This thesis introduces a new model which has the ability to compress an XML document efficiently and retrieve information from the compressed file according to vague queries and even various other types of queries. This chapter will outline the main conclusions of the research as well as the main advantages and limitations of the designed model. Finally, the chapter will also list possible future trends in this research in terms of developing the proposed model.
Top1.1 Conclusion
As the importance of XML usage for storing and transferring data via the World Wide Web becomes increasingly clear, there is a corresponding need to compress the size of XML documents, dealing with them in their compressed mode so as to make them accessible to devices with limited resources. When these compressed documents are used by simple users, in a situation where there is absence of schema, or if such a user has no exact idea of what s/he is looking for, there should be a special technique available to adequately deal with these types of queries. The questions had been raised by this research and their answers are as follows:
The answer to this question is XCVQ compressor. The design of the model showed the best average compression ratio (78.45) among the other XML queriable compressors without the need to the XML schema to be available. This was due to several reasons, such as: (1) limiting the storage of each element and attribute name in the document to only one number, which represents the order of that element or attribute in the XML document, instead of being two numbers, and (2) increase the granularity of the data to be compressed in order to perform better compression ratio. Although this design issue increased the compression ratio, but it affects the time required to compress the document by increasing this time to be higher than the time require to compress these documents using other techniques. However, the compression process usually made only once, while the querying process can be done hundreds of times to retrieve information from the compressed files.
To answer this question, XCVQ Structure Compressor was designed. In the compression process of the XML documents, the research found the strong affect of the redundancy in the structure of the document on its overall size. By succinctly storing the structure part of the XML document and keeping the data part as it is, the experiments showed good compression ratios which were up to 85.43 and averaged 49.47 for the tested XML corpus. This shows the big redundancy in the structure part of the document, apart which is considered to be very important for several purposes and retrieving information is one of them.
Vague queries are one of the important types of queries. They occur in different situations and require special ways to be processed since the existing query languages do not have the ability to answer these queries. The XCVQ-QP can deal with simple and complex queries by forcing each query to pass by two decomposition stages in order to make it easier to retrieve information from the relevant document(s) and then combine the sub-results to be decompressed and submitted to the user. This process required the expansion of XPath query language in different sides: the path expansion, the data value expansion, and the set of functions expansion. The time required to process the queries are very competitive especially when dealing with structure-based queries, since the compressed structure of the document helps in accelerating the retrieving process.