introducing onboarding developers onto the project. Each of these activities introduced a broad topic, such as software architecture, which then transitioned into how it’s applied to Theas’ Pantry. This untraditional form of documentation allows the onboarding developer to interact with the components in a microcosm before they apply their knowledge on the project. Documentation is a vital component of software development. Without supporting texts such as this, clients interested in the project would be met with inaccessibility, and onboarding (in addition to current) developers may not understand the project as a whole which could jeopardize the production of such. 

Link to Article:

https://swimm.io/learn/software-documentation/what-is-software-documentation-types-tools-and-best-practices

-AG

From the blog CS@Worcester – Computer Science Progression by ageorge4756 and used with permission of the author. All other rights reserved by the author.

Anti-patterns are best described as behaviors or approaches to problems that conceptually may help solve the problem, but in practice are a detriment to the process of doing so. In software development, this can come in many forms, whether ‘cutting corners’ by reusing old code or trying to condense behaviors into one class/object. Ultimately these decisions we make as developers come from a place of genuine concern. When these design patterns remain unchecked, they begin to rot in our code and cause many problems, some of which are contradictory to the intention of originally incorporating them.

In the article Anti patterns in software development, the author Christoph Nißle describes several anti-patterns that occur in software development and the consequences of each. Three anti-patterns resonated most with me, as I could see how someone could accidentally implement one of them. The first of which is what Nißle calls Boat Anchor. It represents code that *could* be used eventually, but for the time being, has no relevance to the current program. By keeping this code, the developer is contributing to visual bloat. Not only does this make finding specific lines harder, but once other developers are included on the project they may have questions about how this code will be implemented. To counter this anti-pattern it’s good practice to only keep code that is prevalent to the program’s functionality AND is currently being used by the program. The second anti-pattern I found interesting was Cut-and-Paste Programming. As the title suggests, it occurs when programmers reuse code from external sources without properly adapting it to their current project. This code can also come from the same program. Under both circumstances this code will cause errors, as it’s not a ‘one size fits all solution’, furthermore the code being pasted could have errors. These can be remedied by “creating multiple unique fixes for the same problem in multiple places”(Nißle), but each unique fix requires time and this time could have been spent creating code for the specific problem rather than reusing code. Lastly, the Blob pattern is one that I have personally fallen victim to several times. This pattern has the developer trying to make objects/classes as dense with functionality as possible, but this complexity acts against the single responsibility principle. Classes (and objects) should be solely responsible for one behavior if we include too many then the function of that specific class becomes unclear. The Blob pattern can easily be fixed by dissolving the blob class into several single-responsibility classes. It’s best to catch poor practices such as the Blob early in development to minimize the amount of refactoring that’s needed to fix the code.

As mentioned before, I’ve fallen victim to these anti-patterns as conceptually they save time in the development process. However, the time often saved is eclipsed by the time required to fix errors later in development. Properly following design principles will cause development to require more time, but it should reduce the number of errors that would appear if anti-patterns were used in their place.

Link to Article:

https://medium.com/@christophnissle/anti-patterns-in-software-development-c51957867f27

-AG

From the blog CS@Worcester – Computer Science Progression by ageorge4756 and used with permission of the author. All other rights reserved by the author.

In our work with REST APIs, namely through the HFOSS project Thea’s Pantry, we have implemented new functionality to the database by updating the HTML specifications and creating new endpoints. During this whole process I did not have a concrete idea of what an API was, nor did I understand what made REST APIs any different from their alternatives. 

In the article What is a RESTful API the authors Stephen Bigelow & Alexander Gillis define what an API is, and what components make an API RESTful, in addition to how they can be used. APIs are defined as “code that lets two software programs communicate with one another” (Bigelow & Gillis). This can be seen through our work in Thea’s Pantry as the specification.yaml file provides instructions for the commands which communicate between the backend and database. In a general flow of control the user interacts with software, this piece of software interacts with the API which then shifts control to the external software. From this point the user can directly interact with the external piece of software (in the cases of methods such as delete and put), or the user can fetch information from it which can be returned to their client-side software. REST stands for representational state transfer, this is a type of software architecture that makes communication between two programs more accessible and easy to implement (Bigelow & Gillis). Users can interact with resources from another program using HTTP requests composed of a method, endpoint, header, and sometimes will require a body. RESTful commands, similar to those of databases (get, update, delete.. etc), can be specified by the developers of the API to have unique functionality. This modularity of command functions is one of the benefits of using RESTful APIs. An alternative to RESTful APIs is SOAP. These both achieve similar functionality, but the methods of doing so are different. For example, SOAP is a communication protocol compared to REST which is an architecture style. SOAP is only compatible with .xml files, meanwhile REST can be used with .xml in addition to other file types. It is worth noting that REST and SOAP are not one-to-one alternatives and can be used together. 

