"The MIX project (Mediation of Information using XML) is a collaboration between the UCSD Database Laboratory and the Data-Intensive Computing Environments (DICE) group at SDSC. The goal of the MIX Project is to study, develop, apply and evaluate systems for mediation across heterogeneous information sources. We strongly believe that the Web emerges as a distributed database and XML (or some extension/modification of XML) will be the data model of this huge database. In particular, we envision that in the near future sources will be exporting an XML view of their data along with semantic descriptions of the content and descriptions of the interfaces (XML queries) that may be used for accessing the data. Users and applications will then be able to query these view documents using some XML query language. Along this direction we have developed the XMAS query language."
XML indicates a move towards viewing the Web as a large, semistructured database, consisting of many autonomous sites that will be modeled around XML and related standards for structure and ontology definitions (XML-Data, RDF, DCD, namespaces), APIs, source query capability specifications, and other standards that may describe the transactional abilities of the sites. We focus on wrapper-mediator systems which employ XML as a means for information modeling, as well as interchange, across heterogeneous information sources. The wrapper associated with each source exports an XML view of the information at that source. The mediator is responsible for selecting, restructuring, and merging information from autonomous sources and for providing an integrated XML view of the information."
"We are developing wrappers for a variety of information sources including, relational databases, GIS systems, and Web sites with HTML pages. Mediation is based on the MIXm mediator -- and the associated XMAS query language -- being developed by the UCSD Database Lab component of MIX project. The Database Lab is also developing the BBQ interface. We are pursuing the application of such mediator systems for distributed digital libraries and virtual agencies."
References:
MIX and GIS - Integrating GIS and Imagery through XML-Based Information Mediation By Amarnath Gupta, Richard Marciano, Ilya Zaslavsky, and Chaitan Baru. [San Diego Supercomputer Center, UC San Diego; Email: {gupta,marciano,zaslavsk,baru} gupta@sdsc.edu.] Presented at the NSF International Workshop on Integrated Spatial Databases: Digital Images and GIS. 20 pages, with 31 references. Overview: "In this paper we investigate a mediation-based approach (referred to as lazy evaluation) to integrating information from a spatial information system such as GIS, and a database of geo-referenced imagery that possibly lives in a digital library. As in [other publications], our purpose is to provide a user the ability to issue a single query that involves both information sources, and receive a result that combines information from these sources in a seamless manner. In the same spirit we would like to provide an authorized user the ability to make permissible updates to a source by using conditions involving both sources. The goal of this paper is to demonstrate how a mediated system is designed and to step through the query evaluation procedure in an such a system. We must emphasize that our notion of integration does not rest on the development of spatial algorithms that operate on images or vector data and achieve "physical integration" (e.g., see [11]) through techniques like image conflation. We aim to attain "logical integration" by creating correspondences between related spatial information much in the same spirit as mediation systems do for databases. In this endeavor we do show existing physical integration techniques fits into the fold of our information association methodology, but development of such methods is secondary to the focus of this paper. . . This paper presented an architecture and a logical schema for Web-based spatial information mediation using XML. We have traced a sample query integrating imagery and GIS sources, through its evaluation at the spatial mediator and dispatching to XML-wrapped geodata sources, where query fragments are translated to the language of the source and executed. We believe that, while a first step in the development of scalable and extensible spatial data mediation systems, this exercise helped us elucidate areas of complications which create the context for future research . These proposed research areas include: (1) Development of rules and a cost model for selecting spatial data sources in the spatial mediator. (2) Development of a general cost model for parsing, evaluating, and distributing queries, and for assembling the results. (3) Incorporation of physical integration (alignment) management capabilities in the architecture of the mediator system, and in query planning at the mediator. (4) Supporting geographic analysis 'workflow' as a sequence of spatial queries to the mediator system. (5) Scalability analysis of the mediation of multiple GIS and imagery sources. (6) Supporting alternative mechanisms to associate names with geographic objects. [Etc.]" [local archive copy]
XML-Based Information Mediation for Digital Libraries - C. Baru, A. Gupta, V. Chu, B. Ludäscher, R. Marciano, Y. Papakonstantinou, P. Velikhov, exhibition program, ACM Conf. on Digital Libraries (DL'99), Berkeley, USA. [Postscript
"View Definition and DTD Inference for XML." - B. Ludäscher, Y. Papakonstantinou, P. Velikhov, V. Vianu, Post-ICDT Workshop on Query Processing for Semistructured Data and Non-Standard Data Formats (SSD99), Jerusalem, 1999.
