Namely, he claims that unless a field employs 'hard science predictive theories' and a 'falsify and replace' analysis of said theories, then those in the field are not engaging in science.
He has further explained that 'hard science predictive theories' are not meant to actually explain anything, and requiring them to do so indicates an ignorance of scioence and the scientific method.
What are 'predictive theories' and who uses them? Well, exactly the folks in Bergerson's supposed area of expertise, and that is basically it.
Computer modeling/programming employs 'predictive theories', but these are not 'theories' in the true scientific sense. From what I have gathered, 'predictive theories' in computer science refer to what amounts to an application of previous experience to a proposed new project. That is, one employing a 'predictive theory' would think along these lines: "When this line of code was changed on Program X, 'A' occurred, therefore, if I make a similar change in this line of code in Program Y, something like 'A' should also occur." ( Note - predictive theories are also employed in areas like physics and chemistry, where the ineractions between componants of a system are well defined and controllable, and the 'thgeories' look an awful lot like calculations, not statements. This alone should tell the intelligent person that such theories are at best of limited value in dynamic living biological systems.)
Anyway, a new thread at ARN, started by Bergerson, clearly indicates this transferrence that he engages in. It also shows how he is quite unable to produce anything of substance, for he titled the thread "Examples of IDT's [sic]" and two dayts later, he has provided none. Of course, he implies that he has - in Bergerson's world, alluding to something, hinting at something, or writing about something is the exact same thing as presenting something.
Bergerson gives awa the farm, however, in this post on page two:
Again, there are lots of existing examples of scientific analysis performed using IDTs or the logical equivalent of IDTs. Probably some of the best examples involve the computer programs and the design of computer programs.
When I sit down to design a computer program, I have a goal, I have at least an initial understanding of the constraints under which the program is expected to operate, I have an initial understanding of the output R the program is to produce, and I want to design a program or algorithms of the general type F(S,G) that produces output R. In other words, I want to find a predictive IDT of the type:"Under defined constraints, F(G,S) predicts or determines R"
How did he give away the farm? Well, he states clearly: where he is coming from (computer programming), and what his expectations for a 'predictive theory' are. These are amenable to biological reality (for the most part).
Well, read his writings to see for yourself.