By efunction - a day ago
Showing first level comment(s)
aidenn0 - 21 hours ago
gnicholas - 20 hours ago
When it comes to reading text I honestly prefer the old school academic websites (albeit with a max-width set).
Keeping stuff simple even makes maintaining the site simple, I personally run https://www.discoverdev.io and have used zero JavaScript. Inspired by brutalist design principles :)
deepakkarki - 20 hours ago
joncp - 20 hours ago
> <img src=“https://example.com/image.png” alt=“developer in an office working feverishly to hit a deadline on a software project” />
Is this actually a good idea?
This sounds like a typical "hero" or "teaser" image, to capture the reader's attention or interest. I don't think the alt tag does that in a comparable way, for a screen reader for example. So I'd just leave it blank.
It's really frustrating: all articles on accessibility mention the alt attribute, and that you should use them. But few give guidelines on how to write a good alt attribute that is actually well thought out.
When I write a website that explains things, and uses images, I (nearly) never use only images, but usually images + text. If I put an alt description in the images, I basically duplicate the text, and I guess that will confuse screen reader users more than a blank alt text.
Am I totally off base here?
perlgeek - 16 hours ago
It's the same reason why divs won in the first place. You can't dictate the shape of a document to people and it be anything other than a kludge.
Take the site I work on, Great Big Story[1]. We have sections and headers, but our content isn't defined by text, it's defined by graphics and videos, with little metadata sections on the video player page that I suppose could be semantic if we had the right descriptive tags. I use semantic markup when it seems to make sense, but we have so many weird little needs to introduce elements just for styling purposes.
Take a recent example, I need four elements to make a styled file input form. Two labels, one to look like a button and can be clicked, taking advantage of the browser behavior of treating clicks on the label to be clicks on the form. The other label is the actual label, could be just a span if I wanted, it doesn't have the htmlFor attribute set. The actual form input button is hidden using CSS tricks because the actual button can't be styled. Finally there's the filename display, which is set to the filename of whatever was uploaded or the existing one if present. This uses Javascript callbacks to get set.
None of these elements are semantic, legacy cruft keeps the markup from being simple and readable. I suppose you could demand our video presentation site to be a motherfucking website[2], but I fail to see how that will actually improve much of anything and would just hurt our brand. I'm getting a little tired of coder hatred for web fashion. It keeps us all employed.
But I think we're at least 10 years out from being able to actually use semantic markup. Web standards and browser vendors still have some maturing to do.
vinceguidry - 21 hours ago
camnora - a day ago
another-cuppa - a day ago
cjslep - a day ago
S-E-P - 21 hours ago