[August 05, 1999] "XML-Based Information Mediation with MIX." By Chaitan Baru, Amarnath Gupta, Bertram Ludäscher, Richard Marciano, Yannis Papakonstantinou, Pavel Velikhov, and Vincent Chu. [University of California, San Diego, La Jolla, CA 92093; Email: {fbaru,gupta,ludaesch} marcianog@sdsc.edu, {fyannis} pvelikhog@cs.ucsd.edu.] In ACM SIGMOD Record Volume 28, Number 2 (June 1999), pages 597-599. Published paper from the Proceedings of the 1999 ACM SIGMOD International Conference on Management of Data, May 31 - June 3, 1999, Philadelphia, PA, USA. "As part of the MIX (Mediation of Information based on XML) Project at the San Diego Supercomputer Center and the University of California, San Diego, the MIX mediator system (MIXm) has been developed. MIXm is a fully XML-based mediator prototype whose graphical user interface BBQ (Blended Browsing and Querying) integrates the browsing and querying of XML data. Powerful grouping and ordering operators are supported by means of a novel collection list construct. BBQ automatically generates XML queries from the graphical query specification given by the user. The system solely relies on the XMAS (XML Matching And Structuring Language) query language for extracting XML data (both from the mediator and from the sources). An algorithm for DTD (document type definition) inference for certain views has been developed. In the demonstration, the system is not only presented from the end user's (BBQ) perspective, but also details of the generated XMAS queries and algebra expressions can be shown. . . MIXm uses XML as the common model for data exchange. Mediator views are expressed in XMAS (XML Matching And Structuring Language), a declarative XML query language. To facilitate user-friendly query formulation and for optimization purposes, MIXm employs XML DTDs as a structural description (in effect, a 'schema') of the exchanged data. The novel features of the system include: (1) Data exchange and integration solely relies on XML, i.e., instance and schema information is represented by XML documents and XML DTDs, respectively. XML queries are denoted in XMAS, which builds upon ideas of languages like XML-QL, MSL, Yat, and UnQL. Additionally, XMAS features powerful grouping and order constructs for generating new integrated XML 'objects' from existing ones. (2) The graphical user interface BBQ (Blended Browsing and Querying) is driven by the mediator view DTD and integrates browsing and querying of XML data. Complex queries can be constructed in an intuitive way, resembling QBE. Due to the nested nature of XML data and DTDs, BBQ provides graphical means to specify the nesting and grouping of query results. (3) Query evaluation can be demand-driven, i.e., by the user's navigation into the mediated view."
XML-based Information Mediation - Position paper at QL98/W3C.
Or: "XML-Based Information Mediation with MIX." (ACM SIGMOD 99)
"A Framework for Navigation Driven Lazy Mediators." By Bertram Ludäscher, Y. Papakonstantinou, and P. Velikhov. "This paper presents the framework of the super-lazy MIX mediators. The mediator exports a virtual view XML document into which the client navigates. Client navigation commands are then translated into source/wrapper sequences of navigation commands."
"Enhancing Semistructured Data Mediators with Document Type Definitions." By Yannis Papakonstantinou and Pavel Velikhov. In Data Engineering 99. "Mediators, such as MIX, have to provide DTDs of the views that they compute in order to facilitate query formulation, query processing, and so on. However, a view has more than one DTDs. We propose two quality metrics of view DTDs - roughly similar to "precision" and "recall". The first one, soundness, requires that every possible view document is described by the view DTD. The second one, tightness, requires that the view DTD describes the minimum number of documents that may not appear in the view (ideally, the number is zero). Finally we describe an algorithm for deriving a sound and tight view DTD given a view that has a single variable in the head (SELECT clause) and no recursion (or regular expressions) in the body (WHERE clause)." See also Postscript format.
