4.3 Conclusions

Due to the lack of comprehensive metadata, and incorrect data representation, our work was difficult to commence. Creating and developing a spatial data dictionary at SFU demanded new data standards based on a solid understanding of the nature of the underlying data. After examining what geo-spatial datasets, that have metadata, are available at SFU, we decided to focus on a small number of datasets to keep our study concise. We were all well versed in regards to the general lack of metadata at SFU, and the need for metadata standardization and ontology-based information integration. We attempted to address these issues by creating a metadata scheme based on ISO that is both compact and appropriate for SFU.  Unfortunately, we were unable to implement the schema with ISO elements, thus were required use parallel FGDC tags that contained the same, or similar data.  With respect to the web-based user interface, we attempted to implement a modified version of ESRI’s Metadata Explorer, but this was hampered by serious technical constraints due to bugs being present in the software. Our project also incorporated the conflation of the issues of semantics, ontologies and interoperability in metadata, and used XML to address them.

Even though the importance of standardization of geographical data is well known, the reality is that most geo-spatial datasets have incomplete metadata, and in many cases, none whatsoever. We deeply caution that the effects of accuracy and quality of the metadata we create for geo-spatial datasets at SFU will impact the users, for the metadata could be displaying incorrect information, and mislead the research. With this concern about of the importance of the accuracy of the metadata, we started to experiment with major standards among many information standards, such as ISO, FGDC, and the CGDI, which is dominant data standard in Canada. Since Metadata Explorer appears to be built primarily (and solely) around FGDC, we decide to use with FGDC standard to prevent the future maintenance problem. We outlined the advantages of FGDC and ISO and stated that it can help the automatic data exploring process between data source.  Despite the presence of software bugs, issues regarding the structure of ISO tag in XML files caused much grief in our work.  More work needs to be undertaken to ensure that the ISO standard is implemented in an XML format that is easy to decipher which can be apply in the real world.

The other big challenge for us was to build semantic and ontology-based information integration. Standardization of the metadata using FGDC requires a clear understanding of the concepts of systematic semantics and ontologies.  Uneducated attempts to apply formal semantics to keywords describing the data would create confusion in the meaning of words to both users and the computer. We identified this as a major reason why data searches often fail to find relevant data, and why there are difficulties in exchanging data. Many papers demonstrate that to solve this typical problem, most of them emphasize the application of a web of integrated ontologies in the search engine and metadata structures as a solution because ontologies support the translation and comprehension processes. This is very important because the ability to include contextual meanings supports projects such as a semantic search engine, or a spatial data dictionary.  The main approach to dealing with these problems is to use the XML language to store metadata, and implement hierarchies of classes of tags within the XML metadata storage template.  Complex query systems and simple query scenarios require another domain ontology as we mentioned above. Integration of other data sources through the automatic translation process makes possible the semantic inference search engine and it is quite effective to fulfill the goals of our work. We believe that it may provide better ways to communicate between users and computers and reduce the problems of incompatibilities in data format, software products, and data standards.

Lastly, we propose creating a semantic-search engine with advanced applications of XML tags. XML’s primary utility as a meta-language goes hand in hand with its primary application in the geo-spatial community. As Badard at al declared “The primary use of XML in the context of a spatial data dictionary is for the delivery of updating information,” we noted that we could address the data maintenance problem for the future with XML. It makes new data standardizing of metadata possible, and reduces the gap on interoperability issues pertaining to data sharing across the campus. Since one of our main study issues was that non-interoperability severely limits the use of digital geographic information, XML tags, ranging upwards of 100 different tags in this case, cover the all the possibility of the data processing procedure. In addition, with the combination of programming languages, the application of metadata based in XML is the future for the geo-spatial community. 

Further work requires new architectures of information systems, and demands that the developer have philosophical thought-processes, for dealing with ontologies and semantics in a rigid programming environment is not easy! Furthermore, future studies should research more about synonyms, hyponyms and hypernyms in terms of sharable and reusable knowledge to help move the geo-spatial data community to being object-oriented. In overcoming the ontological barriers to data descriptions, and semantic barriers to search parameters, this study is only the first step.  Likewise, we recommend a much more advanced spatial data dictionary be designed and implemented for general use at SFU, mainly because our current product does not meet the recommendations made in this paper.  Ultimately, regardless of complexity, all problems have a solution, and the issues addressed in this paper are nothing short of complex.  Paradoxes arise, issues with software constraints cause problems; however, a viable solution can prevail in any problem, if and only if, enough careful thought and devotion is invested.