-
Notifications
You must be signed in to change notification settings - Fork 60
Uki core does not use css. At least the part with selectors and separate stylesheets. You may wonder why. To answer that let us first understand why we style pages with css.
This is the task css is great at. You can style pages with thousands of html tags with a relatively short css declaration. You can target any particular element, while retaining the ability to style clusters of them.
But why do we have to style complex DOM in the first place? Because when it comes to web apps html is an extremely low level language. Writing web interfaces with html is like coding a desktop app in assembler. Maybe not that bad but you get the point.
When you build interfaces out of uki views you don’t have to style particular dom nodes. Instead you setup self-contained interface elements like Slider or Table. Thus you don’t need thousands of them. Tens are usually enough. There’s no need to think up a separate technology to style them.
Without css selectour you’d have to write inline styles in html for every element. And this styles will duplicate for similar looking elements. That would increase html page size and thus load time will grow.
With uki we’re in javascript world. When you want to style several elements similarly you can create a function which will add those styles
to any element required. Thus you won’t get any benefit from adding a separate stylesheet.
With css stylesheets you have to write them, merge them and then deploy somewhere. This is a separate tech process you have to manage. And then you have to load css on every page. Thus you have to add <link> for all css files you use. And you have to change this tags whenever you css files are renamed. It all ads to complexity of you app.
There’s no simple way to modify css rules from javascript. Thus you’ll end up with some code that’s unmanageable at runtime.
With uki you create and style views with javascript. Adding a css-selector based styling will split you view logic into two separate places. Two separate technologies would be responsible for styling your code. With a very fuzzy boundary between them.
Thus for every change you’d have to update two things instead of one. And for every bug you’d be debugging two places instead of one.
If you want to support all browsers with css code you have to use conditional comments and/or css hacks. There are several drawbacks with this approaches. First this relies on browser detect, not feature detect. Second you write more code.
When you use uki view attributes it is guarantied by the framework that they’ll work exactly the same in every browser.
Uki does not use css for simplicity reasons. Adding css styling will increase complexity not reduce it.
Though if you feel css is right for you app you can still use className attribute and write .css files.