The optical jukebox emerges as the economic choice for storing huge quantities of data objects in today''s electronic libraries. Sequential access to collections in an electronic library with a very large number of holdings significantly limits library response when requests are frequent, as found in large-scale library systems. Therefore, we consider an alternative scheduling method which takes advantage of specific user and application characteristics found in a library environment. In this paper, we address the performance benefits of cyclic scheduling of user requests in electronic libraries where retrievals are the primary tasks. An approximate model of the retrieval performance is developed in terms of average waiting times for retrieval of objects from an optical jukebox using a cyclic service strategy. This model considers variability in service and its evaluation involves an iterative approach. In addition, the goal of the approximation is motivated by providing a user-interface prompt which indicates the expected waiting times for requested documents. (not available electronically).

SUMMARY Intravenous heparin isfrequently usedtotreatthromboembolic disease, butthe consequencesofstopping heparin havenotbeenstudied systematically. Todetermine whether discontinuing heparin posesa clinical risk, we examined thecharts of378patients treated with heparin fortransient ischaemic attack (TIA), reversible ischaemic neurological deficit, or ischaemic stroke fromOctober 1979 toJune1985. Clinical deterioration, ora new TIAorstroke was more likely (p= 0O01) during the24hours after heparin was stopped inpatients notalready on aspirin or warfarin (10/143, 7%)thaninpatients receiving aspirin or warfarin before heparin withdrawal (3/215, 1%).Stopping heparin inpatients notreceiving aspirin orwarfarin appears toexpose themto anincreased risk forTIA,stroke, orclinical deterioration. Heparinisfrequently usedto treatischaemic cerebrovascular disease. Thecomnplications ofheparin therapy arewell described andinclude haemorrhage, thrombocytopenia, whichisrarely associated with arterial thrombosis, andosteoporosis. Theconsequencesofterminating heparin therapy arenotwell known. Experimental studies havedemonstrated thrombosis after heparin withdrawal. Although anintravenous heparin bolus inhibited thrombus formation inthe extracorporeal circulation ofpigs, compared to untreated controls, thetendency forthrombosis was enhanced 4-5hours later.' Similarly Ashwin2 found thatrats hadahigher incidence ofjugular venous thrombosis fourhours after anintravenous heparin injection thanafterintravenous saline. These experimental findings seemtohaveclinical relevance Alvarez Sabin etal3described apatient with infarction oftheinternal capsule whosubsequently developed a cerebellar infarct three daysafter stopping heparin. Systemic embolisation after inadvertent interruption ofheparin treatment hasbeenreported,4 andwe recently observed several patients onheparin for transient ischaemic attacks (TIA)whodeveloped