The infeasibility of quantifying the reliability of life-criticalreal-time software

This work affirms that the quantification of life-critical software reliability is infeasible using statistical methods, whether these methods are applied to standard software or fault-tolerant software. The classical methods of estimating reliability are shown to lead to exorbitant amounts of testing when applied to life-critical software. Reliability growth models are examined and also shown to be incapable of overcoming the need for excessive amounts of testing. The key assumption of software fault tolerance-separately programmed versions fail independently-is shown to be problematic. This assumption cannot be justified by experimentation in the ultrareliability region, and subjective arguments in its favor are not sufficiently strong to justify it as an axiom. Also, the implications of the recent multiversion software experiments support this affirmation     

Iron supplementation enhances response to high doses of recombinant human erythropoietin in preterm infants

AIMS: To determine whether iron supplementation would enhance erythropoiesis in preterm infants treated with high doses of human recombinant erythropoietin (r-HuEPO). METHODS: Sixty three preterm infants were randomly allocated at birth to one of three groups to receive: r-HuEPO alone, 1200 IU/kg/week (EPO); or r-HuEPO and iron, 1200 IU/kg/week of r-HuEPO plus 20 mg/kg/week of intravenous iron (EPO + iron); or to serve as controls. All three groups received blood transfusions according to uniform guidelines. RESULTS: Infants in the EPO + iron group needed fewer transfusions than controls--mean (95% CI) 1.0 (0.28-1.18) vs 2.9 (1.84-3.88) and received lower volumes of blood--mean (95% CI) 16.7 (4.9-28.6) vs 44.4 (29.0-59.7) ml/kg. The EPO group also needed lower volumes of blood than the controls--mean (95% CI) 20.1 (6.2-34.2) vs 44.4 (29.0-59.7) ml/kg, but the same number of transfusions, 1.3 (0.54-2.06) vs 2.9 (1.84-3.88). Reticulocyte and haematocrit values from postnatal weeks 5 to 8 were higher in the EPO + iron than in the EPO group, and both groups had higher values than the controls. Mean (SEM) plasma ferritin was lower in the EPO group-65 (55) micrograms/l than in the EPO + iron group 780 (182) micrograms/l, and 561 (228) micrograms/l in the control infants. CONCLUSIONS: Early administration of high doses of r-HuEPO with iron supplements significantly reduced the need for blood transfusion. Intravenous iron (20 mg/kg/week in conjunction with r-HuEPO yielded a higher reticulocyte count and haematocrit concentration after the forth week of life than r-HuEPO alone. Infants treated with r-HuEPO alone showed signs of reduced iron stores.     

Cation distribution in oxide fuels — a neutron scattering study

We use diffuse neutron scattering techniques to investigate cation ordering in the mixed oxide U/CeO_2?x — a model system for the oxide U/PuO_2?x. Our results show that there is little deviation from a random cation distribution at 10 or 20 mol% dopant concentration. Significant short-range order appears, however, to be present at 30 mol%. Experiments on samples prepared under reducing conditions suggest that vacancies have little effect on the cation short-range order, and thus cast doubt on the validity of models which assume dopant aggregation around vacancies.     

Growth rate of malignant tumors of the large intestine

The rates of colonic tumor growth were studied on the basis of evaluation of the case histories of 19 cancer patients in whom dynamic endocolonoscopy had been carried out. In some cases, malignant tumors appeared in the colon within 12 months after previous colonoscopy. Tumors were under 2.0 cm in size and were not accompanied by any clinical symptoms. Repeated endoscopic examinations performed within a period of 1-3 years detected large tumors which caused the constriction of colonic lumen. Clinical symptoms developed in such patients approximately 1-4 months prior to tumor detection. It is concluded that although the rates of growth of colonic malignancies are rather high, tumorigenesis is preceded by a long-term stage of preclinical disease.     

Identification and localization in tachistoscopic recognition: The effects of data- and resource-limitations.

Examined the effects of data- and resource-limits on mislocation and intrusion errors in a visual probe task. Mislocation errors were defined as responses to any letter that appeared in the test display but not in the position cued; intrusions were responses with letters that did not appear in the test display. Ss were 42 undergraduates with normal vision. Exp I manipulated the data-limits by changing both luminance and duration, and confirmed that decreasing stimulus energy increases intrusions and decreases the benefit of selective attention instructions. Exps II and III manipulated resource-limits by varying display size and the size of a defined subset of a display and showed that this manipulation primarily affected mislocation errors. Exp IV tested a minor point pertaining to the cue used in the previous studies. Results suggest that these errors may reflect 2 different processing operations, and that complete recognition of any visual stimulus involves both processes. (French abstract) (38 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)