1. Prefix notation shortens the learning curve
2. The only delimiters used in the language are parentheses and commas, which makes the syntax quite easy to deal with. The only potential problem may be more escaping of these three characters than "usually".
3. Only one language for both defining the database (DDL) and maintaining/using the database (DML) is another feature that speeds up learning. I like this a lot.

SI

This is a fantastic tool. My two favorites are:

1. Database enforced constraints
2. Unique attribute names in the entire database, not only in each relation

Having the ability to enforce constraints of arbitrary complexity by the RDBMS engine is priceless. Also, the two shortcuts for constraint specification (key and tuple) save a great deal of typing. Granted, this should be a part of any RDBMS engine but this was the first time I had an opportunity to try it and it's great.

Defining attributes independently of relations (relvars to be precise) is a great design enforcing mechanism. Not that any design should be enforced, it should be a major part of any IT project, but we all know how it goes. With this one and the separation of logical and physical specifications, one can build his/her own "columnar" DBMS (if such a thing exists) using nothing but a good "old" RDBMS engine.

There are two features I miss:

1. Support for quota queries
2. Support of IF/CASE statements in scalar expressions; computational completeness is not a bad idea, but I personally would not mind arranging my statements in SEQUENCES ONLY in a client.

If I can offer a couple of advices to any new users of SIRA_PRISE it would be:

1. Define your coding standards immediately and, if that's not enough to make you comfortable with prefix notation, write your expressions in infix notation first and then translate them into prefix.
2. Have a good text editor with, at least, these features: parentheses matching command and regular expression search/replace.

SI