Cerebral palsy (CP) research has broadened its horizons in the past decades. In addition to the long-term, population-based studies of prevalence of the different CP types, gestational-age groups, and accompanying impairments, data on several aspects of function have emerged. The well-accepted Gross Motor Function Classification System1 has given us a common language to compare gross motor function in different populations and at various ages. Gross motor classification has been followed by classifications of hand function, such as the Bimanual Fine Motor Function2 and the Manual Ability Classification System.3 However, differences in CP classification around the world still exist. This is well illustrated by Rice et al. in their report from the South Australian CP Register in the current issue of the journal. Although a method of classification was proposed along with the most recent definition of CP,4 the traditional systems tend to predominate. This makes comparison between CP registers difficult and important knowledge about trends, differences, and similarities may be overlooked or lost. An essential aspect of the diagnosis of CP is the neurological signs. It is now recommended that classification should be made according to the dominant type of tone or movement abnormality. This has not been the case before in all countries, and has made any attempt to pool data difficult. However, such comparisons have been made possible in Europe after the introduction of the Reference and Training Manual, produced by the Surveillance of Cerebral Palsy in Europe (SCPE). It contains descriptions and video illustrations of the different CP types. It is an excellent tool in the endeavor to harmonize classification of CP in Europe and elsewhere.5 The paper by Rice et al. is a commendable attempt to carry out a comparison between CP registers on two different continents, despite differences in classification systems. The fact that this is feasible is very promising, but the paper also depicts the difficulties of such a comparison. This is especially apparent when comparing the quadriplegia and dyskinetic groups in the Australian and Swedish registers. Although very similar groups are described the outcome regarding classification differed, because of the different traditions in the two countries on classifying dyskinetic symptoms. The prevalence of dyskinetic CP may be different from what we find in most studies today if, or when, classification according to the dominating symptom is fully accepted. As we strive to describe the symptoms and function of children with CP, we have to accept that there are not yet valid and reliable instruments for every measurement. When there is such an instrument is at hand, it should be used to compare studies, and also to make it possible for other researchers to relate to the data. The specific differentiation between dyskinetic and spastic features may indeed be a challenge, as will be, in some cases, deciding which is in fact the dominating symptom. Children with dyskinetic CP, classified according to dominating movement abnormality, have in a majority of cases also signs of spasticity.6 Very few instruments exist to describe dyskinesia. As dyskinetic movements may severely hamper motor function, there is a need for valid and reliable methods to assess this aspect of motor disability. The Barry-Albright Dystonia Scale, referred to by Rice et al., deals with the dystonic, but not the hyperkinetic symptoms.7 This scale is seldom used in CP research, but might be considered a useful tool.6 Also, in children with predominantly spastic CP there may be involuntary movements of a choreo-athetotic type that interfere with activities, and we have little means to describe this phenomenon in a structured way. Despite all this, Rice et al. take on the challenge to describe and discuss the dyskinetic features in children with CP and find this in almost every fifth child! CP research is carried out in various ways around the world. Many CP registers rely on reports from local physicians on what CP type, gross and fine motor function, and accompanying impairments a child might have, while others, such as the South Australian CP Register, has the advantage of a clinical examination in addition to the data from the records. Here is an opportunity to apply a variety of assessment instruments to ascertain information that is not possible to assess from records, but that is required to investigate the child. This may be one way to develop further population-based CP research. Future studies from the South Australian CP register will surely benefit from the access to clinical investigation using valid and reliable instruments to gather more population-based knowledge on CP.