- From: Tab Atkins Jr. <jackalmage@gmail.com>
- Date: Tue, 15 Sep 2015 14:01:07 -0700
- To: Henrik Andersson <henke@henke37.cjb.net>
- Cc: www-style <www-style@w3.org>
On Mon, Sep 14, 2015 at 11:05 PM, Henrik Andersson <henke@henke37.cjb.net> wrote: > I have recently seen a talk about a problem with the web [1] that > brought up a to me previously unknown issue. The issue is about content > knowing if it is visible or not. Right now content does not know when it > is visible. It leads to several problems. > > The first problem is performance. Some content is expensive to render. > While Flash has traditionally been the punching bag in this aspect, it > is far from the only culprit. Knowing when content isn't visible would > allow web applications to be good citizens, resulting in far better > results than any automated solution. > > The second problem is measuring impressions. Some content, mainly ads, > need to know when the user has seen it. > > The third problem is a security problem, clickjacking. For all zero of > you who don't know what clickjacking is, it is when one website embeds > another in a deceptive way in order to fool the user to interact with > the embeded content in an unintentional way. Such as pressing buttons > they didn't mean to. > > The speaker shows a POC solution to these problems that involves some > rather internal knowledge of the user agent implementation (Chrome > specifically IIRC), but he does not propose any form of API so it is > hard to judge how it would work on the web. > > Can, and most importantly, should, CSS be used to solve this issue? This is being worked on, just not within the CSSWG: <https://github.com/slightlyoff/IntersectionObserver/blob/master/explainer.md>. This is API is designed to solve your first two use-cases, at least. It might be useful for the third, but it's not intentionally solving that. ~TJ
Received on Tuesday, 15 September 2015 21:01:55 UTC