> There is, however, a major problem with Twitters 
  > bootstrap framework... [a] return to all the problems 
  > that came with tables. A complete rejection of the 
  > agreed upon separation between semantic content and 
  > visual design.

First of all, the semantic web was never widely agreed upon in practice [1], simply because almost everyone is interested in generating a particular page layout first. For whatever reason, the standards bodies have tried to pretend that prioritizing the semantic web in the HTML standards would somehow change the behavior of most web developers. It hasn't.

The fundamental issue is that there was a "tight coupling" between page layout and data structure in early web browsers, and while CSS has helped separate layout from HTML, it has not changed the behavior of web developers en masse, who will hack HTML to kingdom come if it gets them the page layout they want. (If you only every use <div> tags, you're not doing semantic HTML.)

But things are actually looking quite bright, because we now have RESTful web services, that operate solely with data structures (typically JSON), so REST is essentially the modern semantic web. That frees HTML from the burden of being a data structure at all.

Now if we could just replace HTML and CSS with something that actually gave you a WYSIWYG page layout, we'd be all set.

[1] As evidence of this, I point you to the HTML parsers of all major browsers, and the Google web crawler/parser. Neither of these assume that pages are valid, let alone semantic.

The web changed. Once again.

Now, the web is responsive. You no longer have to take into account one state for your website, but up to 4 or 5 (if not more). Mobile, tablet, desktop, large desktop.

The reason we changed from table to semantic driven HTML is that tables were not suited for its task anymore. Right now, semantic driven websites, while completely valid, are the the right tool for task anymore.

Sure, you can name all your classed to be 100% semantic and have the whole thing fall down like a waterfall when you turn off the CSS. It's so easy to change the layout of the website!

The problem is, that's not the main issue anymore. Most of the energy is put into making a website... 4 time during development. You can't simply style your blocs all over again and again. The need for a framework get felt, a lot.

Then, if you really need to move your website around during a redesign... why, it's easy! Simply change "large-4" to be "large-8". You want to move it around? No problem, simple copy it into another row. As long as your rows are semantic, you shouldn't have any issue restyling the whole thing.

This must be a troll. Twitter is destroying the web because 2 employees released some open source? Really?

> Bootstrap's documentation however does not encourage this method of use, and instead repeatedly demonstrates classes intermixed with HTML.

Since there's not a link to one in the blog post, are we to assume that the author wrote this rant without submitting an issue or pull request to Bootstrap?

https://github.com/twbs/bootstrap

i would be a lot more sympathetic to all the "don't use tables for markup" people if css had, in all the years of its existence, said "you know what? grids and boxes are a really handy layout tool!" and made them easier to do. spans and flows simply don't cut it. the fact that there are css grid frameworks is ample evidence that people are viewing css's "right way to do things" as damage and routing around it.

a more constructive thesis for this post would have been: Why You Should Use Mixins w/ Bootstrap.

e.g. http://ruby.bvision.com/blog/please-stop-embedding-bootstrap...

If this gets voted to the top I won't sleep well.

First of all, I thought Bootstrap was no longer part of Twitter, officially?

Second, enclosing elements in tables is a whole other level of violating the divide between content and visual design. There's not even a close comparison between row/spans and the restrictions of a tabular element.

...very interesting...