David Weinberger has an interesting post, Busted, which links to Michael O'Connor Clarke's Paging Hagbard Celine... .

Michael's post talks about what looks like some organization using a text generator to create websites that show up in Google and can be used to influence the ranking of search results. Michael links to the Markov Babbler, by Geoffrey Plitt, as an example of the kind of text generator that could be being used.

Besides the things about David's and Michael's posts that I found interesting from the social, political and legal perspectives, I had fun playing with Geoffrey's Markov Babbler. Here is some of the totally insane, though almost meaningful, text it spit out, using some of my post on RSS = Readily Significant Schizophrenia as input:

And, and site. In particular, I think a site.

And, and site. In particular, I think a site. But, say, this Echo Project effort to work for the subsequent comments from Dave Winer on the current RSS is designed for the data and RSS where the link element can function as if it happens to work for that in some ways.

And, and RSS, the vagueness in looking at everything. In particular, though they shouldn't be complimentary: content use is designed for functional content use RSS data model affects the same thing. And, is like a schizophrenia.

In other, the vagueness in looking at everything. RSS is designed for the same thing. People use is like a good next step might overlap but aren't the application modeling respectively. But, though they shouldn't be complimentary: content use is designed for the metadata and vice versa.

So, though they aren't blended, and metadata / data are up against each using an appropriate methodology, say, and RSS is designed for the metadata / data and it were in looking at everything. People use and it happens to improve each using an appropriate methodology, is designed for functional content use and vice versa. So, and functional content uses.

But, and application model, this Echo Project effort isn't off track in the metadata or for that in the vagueness between pure metadata of syndication might be complimentary: content uses. And, data model, and vice versa.

In particular, is a site summaries, and site summaries, I think a good next step might be complimentary: content use RSS as either pure metadata and I think that in the metadata or for that site syndication might overlap but aren't the action of site syndication and RSS data modeling and application modeling respectively.

Sam's posts and it happens to resolve these altogether are up against each using an effort to work for that in the same thing. So, RSS as RSS is like a good next step might overlap but aren't blended, RSS data and it happens to syndicate a mechanism to work for functional content uses. In other, and then look to work for the same thing.

In other words, this Echo Project effort isn't off track in the action of syndication might be complimentary: content use is designed for that in the application model, the subsequent comments reference comments from Dave Winer on the application model affects the subsequent comments from Dave Winer on the metadata and RSS as if it happens to syndicate a mechanism to first make sure they would ideally be blended, and RSS is a site summaries and content use, I think that in looking at everything.

********

A couple notes on how this was generated and what is displayed: I experimented a little with the Markov Babbler to see what it did with things. And, I found that it separates things around periods, for example 2.0 might get babbled as 2 and 0 separately. So, I edited my input taking RSS 2.0 and RSS 1.0 and changing it to just RSS. I also removed parentheticals and abbreviations like i.e.

Finally, the output above was originally in the form of a single long paragraph, which I excerpted and broke up into smaller paragraphs for readability and emphasis. However, no sentences were edited.