Types of granularity and the TOG to facilitate modelling for the GIS domain

During and in between all the research traveling over the past half a year I also managed to write some papers. One of them is an invited book chapter [1] based on chapter 2 of my PhD thesis, i.e. the taxonomy of types of granularity with some additional material to make it self-standing to read. This book chapter that will appear in Novel Developments in Granular Computing (edited by JingTao Yao) early next year is to some readers, however, still rather abstract. To try meet feedback on how to apply these types of granularity and the TOG, I applied it and wrote a paper about using them to improve representation of granulation hierarchies in the subject domain of geography and ecology. This case study for representing semantics of granularity in Geographic Information Systems [2] will be presented early next year at the Geomatics’09 conference. The abstract of what I like to think to be a, from a potential user perspective, very readable paper (pdf) is as follows:

Dealing with granularity in the GIS domain is a well-known issue, and multiple data-centric engineering solutions have been developed to deal with finer- and coarser-grained data and information within one information system. These are, however, difficult to maintain and cumbersome for interoperability. To address these issues, we propose eight types of granularity and a facilitating basic theory of granularity to structure granulation hierarchies in the GIS domain. Several common hierarchies will be re-assessed and refined. It illustrates a methodology of first representing what one desires to consider for a GIS application, i.e., at the semantic layer, so as to enable reaping benefits of flexibility, reusability, transparency, and interoperability at the implementation layer.

A nice extra, for me at least, is that the Geomatics conference will be held in Havana, Cuba, as part of Informatica’09. Though I can understand Spanish and speak it a little (well, by mixing it with Italian), I do appreciate they are making it into a bi-lingual event with simultaneous translation Spanish/English. Looking at the preliminary programme (details online here soon), the following topics and people are already booked in: Oscar Corcho on semantics and the grid, Joep Crompvoets on spatial data, Michael Gould on data infrastructure, Robert Ward of the International Hydrographic bureau about marine data and information, several ISO representatives on various topics, and more researchers on open source geoinformatics, precision agriculture, managing remote sensing data and other topics, which will be presented by people from, mainly, the Americas and Europe (Cuba, Brazil, Chile, UK, Belgium, Spain, Switzerland, Italy, and Germany, among others).

For those of you who cannot physically attend—be it for financial reasons or due to the blockade—but would have liked to be there: you also can register as a virtual participant.

[1] Keet, C.M. A top-level categorization of types of granularity. In: Novel Developments in Granular Computing: Applications for Advanced Human Reasoning and Soft Computation. JingTao Yao (Ed.). IGI Global. (in print, not yet online—contact me if you want to have a preprint).

[2] Keet, C.M. Structuring GIS information with types of granularity: a case study. VI International Conference on Geomatics, 10-12 February 2009, Havana, Cuba.


New book on innovations in information systems modeling

To give my bias upfront: the book that contains my first book chapter is released today, in Innovations in Information Systems Modeling: Methods and Best Practices (part of the Advances in Database Research Book Series), which is edited by Terry Halpin, John Krogstie, and Erik Proper. To lazily copy the short description, the book has as scope (see title information sheet):

Modeling is used across a number of tasks in connection to information systems, but it is rare to see and easily compare all the uses of diagrammatical models as knowledge representation in one place, highlighting both commonalities and differences between different kinds of modeling.

Innovations in Information Systems Modeling: Methods and Best Practices provides up-to-date coverage of central topics in information systems modeling and architectures by leading researchers in the field. With chapters presented by top researchers from countries around the globe, this book provides a truly international perspective on the latest developments in information systems modeling, methods, and best practices.

The book has 15 chapters divided into four sections, being (I) language issues and improvements, (II) modelling approaches, (III) frameworks, architectures, and applications, and (IV) selected readings, containing altogether 15 chapters. The book chapters, whose abstracts are online here, range from refinements on subtyping, representing part-whole relations, and adapting ORM for representing application ontologies, to methodologies for enterprise and active knowledge modelling, to an ontological framework for method engineering and designing web information systems. The selected readings sections deal with, among others, a formal agent based approach for the modeling and verification of intelligent information systems and metamodelling in relation to software quality.

The chapter that I co-authored with Alessandro Artale is called “Essential, Mandatory, and Shared Parts in Conceptual Data Models” [1], which zooms in on formally representing the life cycle semantics of part-whole relations in conceptual data models such as those represented with ER, ORM and UML. We do this by using the temporal modality and some new fancy extensions to ERvt—a temporal EER based on the description logic language DLRus—to cover things such as essential parts, temporally suspended relations, and shareability options such as sequentially versus concurrently being part of some whole. To aid the modeler in applying it during the conceptual analysis stage, we also provide a set of closed questions and decision diagrams to find the appropriate life cycle.

