Michael;

I quite agree.

The problem is that those who could do it, don't.

Canie was gracious enough to make all the archives here accessable but the task of copying all that information and reformatting it is overwhelming.

I've been trying it with Shan's predictions and even though he used the prediction input form there's still enough variation to make it difficult for a program to read.

So by all means, if you have the time and energy to create and maintain a prediction registry with flexible output routines, go for it.
My past objections have been based on my experience with predictors who really don't want to be pinned down. The great Japanese debate for a recent example.

Roger