APIs allow developers to extend the functionality of their programs by communicating with other programs. This can be achieved through HTML requests (in the case of RESTful APIs) and nodes (in the case of SOAP APIs). REST APIs favor flexibility and modularity, on the other hand, SOAP APIs are more rigid and require concise specifications. Due to its accessibility, RESTful APIs are more favorable in projects such as Thea’s Pantry. I cannot see SOAP being implemented in Thea’s Panty due to its rigidity as seen through the types of files it uses. REST is much preferred here as we can use javascript files to define the HTML requests that the API will use.

Link to Article:

https://www.techtarget.com/searchapparchitecture/definition/RESTful-API

-AG

From the blog CS@Worcester – Computer Science Progression by ageorge4756 and used with permission of the author. All other rights reserved by the author.

Hello to all readers new and old, thank you for taking your time to read this brief introduction to my blog. My name is Andrew George, and I am a senior at Worcester State University. For the past couple years I’ve been studying Computer Science and slowly adding skills to my repertoire. I am once again looking forward to researching topics pertaining to my major, this time with a focus on the Software Design process. I do encourage you, if you have not already, to read my previous introduction pieces found on this blog as those have a more in depth explanation of where I started and how I got to this point today. Thank you once again and until next time!

-AG

From the blog CS@Worcester – Computer Science Progression by ageorge4756 and used with permission of the author. All other rights reserved by the author.

Throughout the cumulative experience I’ve had with testing, most of it focuses on the program’s logic and ensuring that it yields a correct result. One aspect of testing I have no experience with is security testing. Here one must find flaws within a system/program’s security and report them to developers so they cannot be exploited later in the product’s lifespan. Security testing has much higher stakes than that of unit testing as vital information such as consumer personal information and system source code may be leaked if there is a security breach. Therefore, testing security is of utmost importance when releasing a service to the public, as failing to do so will damage the service’s integrity.

Security testing hosts several different types of tests, each of which focuses on different aspects of a system. The article Security Testing posted by user pp_pankaj highlights the principles upheld by this testing and what each test achieves. Some of these tests, Posture Assessment, I found quite interesting. Posture Assessment combines the testing methods of ethical hacking, risk assessment, and security scanning into one report to provide an overall security posture of a system (pp_pankaj). Each of these subtests has a shared goal of having a hacker, hired by the development team, find security vulnerabilities within their system and report it to the team. Another form of testing I found interesting was social engineering testing. This deviates drastically from what we as programmers come to understand tests as. These are emulated attacks through communication such as email. The purpose of this test is to train developers to avoid suspicious engagement and to find new ways to breach a system without making direct contact. Whether a development team gets successfully breached through the socially engineered test is dependent on the team’s understanding of who they must respond to. A few weeks ago I was researching a data breach that happened earlier this year at Microsoft. Hackers were able to control a testing account and had direct access to employees on the project. From here they were able to obtain information they naturally should not have access to. All of this occurred due to developers not knowing that they must not communicate with a testing account.

A general metric as to whether security testing is vital to a project is to consider whether your product is liable for holding personal information that is not your own. If this is the case, then it’s in the development team’s best interest to uphold their principle of confidentiality and integrity by running security tests throughout the lifespan of your product.

-AG

Source: https://www.geeksforgeeks.org/security-testing/

From the blog CS@Worcester – Computer Science Progression by ageorge4756 and used with permission of the author. All other rights reserved by the author.

