aron@sri-unix (12/07/82)
This message is an update for those of you who followed my saga re: S and statistical routines. We have definitely decided to go the route of integrating S and BMDP. We have source of both packages on site (we are an educational institution) and we are in the process of deciding what statistics we need, and in which package they currently reside. When our system is totally in place it will work somewhat as follows: Patients will be entered into the database when they are scheduled for appointments. Results of all tests and procedures are entered screen forms into the database. A generalized functional query language will be used to extract subsets of data from the database. This data will then be stored as an S dataset. S (with integrated BMDP statistics) will be used for analysis and graphics. The new troff (with PIC) will be used to write abstracts and research papers. The database, by the way, is a home-grown system (essentially multi-key indeces) written in C. How's that for a unix based research environment? I will let you all know how things progress. The two missing pieces now are the query language and integrated S/BMDP. aron shtull-trauring harpo!seismo!presby!aron