A disadvantage of publishing with IGI is that they don’t accept latex files, but the poor lad from the typesetting office was patient and did his best to make something presentable out of it in MS Word (ok, I wasted quite some time on it, too). I don’t have a soft copy of the final layout version, but if you would like to have a latex-ed preprint, feel free to drop me an email. Alternatively, to gain access to all the chapters: the early-bird price (until Feb. 1, 2009) knocks off $15 of the full price of the hardcover.

[1] Alessandro Artale, and C. Maria Keet. Essential, Mandatory, and Shared Parts in Conceptual Data Models (chapter 2). In: Innovations in Information Systems Modeling: Methods and Best Practices, Terry Halpin, John Krogstie, and Erik Proper (Eds.). IGI Global, 2008, pp 17-52. ISBN: 978-1-60566-278-7

Additional suggestions for conference blogging

Lunch Over IP has an interesting blog post with tips for conference bloggers (pdf) covering twelve topics: tools, location, preparation, software, speakers, style, quotes, audience, context, linking, tagging, timing, mistakes, collaboration, tagging, and timing. These suggestions by Ethan Zuckerman and Bruno Giussani are useful suggestions for blogging about ‘general’ conferences, but I would like to add a few suggestions for scientific conference blogging, and those of computer science in particular, which are the principal outlets for the latest research (as opposed to journal articles in other disciplines).

The main modifications concern preparation, speakers, and timing, which is based on the conferences and workshops I did blog about (ORM’06, DL’07, OWLED’07, AI for cultural heritage 2007, AI*IA’07, IFIP TC9 ICT for warfare, OWLED’08, ISWC’08, ICT for Peace Symposium’08), the differences in quality of those post, the ones I started writing about but abandoned, the ones that I intended to blog about but did not do, and why for some I did not even start the process.

The, by far, most important point is preparation. Look up the accepted papers, decide on a theme, try to get the relevant papers beforehand, and read them. Split them into the stack of ones are worthwhile the “blog-attention” regardless, which to have as “potentials” and which to “discard”. For those where there is no paper available before the conference, skim through the paper upon receiving the proceedings, or at least mark them to go to the presentation and check the paper after the presentation.

Then, at the conference, attend the presentations and make notes primarily for those ones you have pre-selected and only by exception one that seemed unexpectedly interesting or that generated quite a bit of debate from the audience; in a good conference, there is too much new information to digest properly to summarize all presentations adequately, so not only preparation but also selection is important. Lunch Over IP mentions collaboration, which might be useful provided you team up with someone who has different interests or attends a parallel session. On the other hand, it also can be useful to have multiple reports and arguments about the same paper & presentation, in particular if one is attending an interdisciplinary conference. Further, even a lousy presentation but good paper should be worth mentioning: a scientific conference is not a marketing exercise where better-sold goods deserve more attention, but instead those papers that add something significant (the presenter could be a brilliant but nervous PhD student, humble researcher, or socially-challenged professor). Vice versa, a good presentation may mask a lousy paper; if there is such an ‘unexpectedly interesting’ presentation, then before blogging about it be sure to check the paper and consult an expert if it is not precisely your area of research.

The third point, timing, which Lunch over IP would like to see as liveblogging: posting within 10 minutes after the presentation. Well, no; let us call the opposite lagblogging. Aside that new things may pop up during the presentation—e.g., having misunderstood a section, newer material has been presented, criticism from the audience you had not thought of—one should back up any posted comments with an argumentation, which takes time to write, or compare it with another paper on the same topic that might be scheduled afterward, or even in another timeslot. Or perhaps there are links between papers one has not thought of before. Such papers should be synthesized into one analysis and not processed and published in a piecemeal fashion. Being able to connect dots is important in science, and when you do it in your post, the readers will appreciate that: not being at the venue, your blog readers were not exposed to the amalgamation of topics and papers, so your synthesis will give added-value. Make a connected ‘flow’ out of the selected papers and presentations. In that case, being a day or two (or three) later is fine.

Last, but this may be just my personal opinion, when I read other people’s conference blog posts, I really do not care who you rubbed shoulders with. First and foremost, I want to know what is useful to check out (and why), what was the ‘vibe’ of a panel session to get an idea of what lives in that research community, and what was deemed worthy of ‘keynote speech’ by the organisers (and was it really worthwhile listening to?).