By design, the web is a "1. send me the document <-> 2. here it is" transaction, not a series of many small notifications. By design, the url() property almost certainly wasn't intended to be dynamic. This is clearly 'bending the established rules' — cleverly, admittedly.
> By design, the web is a "1. send me the document <-> 2. here it is" transaction, not a series of many small notifications.
That was true in 1998, but most of the web has been turning, by design, into what you might call a series of many small notifications ever since then. Gmail's been doing it since 2004. Today, almost all large sites are running Google Analytics or something like it, which track everything this article discusses, and operates on constant micro transactions. All web apps are built on many small notifications, and many of them even use websockets which was explicitly, by design, built for streams of micro transactions.
> By design, the url() property almost certainly wasn't intended to be dynamic. This is clearly 'bending the established rules'
There was never a rule against, even if dynamic usage wasn't expected or imagined (which I find unlikely). CSS allows it, therefore it's allowed by design.
The `url()` function wasn’t intended for tracking, sure. But my point is that it doesn’t matter, since it is accepted that the website you’re on can track you to begin with. I don’t think anyone in the standards bodies is trying to prevent that.
By design, the web is a "1. send me the document <-> 2. here it is" transaction, not a series of many small notifications. By design, the url() property almost certainly wasn't intended to be dynamic. This is clearly 'bending the established rules' — cleverly, admittedly.