Beyond Unit Testing, there are several ways one may go about testing a developed system. One of these ways is through combinatorial testing. While researching this topic I’ve seen a couple of steps that I’ve taken in testing that match with this testing method. For example, earlier this semester I learned about how behavior tables can help guide Unit Testing by showing what aspects of a program will be covered by one test. Combinatorial testing achieves a similar effect by first taking all possible inputs (from a pool of predetermined inputs) and then creating a set of tests that will test each unique combination from the pool. A source that helped me grasp this topic is Combinatorial Testing by Shanika Wickramasinghe. In this article, Wickramasinghe provides an example of how they would develop tests. It’s important to note that in this example only combinations are created, NOT permutations, meaning that overall much fewer tests would be needed to fulfill a combinatorial test. This does raise a question for future reading of whether there is such thing as “permutative” testing and how that and combinatorial testing differ.

Using combinatorial testing does provide benefits despite the time it may take to achieve a successful test. Combinatorial tests are all designed to try multiple inputs simultaneously, meaning that both single-fault and multi-fault assumptions will be made in a full combinatorial test. Once these tests are complete, the developer can better understand which inputs cause a problem within their code. Additionally, once the pool of potential inputs is determined, the tester will have a set number of tests they must conduct. These tests may find faults in the program that require specific input that the development team may not have accounted for. Through feedback such as this, the development team can resolve the bug and create ways of handling errors caused by unexpected input. These benefits do come with equally heavy drawbacks. Manual combinatorial testing is possible, however the testers may struggle with creating combinations from a larger input pool. A way this can be solved is by using an automated combinatorial tester. It’s important to note that this can be limited by how intensive the tests are on the hardware of the automated tester. Lastly, the combinations that the test may provide could be so random that it’s nonsensical to test such a thing. This becomes an issue of resources which will vary from developer to developer. Ultimately whether one uses combinatorial testing or not is up to the developer. There are some instances where the cost of conducting one of these tests would be beneficial to the development process, but this is not a “one size fits all” type of test. By using some of the team’s resources, whether it be labor or hardware-bound, combinatorial tests will yield meaningful results as to which areas of the program need further testing.

-AG

Source: https://testsigma.com/blog/combinatorial-testing/

From the blog CS@Worcester – Computer Science Progression by ageorge4756 and used with permission of the author. All other rights reserved by the author.

Performance Testing is a subject I just recently discovered is included in the Software Quality Testing umbrella term. Outside of casually monitoring CPU and GPU usage while using demanding software, I have little knowledge of what Performance Testing is. As a future worker in the Computer Science field, knowing how to report and minimize hardware stress must be an invaluable skill.

During my search to understand what Performance Testing is, I found an article written by Tricentis that included an in-depth explanation for each step (or deviation) of Performance Testing. Before going into detail, Tricentis highlights that performance testing is a direct solution to ensuring the stability and scalability of software. If performance testing is ignored, then users may encounter issues with speed, crashes, or data loss within their application. Testing seems to flawlessly integrate with CI/CD models, as testing should occur whenever a working model is available. This allows the development team to fix issues early before they sink too deep into the product’s foundation. Within the realm of performance testing, some subtests aim to find results for specific metrics. A few that I was unfamiliar with are as follows. Regression tests seek to test two states of the program, one being the current model, and the other being one with experimental changes. Both of these models work together to compare the hardware performance. These are used to highlight whether changes from the most recent version are more damaging to the product. Functional tests focus more on the software aspect of the product. Using this test will yield results on whether the UI elements, APIs, etc are all functioning correctly. Lastly, load testing seeks to put the product under great stress by simulating several concurrent users. Testing with this will give the developers insight as to how their product will function once published to the general public. Metris from load testing will help reinforce the product before it can fail after its initial release.

Performance testing will yield several useful metrics to guide the development of the product. These metrics include hardware bottlenecks, response time, and load time. Issues about these metrics must be addressed during the development process as they directly impact the reliability of their product post-release. Untested hardware is an active detriment to the developers, and will further damage the development process. Following the idea of testing software ‘frequently and consistently’, developers must also view how the current state of their product affects the hardware it will operate on.

-AG

Source: https://www.tricentis.com/learn/performance-testing

From the blog CS@Worcester – Computer Science Progression by ageorge4756 and used with permission of the author. All other rights reserved by the author.