which is a paper about the first six years of string theory, 1968-1974. As you can see, the letters identifying the archive have disappeared, and the numbers - whose format is very different from the previous one - is shared by all fields of science.
Moreover, the purely numerical format of the numbers looks extremely incomprehensible and doesn't make it clear that the number is a preprint number rather than a random sequence of digits. To make this fact clear, you really have to attach the code "arXiv:" at the beginning which is a waste of space because the same five or so letters that were used before were also able to indicate the subfield which is kind of more useful and information-rich than just saying that it's from the "arXiv".
At any rate, I would recommend the arXiv staff to switch back (and rename the papers that were posted under the new ugly identifiers). If they run out of numbers in the old format, they can add 2600 extra papers in every arXiv by using identifiers like astro-ph/0704A00 ... 0704Z99 after 0704999. If 3600 papers is not enough either, one can add other analogous codes with more letters in them.
Alternatively, if an archive became too big, it should simply split into several archives. For example, astro-ph could be split into astro-ph (new) and cosm-ph etc.
Figure 1: Apple's arXiv widget: click to download. Does it still work?
I doubt that people will get enthusiastic about the new format. Imagine that Tim Berners-Lee suddenly decides that http://www.cnn.com/ was imperfect and should be replaced by interWeb:\\\64.236.24.12\ or something else. I guess that the world wouldn't be too excited.