iMeMex:Personal Dataspace Management System

NEWS

Please check our external web-site http://www.imemex.org for detailed information on this project. An initial version of our software can also be downloaded from that side.

If you are looking for a "Semesterarbeit" or a "Masterarbeit" take a look at the available topics or contact directly Dr. Jens Dittrich.

What is iMeMex?

Desktop computers provide thousands of different applications that query and store data in hundreds of thousands of files of different formats. Those files are stored in the local filesystem and also in a number of remote data sources, such as network shares or as attachements to emails. To handle this heterogeneous and distributed mix of personal information, data processing logic is reinvented inside each application. This results in an unfortunate situation: most advanced data management functionality, such as complex queries, backup and recovery, versioning, provenance tracking, among others, is (at least partially) performed by endusers in tedious, manual tasks. To solve these problems we propose a software platform named iMeMex that brings physical and logical data independence to the desktop, freeing users from low-level data management considerations. Unlike in relational DBMSes, our platform does not assume full control of the data, but rather manages the complex dataspace of one’s personal information. We attack several research challenges encountered building such a platform: (i) the definition of a unified data model that allows the integration of information in distinct representations and locations without requiring semantic data integration, (ii) the development of a new search&query language over this data model along with algorithms for the efficient processing of complex queries and (iii) the need for soft update and recoverability techniques to provide durability and consistency of personal information. (See the position paper "iMeMex: A Platform for Personal Dataspace Management", SIGIR PIM 2006 for an overview on the current status of the project.)

What does iMeMex mean?

iMeMex is an abbreviation for "integrated memex", where memex refers to the great vision provided by Vannevar Bush in 1945 (As we may think). In contrast to previous systems iMeMex does not try to implement yet another application on top of an existing operating system. In fact, we think that the myriad of different information management solutions is not part of the solution but, on the contrary, the root of the problem. Our system takes a new approach in that information management is deeply integrated into today's operating systems. In contrast to other approaches our system is not restricted to a single operating system. Currently iMeMex works on windows, linux and Mac OS X. The integration on the storage level enables exciting new information management capabilities - unseen before in state-of-the-art information management and operating systems. Our system iMeMex is a first implementation of a new breed of system termed Personal DataSpace Management System (PDSMS).

Why is there a Potato?

iMeMex handles many different types of data: tuples, XML twigs, text, events, binary data, file content, tokens, etc. The "row" or "tuple" abstraction does not work for all of these different formats. Therefore we came up with a generic data abstraction named Portable Data Token = PoDaTo. A PoDaTo may wrap arbitrary data of arbitrary granularity. (See the paper "iDM: A Unified and Versatile Data Model for Personal Dataspace Management", VLDB 2006 (IIS Track), Seoul, Korea for details on our data model.)

Note: we are always looking for motivated students that want to write there Semesterarbeit or MSc thesis in an exciting new research project (topics). You should have sound programming skills and a team-oriented working style. We have a broad range of topics available. Do not hesitate to contact us.

Contact

Dr. Jens Dittrich