I'm planning/hoping to do three articles after the 3NF one: one on normal forms intermediate between 3NF and 4NF and how the representation principle may get violated somewhere in that space (which is perhaps the most important article in the series, because understanding the representation principle in its various forms is the core concept of normalisation), one on 4NF and one on 5NF which will probably include a definition of 6NF but certainly no examples of 6NF (that will be Date's 6NF, not Fagin's 6NF - Fagin's 6NF aka DKNF may get a mention in the representation principle article, or not, depending on how I work the article out).

The article describing the algorithm was published in ACM ToDS 1/4 (December 1976) beginning at page 277, the author was P.A.Bernstein and it was titled "Synthesizing Third Normal Form relations from functional dependencies"; I don't know of any way of getting it without paying the ACM (they deserved paymenty for the original publication, but given that they are a learned society and that ToDSis supposed to be a genuine technical joyrnal I think that this long afterwards http://www.gnu.org/philosophy/right-to-read.html is relevant). A related article is "A New Normal Form for the Design

of Relational Database Schemata" by Carlo Zaniolo, in ACM ToDS 7/3 (Sept 1982), which is available without charge somewhere on the web (I'll not say where, in case the US copyright mafia sue me - stating facts is now a serious act entailing punitive damages in American law, and since it is also classified in the USA as a criminal offence and we have a crazy extradition treaty with the US the lunatics in charge of the US Asylum can extradite Brits from Britain for offenses committed in Europe against US law which are perfectly legal acts in Europe).