Google doesn’t look at it that way, they don’t index and rank “sites”, they index and rank individual “pages”. So it generally doesn’t matter to Google that all of those pages are on one site or many. What matters is the diversity of links within the web of pages that your page is connected to.
So, actually the opposite of your assertion, “you are better off using the blog creation sites”, is true. Since community websites like Squidoo and Hubpages, typically fair far better than the disconnected blog sites, especially when there isn’t a community sharing situation. The fact that it is the same, or different domains, has no positive nor negative effect, except where you are competing on the same page for the same keyword against a page on the same domain. But for backlinks, this doesn’t make any direct difference.
I have never seen any credible evidence that suggests Ip diversity, nor domain diversity are factors. Backlink diversity yes, but it has absolutely nothing to do with IP addresses nor domains.
You do need diversity in backlinks sources, but it has nothing to do with a particular domain name. If your extended web contains backlinks from many sources, it doesn’t matter at all if they are all on the same domain or not. The limiting factor is the source of the backlinks on your web. It’s based on the random surfer model and is agnostic towards domain names.
If you were to get backlinks from 1000 different pages on a website and each of those pages, on that website got their links from as many different sources it works just as well as if they had been all from different domains.
It might be true when you are talking about a small website, or about pages that all share the same backlink sources, but the link diversity will usually be far greater when garnering links from larger websites like Squidoo. Your WordPress and blogger subdomains have virtually no link diversity, except what you go out and create on your own or through organically earned backlinks. There is no comparison in that regard. Link diversity beats lack of link diversity every time, at least when you are talking about allinanchor ranking.
That is my opinion, it has held true in all tests I have done. If you have any evidence that shows IP address or domain diversity as causation, rather than simply correlation, I’d like to hear about it.
Web 2.0 is a rather broad term that was coined to describe websites that allow the “users” of the website to contribute content. For example, this forum is considered a Web 2.0 technology when using that term in the broadest sense of it.
Since there are many types of website, like forums, blogs (via commenting), microblogs (Twitter), social networking (Facebook, Linkedin), Article syndication (Ezinearticles), social bookmarking (Digg, delicious, et al), etc., that all are Web 2.0 technologies we usually refer to them with more specific terms. However, there is a group of websites that, for the purpose of SEO, are typically referred to as Web 2.0 as there are many different labels and it is more convenient to simply group them into a group with the generic label of Web 2.0 properties.
Sites that are typically included in this generically labeled group are social community sites like Squidoo, Hub Pages, Myspace and Facebook. Also typically included are community blogging sites like WordPress.com, Posterous, Blogger, Xanga, Typepad, Tumblr, etc.