Starting from semantic GIS and Knowledge Expression

Semantic (semantic) and Its Semantic Web

First, let's talk about semantics. For personal understanding, semantics is meaningful information. For example, if we say "B-u-s" separately, it doesn't make any sense, but if we say "bus ", in English, it is a bus. In academic research, the semantics (semantic) was proposed as the "Semantic Web" (Berners-Lee, T ., hendler, J ., lassila, O ., 2001. the Semantic Web. scientific American, 28-37 .), maybe earlier.

What is Semantic Web? The so-called Semantic Web means that the existing HTML-based Web is based on Expression and visualization, and it is meaningless or difficult to extract meaningful information from it. On the contrary, XML is good because the self-descriptive XML can represent meaningful information. Therefore, the research on semantics is a technical perspective, that is, the research on XML application.

Origin and significance of semantic GIS

Back to the GIS discipline, the existing data models of the GIS discipline are actually based on the expression or visualization, which is determined by the basis of the discipline-cartography and computer graphics. Therefore, the existing data models, whether it is an object model or a domain model, it is not the information of the real world, but the visualized parameters of some of its properties (basic, not completely). Therefore, problems such as system interoperability and Data Metadata cannot be well solved. Therefore, the concept of semantic GIS is proposed.

Then, the relevant research and implementation are probably based on the relevant OGC specifications, especially the GML. The core concept of GML is feature, which can be translated into objects or features and can be customized to express knowledge in different fields. This object not only contains its attributes and capabilities, but also has spatial attributes. The criticism of GML is that it cannot express three-dimensional information and has no topological relationship (3.0 how to carefully query data ). GML 3.0 can already display Domain Information and temporal information, at least in terms of specification. You can also discuss whether the topology is a required attribute of space information.

Therefore, such standards and research for the GIS industry are also of great significance, such as data exchange, system interoperability, and even the standardization and modularization of the GIS platform mentioned above. But the premise is the fact standard, rather than some rules that are too academic. The GIS system and a good software architecture, why not and how? This issue is also mentioned in the article, which will be further discussed later.

In existing platforms, ESRI's feature is closest to the OGC specification, but ESRI's long-standing "anti-usability" makes these features or features basically unavailable in practice.

Contents that can be studied in semantic GIS

In my personal documents, foreign semantic GIS research basically focuses on GML, while GML is also an application of XML. Therefore, from a technical point of view, it is easy to think that these studies are meaningless and meaningless, but in fact, in the software industry, the main driving research should be "industry knowledge", and the academic name is "professional knowledge ".

I personally think that the following aspects need or should be studied:

○ Data expression, based on GML 3.0, studies its data expression methods in different fields and how to express existing professional knowledge;
○ Research on data storage technology related to GML and how to store it in relational databases;
○ Research on the Integration Technology of GML, how to seamlessly integrate with the existing data structure, that is, how to seamlessly convert the existing data into GML and switch back;
○ Research on GML-based spatial analysis;
○ Research on visual interaction technology of GML, such as SVG, especially in Web and network environments. SVG personally thinks that a key issue is the issue of Spatial Indexing and efficiency. How can we solve it?

Chatting

Article 1ArticleWithout special format descriptions or metadata, we can understand its knowledge and obtain relevant information from it, because the text is based on semantics rather than expression, but the existing is system or GIS system, is completely based on expressions, or even visualization. This is probably the initial driving force of semantic web or GIS.

XML is popular because it is self-described and oriented to semantics rather than expression (HTML is based on expression technology). Therefore, it is proposed to have a semantic web. However, semantic web may never come. RSS is popular, or the next technology is popular. The final results are consistent, and the convergence and transition of technologies and environments are maintained.

However, we may not see the emergence of semantic GIS, but can we have RSS-like technologies for data exchange and interoperability. For research, one inspiration is to start from the demand, not to "ignore the ears ".

The popularity of RSS is that a technology requires not only needs but also needs to be rooted in the environment. In the study of Semantic Web, RSS is popular in advance in the trend of optimistic about and promoting web services. One important reason is: first, the simplicity of RSS; second, web is the mainstream, and software is already an accessory or infrastructure. Third, the rapid increase in the amount of information leads to a greater dependence on search rather than organization for information retrieval, google's popularity is the result of interacting with this trend. RSS is more suitable for search. Fourth, for semantic web, the concept of Web is not only information, but also software and user interface, it is a medium for users to interact with computers. Semantic Web only shows the original meaning of web.