CSS3 proven to be Turing complete?

Re:CSS3 proven to be Turing complete? by PeCaN on Tuesday December 15, 2015 @01:13PM (#10739356)

Heck, it means that theoretically you can't verify a CSS3 stylesheet for correctness because you don't know the stylesheet will ever finish rendering.

In practice it's probably not a problem since you don't have infinite HTML, but it is theoretically slightly disturbing.

I'm an advocate of provably-correct, non-Turing-complete languages, which really should include CSS....

[ Parent | on HN ]

Re:CSS3 proven to be Turing complete? by Dylan16807 on Tuesday December 15, 2015 @01:39PM (#10739527)

This proof does not imply that. It cannot loop without you clicking all the boxes.

[ Parent | on HN ]

Re:CSS3 proven to be Turing complete? by PeCaN on Tuesday December 15, 2015 @03:47PM (#10740355)

A static analyzer doesn't know that you won't click all the boxes though.

[ Parent | on HN ]

Re:CSS3 proven to be Turing complete? by Dylan16807 on Tuesday December 15, 2015 @04:17PM (#10740539)

It still renders in O(n) time where n is the number of elements.

You don't need Turing completeness to have trouble caused by user input. Imagine a million check boxes where the only rule is that they shrink by a pixel when checked. That should be about the same difficulty to analyze as the 110 rule checkboxes. Each input only affects the styling of 1-3 nodes.

[ Parent | on HN ]