This study investigated schoolchildren's posture and discomfort while working at computers. Sixty-eight children (mean age 9.5 years) were observed at school during normal computer sessions lasting 15-25 min. Rapid upper limb assessment (RULA) was used to evaluate posture, and a body discomfort chart (BDC) and a modified visual analogue scale (VAS) were used to record site and intensity of discomfort. Computer tasks were noted and in accordance with RULA, postures were classified as Action Level (AL) 1 (acceptable) to 4 (needs immediate change). Most children adopted postures at an unacceptable level while working at computers. None of the postures were in AL 1; 60% were in AL 2; 38% were in AL 3; and 2% were in AL 4. Posture became worse over time. Poor posture was associated with discomfort, but it is not clear if it was related to the sitting posture or to the computer use. Children who reported discomfort had a higher mean RULA grand score (5.0) than those who did not report discomfort (4.4). The type of computer task influenced the children's posture. RULA proved generally to be a suitable method for evaluating children's posture.