(I). Evaluation of input data, adjustment procedures

This paper is the first of two parts presenting the result of a new evaluation of atomic masses (AME2003). In this first part we give full information on the used and rejected input data and on the procedures used in deriving the tables in the second part. We first describe the philosophy and procedures used in selecting nuclear-reaction, decay, and mass spectrometric results as input values in a least-squares evaluation of best values for atomic masses. The calculation procedures and particularities of the A ME are then described. All accepted data, and rejected ones with a reported precision still of interest, are presented in a table and compared there with the adjusted values. The differences with the earlier evaluation are briefly discussed and information is given of interest for the users of this A ME. The second paper for the AME2003, last in this issue, gives a table of atomic masses, tables and graphs of derived quantities, and the list of references used in both this evaluation and the NUBASE2003 table (first paper in this issue). AMDC: http://csnwww.in2p3.fr/AMDC/