The topic of the paper is a methodology and models for interactive knowledge creation processes. It is assumed that a knowledge system is being evolved by a team of collaborating experts that creates new evaluation indicators for research and their meanings. We proceed from the notion of a goal-oriented knowledge system. The experts team creates this system to fill a knowledge gap in an evaluation area. The gap can be identified through observation of the subject area. We define a goal-oriented knowledge system as an expert-created system of concepts generated for filling the knowledge gap. The main emphasis is placed on a methodology and models for new knowledge creation by collaborating experts. We have studied relationships between emerging indicators' meanings and their volatile denotata as new knowledge sources. In our study denotata are represented by indicators data together with computer programs for indicators calculation. According to the methodology, experts using computer programs can interactively fix their own emerging indicators' meanings. Emerging meanings are temporal and changeable concepts opposite to conventional concepts that have time-stable signifiers. The goal-oriented experts' cognition and interaction are the basic mechanisms for the new indicators' meanings creation. The basic conditions for their creation and propagation are a computer dictionary, which is a tool for new meanings propagation between collaborating experts, as well as information and communication technologies. According to the methodology, experts can specify time-dependent states of new indicators in the computer dictionary. Each state of an indicator is described by its digital denotatum, a meaning definition, and a name as an indicator signifier, which are used by experts to describe new knowledge by dictionary descriptors. In order to show feasibility of proposed methodology, we are designing the computer dictionary as a tool for indicator development. The dictionary is a part of the evaluation system prototype for verified monitoring and evaluating implementation of a research programme. Our dictionary design is based on two semiotic models for a description of indicators development stages, including generation processes of expert knowledge about developed indicators by collaborating experts. The computer dictionary based on these models of knowledge creation has shown new possibilities for developing indicators for evaluating research programmes. Collaborating goal-oriented experts could interactively fix their own emerging indicators' meanings, specify the states of developed indicators as dictionary descriptors, and harmonize emerging indicators with research programme's goals using information and communication technologies. [ABSTRACT FROM AUTHOR]