I always like to suggest to try Vaading before truly write a new UI library from the ground up. They have been nice enough to expose the individual behaviours as mixins so you can easily compose your own component so at least that heavy-lifting is done.
Hey Adalis, and thanks a bunch for commenting. Indeed, building a UI library from scratch is an incredibly expensive process, and in most cases, I'd personally advise against it. The only exception, as discussed in the video, is that making a really bad initial decision can lead to one simply having no choice, but either building one from scratch, or picking an exiting one and modifying it. Today, the latter is easily possible, due to utility libraries like Tailwind, etc. At one point in time, the developer experience of the initially chosen tool can become a huge pain, and bad developer experience directly translates to frustration, loss of time, regressions etc. which directly translates into loss of money. Thanks again for sharing your views. Cheers
@@programmer-network Tailwind is nice, though I don't like how you forced to copy-paste divs with 30-50 classes. While I appreciate that its there, I much prefer the direction with Daisy UI where those initial classes are collapsed but you still have an overwrite them if you wish. One thing I do wish is for Daisy to show the used classes on their docs so that its possible to know exactly what is used and how to overwrite, but that is mainly for juniors. Anyhow, one extra contributor I seen is people underestimating the modern UI libraries and assuming that you cannot customize them. Like, they do not give the middle-step a try before deciding to write their own library. I've sadly seen way too many internal UI libraries which quite literally Material with a slightly adjusted padding and corners :D
I always like to suggest to try Vaading before truly write a new UI library from the ground up. They have been nice enough to expose the individual behaviours as mixins so you can easily compose your own component so at least that heavy-lifting is done.
Hey Adalis, and thanks a bunch for commenting.
Indeed, building a UI library from scratch is an incredibly expensive process, and in most cases, I'd personally advise against it.
The only exception, as discussed in the video, is that making a really bad initial decision can lead to one simply having no choice, but either building one from scratch, or picking an exiting one and modifying it. Today, the latter is easily possible, due to utility libraries like Tailwind, etc. At one point in time, the developer experience of the initially chosen tool can become a huge pain, and bad developer experience directly translates to frustration, loss of time, regressions etc. which directly translates into loss of money.
Thanks again for sharing your views.
Cheers
@@programmer-network Tailwind is nice, though I don't like how you forced to copy-paste divs with 30-50 classes. While I appreciate that its there, I much prefer the direction with Daisy UI where those initial classes are collapsed but you still have an overwrite them if you wish.
One thing I do wish is for Daisy to show the used classes on their docs so that its possible to know exactly what is used and how to overwrite, but that is mainly for juniors.
Anyhow, one extra contributor I seen is people underestimating the modern UI libraries and assuming that you cannot customize them. Like, they do not give the middle-step a try before deciding to write their own library. I've sadly seen way too many internal UI libraries which quite literally Material with a slightly adjusted padding and corners :D
So simple yet so hard for so many.