I usually like Wikipedia for the general accuracy of their definitions and more importantly, for the amount of them that are available. However, I find that their definition of
processor is vague at best, and I feel left with a feeling of misunderstanding after reading the article. I felt more knowledgeable after reading what AMD and Intel had to say on the subject (their articles were more technical but also much more precise).
One thing remains : whether there are two cores or two processors, they both share the same bus (at least current ones such as the Pentium D and the PowerPC970MP - aka G5 - and maybe the all-new Dual-core Xeon - interesting that Apple launch their new G5 within days of its release, it makes me wonder if it is coincidental or not - but I think that at least the next generation Xeon will have two separate buses, one for each core) which means that when buss access is required, a Dual-core chip will have twice less bandwidth than a Dual-processor, which could result in some performance decrease. Its advantage over the Dual-processor, however, is that with the "old" design, when the two processors needed to communicate, they were forced to do so via the FSB, whereas with the new, Dual-core design, they do not need to go through the bus to exchange data.
This definition is consistent with the performance increases that Apple is claiming.
Interestingly, Apple is claiming a performance increase from their Quad G5 over the previous Dual 2.7, not an increase from the Dual-core 2.0 and 2.3 over the Dual 2.0 and 2.3 G5. I did not see any comparison between these old ones and the new Dual-cores on their website, just a comparison between the Dual 2.7 and their Quad Core 2.5, which is obviously faster. I did not look at the updated PowerMac pages very long, though, so I could have missed something.