Not So Quick Styling

08 Oct 2020

Styling made easy through UI Frameworks but no so quick to use

As the minutes go by, and the sun rises and sets cyclically, I find myself learning a UI framework called Semantic UI for the past week. I must say, Semantic UI is definitely a great alternative to creating styles from scratch. This is especially great for people who are not looking to be front-end developers and just need a quick touch up on the design of their projects. However, from my experience I did not enjoy using Semantic UI as much as I thought I would, felt that I could have achieved the same results with raw HTML and CSS, and think that Twitter Bootstrap is a better UI framework.

First off, I had to spend a good amount of time to learn the basics of Semantic UI and then built up frustration whenever I wanted to style an element with specific styling. Compared to other UI frameworks like Twitter Bootstrap, Semantic UI’s philosophy and structure of their framework is all about using the “natural language.” For those who are unsure about what it means to use the “natural language” when it comes to coding something, this means that the code and structure of typing reads like a regular English sentence. To Semantic UI’s credit, they were definitely able to achieve that goal of making a “natural language” UI framework. However, the main gripe I have with this framework using the “natural language” is that the keywords become so simple to the point where things can get messy. For example, when you want to style an element, sometimes an element will require the “ui” class and in other times an element does not need the “ui” class and only needs the “fitted item” classes. This can be very confusing for people who are just beginning to learn web development and UI frameworks. I think that this is where Semantic UI oversimplified styling and split up styling into too many small pieces to a point where remembering the terminology becomes a headache.

Secondly, I felt that I could achieve the same results with raw HTML and CSS when I was experimenting with Semantic UI this past week. For some backgound, up to this point I have now experienced using Semantic UI and Twitter Bootstrap. When I was creating a few simple web pages and replica projects, I once again found it very frustrating whenever I wanted to create my own bespoke styles and override certain built-in styles from Semantic UI. Notice that I said “once again,” this is because that phrase ties into my overall opinion on UI frameworks. I understand that I can easily override styles by referencing the Semantic UI class names in my CSS file, however through experience and research, I found this to be bad practice for larger projects and a bad habit to keep in general. This is because if I were to build a larger project rather than a simple functionless web page, targeting the UI framework’s class names directly will cause problems like applying the style to many other elements that are also using those class names but are meant to have different styles. That problem can also lead to confusion if I try to track the styling for certain elements and try to read the style sheet another day.

Lastly, if I had a choice, I would choose Twitter Bootstrap over Semantic UI. Even though Twitter Bootstrap’s class name structure does not read like the “natural language,” the naming scheme is much more clean and organized in comparison to Semantic UI’s naming scheme. The main factor that makes me prefer Twitter Bootstrap is that there is no need to remember which situations require me to include the “ui” class name. Furthermore, if I ever needed a style from Twitter Bootstrap, all I need to do is add the class name to the list of classes. For example, if you want to make a container of elements, all you have to do is give a div element the class “container,” whereas in Semantic UI, you would have to remember that creating a container requires you to give the classes “ui container.” I also like that Twitter Bootstrap allows you to essentially edit styles in the class attribute. For instance, say you have a container:

<div class="container">Hello</div>

Then you decide that this container should have a width of 75%. All you have to do is add the class “w-75” where w is the class for changing the width of an element and 75 represents the percentage value:

<div class="container w-75">Hello</div>

This small feature about Twitter Bootstrap is really great for styling elements on the fly without needing to touch any CSS file. Some people may say that this naming scheme like naming the class for width with just the letter, w, is vague and confusing, however I think that this is great organization and structure for adding classes to an element. Instead, I think that Semantic UI’s naming scheme is too simple, to a point where adding many classes can lead to confusion when revisiting the code.

Overall, in my opinion, I did not enjoy using Semantic UI, feel that you are better off using raw HTML and CSS, and think that Twitter Bootstrap is a better UI framework. However, this is not to say that I think Semantic UI is a bad UI framework by any means. I still think that Semantic UI is purposeful and it is clearly good as people stil use it. However, I found that the schema of Twitter Bootstrap is much more efficient and productive than Semantic UI. In the end, I think that UI frameworks is great for those who are not the best at UI design or have a project where styling is not the main purpose, and I very much prefer using raw HTML and CSS to style any project I work on.