Luca Dan Șerbănați at the baccalaureate, 1961
Luca Dan Șerbănați in the early 1970s
Luca Dan Șerbănați giving his last fifth-year lecture, April 1989

Luca Dan Șerbănați

Emeritus Professor at Politehnica University of Bucharest

Research, teaching, industry and memoirs

RO | EN | IT
Luca Dan Șerbănați in Venice, 1990
Luca Dan Șerbănați in New York, 2005
Luca Dan Șerbănați

1. The interform concept

From intermediate form to design object

Initially, the idea was relatively simple: during design, a program does not appear all at once, but passes through a succession of intermediate forms. These forms must be identified, preserved, compared, and transformed. Gradually, however, I understood that such an intermediate form could not be reduced to a single representation. It contained several aspects of the same software reality.

Thus, the term interform came to denote not merely an intermediate state, but a richer conceptual unit: an evolving design object carrying structure, meaning, relationships, and history.

Facets of the same software reality

The idea of facets was essential. A software product is a complex reality that cannot be exhausted by a single description. The same interform may be viewed as text, structure, image, a network of relationships, a semantic object, or a moment in a refinement process.

Each facet provides a distinct view of the same reality. The textual view supports writing and modifying the description; the structural view supports control of internal relationships; the graphical view can make the organisation intelligible; and the semantic view enables verification, evaluation, and transformation. None is sufficient by itself, but together they form the operational image of the software product under construction.

Levels of representation

An interform could be represented through:

These levels were not merely alternative presentations of the same text. They enabled different operations on the software product: editing, navigation, verification, transformation, comparison, symbolic execution, or code generation.

Synchronic and diachronic relationships

An interform also had to be understood through its relationships. Synchronically, it was linked to other interforms through inclusion, invocation, use, composition, specialisation, or cooperation. Diachronically, it was linked to its own versions: previous forms, refinements, variants, corrections, and abandoned alternatives.

Design thus became not merely the writing of a sequence of texts, but the management of a dynamic network of intermediate forms, each with its own structure, facets, and history.

The interform as an active object

In my conception, an interform was not a passive document. It had to be inspectable, transformable, verifiable, and relatable to other interforms. A design environment should therefore manipulate not merely files, but design objects endowed with structure and meaning.

This idea opened the way to functions that became concrete components of the INTERFORM project: syntactic and semantic editing, symbol tables, symbolic execution, version management, design-decision history, and collaborative support.

The interform and the abstract machine

The interform concept was related to that of the abstract machine, but it was not identical to it. The abstract machine denoted the conceptual level of execution or imagined behaviour; the interform was the manipulable representation of a design stage, capable of expressing that level as well as its relationships with other levels and facets of the product.

2. From concept to the INTERFORM project

After years of conceptual modelling of software design, it became clear to me that a genuine support system had to assist the entire design activity, not merely final coding. By the mid-1980s, compilers, text editors, and isolated tools existed, but there was no environment that organised the development process according to coherent principles: preserving design decisions, controlling versions, performing checks and symbolic execution, and supporting collaboration.

INTERFORM was my response to this problem. It was based on the idea that a software product under construction should be represented by manipulable design objects that could be edited, analysed, transformed, and validated. The system was intended as a design workshop, not merely a program editor.

3. System architecture and functions

Layered architecture

The system architecture was conceived in layers with distinct roles, from manipulation of language structures to project organisation and collaboration:

Essential functions

State-oriented access

INTERFORM was conceived as a workshop in which the user worked with concepts and relationships, not merely text. Interaction was organised into states, each with its own command set:

System components

Internally, the system included a dialogue moderator and a model processor operating directly on interforms. It integrated editing, checking, symbolic execution, expansion, and code-generation functions, together with auxiliary components for management, windows, macrocommands, and interactive support.

4. Implementation and contracts

The implementation was carried out on a PDP-11/40, under the technical conditions available at the Polytechnic at that time. The system evolved from support for design and coding towards a broader view of the software-product life cycle. Seen today, INTERFORM was more a prototype of a conceptual workshop than a finished industrial product, but this very ambition made it important in my work.

Contracts that marked its construction

5. What the concept anticipated

In retrospect, the interform anticipated several directions that later became important in software engineering: integrated development environments, rich intermediate representations, multi-view modelling, model-to-model transformations, code generation, preservation of design rationale, and model-driven engineering.

I do not make this observation in order to rewrite the history of the field, but to clarify the concept’s place in my own path. The interform was an attempt to make visible and manipulable the most fragile area of software construction: the space between idea and program, between specification and code.

6. Connection with the articles

The INTERFORM project concentrated and continued the ideas formulated in my articles of the 1980s on systematic program design, CAD/CAM systems for software, and the interform as the central object of the development process.