Despite the fact that dynamic XML labelling\udschemes have been investigated widely, some challenges still need\udto be tackled. Dynamic XML documents are subject to change. An\udefficient dynamic labelling scheme is able to maintain the node\udrelationships throughout continuous changes to the XML tree\udstructure. Such a scheme generates labels for new nodes to avoid\udthe need to relabel the whole tree. The main problem for dynamic\udXML is overflow that occurs when the label length of the new\udnode is over the reserved space limit. There has not been sufficient\udanalysis to determine the class of labelling scheme which faces this\udproblem in the early stages of update. To this end a series of\udexperiments were performed when updating the Nasa XML\uddatabase, which contains real data. Five sets of new nodes (50, 100,\ud400, 800, 1200) were inserted into this dataset using two versions\udof XML node indexing system: a Prefix and an Interval labelling\udscheme. It was found that Interval falls victim to the problem of\udoverflow after the insertion of only 100 nodes whereas Prefix has\udno problem even when adding 1200 nodes.
展开▼