News for hackers. Stuff that matters.
Re:CSS3 proven to be Turing complete? by fixermark on Tuesday December 15, 2015 @01:19PM (#10739390)
I think there's a flaw in the model.

"Assume that the amount of HTML currently loaded is finite but sufficient for all of the state to be properly rendered."

So consider <html> <div class="myContainer"> <div>1</div> <div>2</div> <div>3</div> . . .

I've never met an HTML streaming solution that won't just stream this sequentially, which means no </div> for myContainer will ever be emitted (and therefore the HTML will never be well-formed, and the CSS will never have sufficient information to lay out).

If the HTML were streamed such that <html><div class="myContainer"></div></html> were received, and then the interior of myContainer were streamed, that'd be a different story, but that doesn't exist. So I don't think this blog post's argument that CSS3 is Turing complete works for any real implementation.

Re:CSS3 proven to be Turing complete? by d-Pixie on Tuesday December 15, 2015 @01:26PM (#10739443)
What relevant CSS rule can't be applied to this? We know our previous context and everything before which is what CSS acts on anyway, no lookahead rules in CSS :)

And yes, there could come an yet as to unknown element later that changes the display, displays on top of using position:absolute; for example, of current content. But not that changes how current content itself is displayed.