John, you wrote:

> 'Peer Review' does not, in itself, necessarily result in any sort of > 'standardisation'. On the contrary, those 'peers' may find themselves > reviewing a totally different thing each time.

I am referring to the reviewing process, which is in fact standardized. By the way, any published paper is based on subjective reasoning, yet the output should be 'standard quality' result. Let's try to go a step further. I hope we could see beyond SAS. If this discussion was only to say: "No future without SAS", OK, I give up. In the opposite case, I'd like to ask:

"WHO TESTED PROC GENMOD ?"

Clinical trials based upon longitudinal repeated measures are common. "WHO CONTROLS THE WAY IT HANDLES GEE MODELS?". I found that results were different from existing (high quality!) software in C, and later SI made patches available.

There might be extension to highly advanced technologies, such as bayesian models or neural networks. Who will test the algorithms?

S+ programs, are these not good? What about extensive SAS macros as a way to develop new procedures? And C++ programs ?

There is always an open part and a closed one, otherwise one had to develop only in assembler ! But, don't you believe that SAS is becoming too much close-minded ? ....maybe the Institutes and CRO should buy a standardised SAS/FDA, no problem with it. ..but I would be interested in knowing more about the hidden part.

Cheers Fabrizio