I suppose that's right in the abstract, since blogs are the most ubiquitous posters of links on the Internet. But academic researchers, other journalists, media critics, users of Digg-type services, and others would surely prefer to have a snapshot-in-time view of a given story they're commenting on.

The current practice of writing over stories strikes me as intellectually dishonest, frankly, since there are people analyzing their coverage in near-real-time.

Interesting, that they should come up as a subject, here, so close to the post about attributions informal writing.

That said, there may very well be some technical issues, particularly with some of the smaller papers, given the sheer data load.

But there's another part of this that cannot be ignored; if the story is no longer likable from anywhere, and can be found, it can't be used to track the progression of the story.

As an example; stories about that long line of trucks moving from Iraq in this area less than 24 hours prior to our invasion of Iraq, have been hard to find these last few years. Yet the majors spent almost a full day tracking that story.