4 ms·
As a web developer, I see two main issues with web styling: First, the web was built around sharing technical papers. That means HTML structure focuses on tho
by ergothus 10y ago
As a web developer, I see two main issues with web styling:
First, the web was built around sharing technical papers. That means HTML structure focuses on those elements that are relevant to papers (outline layout via H* tags, tables of data, not much else), and not the sort of things that marketing and sales want to push. (ads, rails/gutters, etc). Those of use that suffered through the early "slice-and-dice" method of making web pages are painfully aware of that. I'm a big fan of technical papers, and my expectations of flash and glitz are minimal (I, for example, hate the modern trend of not using the full width of my window.). Despite this, I feel we keep trying to stay true to the origins of the Web rather than allowing for the actual USE of the web.
Second, in an effort to keep the content machine-parseable as well as allow for agents of different devices, CSS is applied separately from content/structure (theoretically). Specifically, the concepts used DO NOT MATCH the concepts used in developing desktop applications. Even Flexbox, the most recent attempt to fix this, only loosely relates to the way desktop applications would layout the content.
I'm a huge fan of the GOALS involved in HTML/CSS, but after working on web stuff for over 20 years (not using CSS quite that long), I feel I can say it's been a failure. We've spent that all or most of that time with painful workarounds for basic tasks like: center content (particularly vertically!), Adding a left rail/right rail, filling the height of a container, matching the height of the visible window, making sure layered content is zindexed properly, and those are just the ones off the top of my head. We've invented and reinvented ways to do things like drop down menus, toggleable buttons, modal windows. Heck, from the very start people implemented their own authentication windows because the appearance and capabilities of the browser-based solutions didn't match the demands.
After 20 years, and with the benefit of all the experience of desktop development to add in, I feel like we shouldn't be fighting to manage such basic requests, that we shouldn't be reimplementing field validation and error messages YET AGAIN because even the latest advanced offerings just don't cut it.
We should be able to have:
* "flexible" content (appearances adjusts to visible space)
* machine parseable content
* attractive UI
...without it requiring the dramatic hoop-jumping we have today.
- talmand 10y agoHow else would things go when people use a tool for which it was not designed to do? It's an evolutionary thing. A need is identified, a feature addressing the need is proposed. Then new needs are identified, new features are proposed. The problem wasn't in the tool, the problem was the tragic slowness in the browser vendors implementing the features into the actual browsers. Don't forget that for many of those 20 years we had competing standards of different browsers doing their own thing. If the vendor didn't like a proposed feature of CSS, it didn't get implemented. I wouldn't blame CSS for that. Every complaint that I see about lack of features of CSS or how long it took for the features we do have to get implemented I blame the browser vendors.