Introduction

This year, it is 20 years that the initial design of EMME/2 was decided upon. Since that time, EMME/2 has continuously been enhanced and improved, adding many new features and modeling capabilities. However, the basic design described in the initial publications on EMME/2 [1,3] is essentially still valid today. On the one hand, this proves the importance of building on a solid conceptual base - without it, EMME/2 would never have gone this far and kept growing all the time. On the other hand, times have changed a lot since the early eighties, and interactive computing is a totally different thing today than it was then. While the ``age'' of the basic concepts behind EMME/2 was never felt as a hindrance to enhancing the software with new modeling features, the ``age'' of the graphic subsystem of EMME/2 is evident, both to the users who have to operate the system, and to us developers who are seriously limited by the outdated technology assumptions on which EMME/2's graphic subset is based.

While it was possible to improve the graphic subsystem of EMME/2 in many ways over the past 20 years, these enhancements had always to be done in a way which did not affect compatibility with older versions. While this always was (and still is) an important criterion for EMME/2 developments, over the past years it became clear that the need for a new graphical interface cannot be satisfied by gradual and upward compatible enhancements of the current EMME/2 system, but that only a completely new and radically different design would give us the necessary leeway to satisfy the interactive graphic expectations of the EMME/2 users now and for the years to come.

This cannot and does not mean that the traditional EMME/2 software is to be discontinued and its functionality replaced by a new and incompatible piece of software. Rather, we foresee both the traditional EMME/2 and the new software based on the new concepts will have to coexist for quite some time. This is not only necessary for the continuity of the large existing EMME/2 application base, but the huge fundus of tools and modeling capabilities that EMME/2 offers is too valuable to just throw away and start from scratch.

Thus, the project Enif, which is presented for the first time in this paper, is not aimed at replacing the traditional EMME/2 modules in the foreseeable future. But it is aimed at providing a new way of interfacing graphically with EMME/2 data within a reasonable short time frame and it is designed to be solid enough to be able to grow into the future and eventually provide a functionality equivalent (but not the same!) as does EMME/2 currently.

In this paper the concepts behind Enif are explained and they are illustrated using examples from the current implementation, which can be used to produce graphic and list output based on EMME/2 data banks. The emphasis is put on making the reader understand the foundation on which Enif is built. The user interface itself, i.e. the details of the interaction between the user and Enif are not dealt with explicitly, since a) the current implementation is not final and the external aspects may still change and b) this paper is not intended to be a reference manual.