The State of the Art decides, whether an invention is new or not. (see Patent Law )
State of the Art are all publicly available documents.
Additional to books or magazines, there are all websites, programs, and sourcecode that are freely available in the Internet
The major problem is to give evidence that some online ressource was already available at some given date. (see 1)
Timestamping (as used here) is one way, to create evidence to be used in court later (hopefully still in 10-20 years) to certify something :)
If some printed publication contains an article that e.g. says: “The program xyz has released version abc and can be downloaded from http://x.org”, then: * the evidence that this has been published is given
The goal is to provide a guideline for online date forensics
There are problems with paper publications too.
There is lead time, sometimes as much as three months or more and someone body has to review what is being printed. And this certainly applies to peer review publications.
The date of publication can be inaccurate too. It's not uncommon for the sunday dated newspaper to actually be published and on newsstands the day before and likewise the publication date can easily be before any public actually received and views an article in a paper magizine.
Publishing thru electronic means can be more instanious and timestamped by a path of computers it is passed through and by. Certainly the technology of computer communication is going to be faster than print media. This is actually one of the concerns of the print media industry, not being able to keep up.
The way to establish a time stamping verification is to pass it thru multiple and unrelated systems as well as having the IP graniung organizations maintaining their own record of such traffic.