summary Despite the adoption of Arabidopsis thaliana as a model plant system and the plethora of molecular information being obtained from its use, it is disappointing that the scientific community has not devised a cell signalling model integrating and visualizing these data. Lack of common systems of nomenclature and the sheer size and complexity of the task inhibit any individual from bringing together the knowledge into a unified structure. There are clearly many aspects of cell biology that are similar, even between plants and animals, that could facilitate development of a generic model. A gene-coding or nucleotide classification system which is 'user-friendly' would be beneficial to building such a model and enable rapid identification of orthologues of genes from different organisms. Whilst some international projects seek to address the problem of assigning unique numbers to genes, none suggest a nucleotide classification system that provides biological information that is transparent within the code. This paper discusses these issues and identifies the need for a more formal, semi-automated approach to modelling signal transduction utilizing the strengths of the proposed classification approach. By way of illustration, an example of a possible nucleotide function code is suggested, to demonstrate more clearly the benefits of such a system. Further discussion of this topic will be encouraged on websites (<www.bspp.org.uk> and <www.drastic.org.uk>).