Hi guys, I join the discussion short introduction: I have been adding user context support into my "Gnowsis Semantic Desktop" prototype in 2003, based on input that Sven Schwarz has given me back then. Later I initiated the nepomuk project together with Stefan Decker. I love to say wise things, but I don't have time to code KDE...
I propose that we learn what the problem is, ask ourselves some wise questions, and find wise answers, preferably backed up by working code or good "scientific" evaluations. And we should start a wiki page for this. which wiki? kde? dev.nepomuk.semanticdesktop.org Context is a great opportunity and its really simple to make it work well, but we all have to learn a little before (I have to learn waht the problem on your side is, for instance) Its also a great opportunity to fuck it up, so lets not do that :-) It was Hari krishna Anandhan who said at the right time 21.08.2008 05:15 the following words: > Wow! Looks like a great deal of research has already gone into > contextual thinking! > and that is only scratching the surface of a single diploma thesis student. now think that there is a lot of software products out there that do this already, and many more finished PhDs and PhDs underway. I do not say that all that is is good, but at least the discussion about Context and user observation in KDE should not start out of the blue, but based on solid knowledge of what works and what not. note on this: > Though the concept of widgets described in the paper already matches > with the capabilities of plasma "widgets" and that plasma already has > a dashboard that is better than Apple's IMHO, i think i can safely say > that what we are trying to achieve with plasma is a lot more than > "Context-sensitive Dashboard" presented in thesis. > > But, i guess that I will leave that part for Aaron seigo to explain as > he is the lead developer of plasma and the man behind the big vision > ;) > The widgets as such, the presentation, is important, but not as important as the underlying data and messaging. What you usually need is multiple components that work together (I try to translate from academia to DBUS) * applications observe the activities of the user and communicate them, to other applications and to a central "user work context" daemon that gathers activities also, all activities may be logged for later use for example: the user opened file "My Plan to get famous" in KWrite and starts editing * daemons report background context for example: a GPS daemon connected to a GPS mouse sends the position of the user as context information * the "user work context" daemon receives these messages and computes what the user is probably doing, this is important to have a daemon for this, as user often press alt-tab and you need something clever to keep track of what the user is really doing * applictions use the "current work context information" to give assistance or other features: the task management application says "your plan to get famous could be related to IMDB.com, you did similar surfing last week" The last part, applications, is where the widgets come into play. We also know of other applications, here is a snippet from a document we have written: ==== snip ==== On the desktop, despite the difficult task of identifying a context, we can ben- efit if this in many scenarios: receiving suggestions about useful documents related to what we are currently working, searching and receiving relevant documents within a context, being able to automatically provide annotations for documents, and also for better organizing personal work into tasks. * Context Assistant Bar A straightforward application for the detected contexts is the one in which the user is working on his computer and the algorithms watching his actions are able to identify the context in which his actions take place and display those contexts in an "assistant bar". For example, a visualization of the current context can be imagined, where documents, places or people relevant to the context are shown, and the user is able to browse through them. Also, possibly relevant documents can be automatically opened to reconstruct a previous working context. For example, a user working on a deliverable, always opened the relevant publications and also the emails regarding directions on how the deliverable should be structured. * Information Retrieval: Search for Resources When we think of searching and retrieving documents, context can help us by emphasizing the importance of the resources within a certain working context. For example, when working on something related to context representation, and submitting the query "user context" to the search engine, the resources that are also related to context representation but not direct containing the term would also appear ranked higher. * Suggestions for Document Annotations A nice application would be for the system to be able to suggest annotations to the working documents. For example, the user is working for a project, and the algorithms would then annotate the current document with the name of that particular project. Of course, a document can be part of several contexts. In this case, a document would receive annotations from each of the contexts. === end of snip ==== There is a load of other applications, and all this is implemented somewhere and there is more, etc. I repeat: let us make a wikipage and gather questions and answers, then go the next steps. Please don't hurry, you can only fuck it up once. best Leo > Side note: Should I continue adding all the above email ids to further > discussions or just mail to nepomuk-kde ? > > > On Wed, Aug 20, 2008 at 7:46 PM, Sven Schwarz <[EMAIL PROTECTED]> wrote: > >> Hi guys! >> >> Just reading over http://www.kde.org/announcements/4.0/desktop.php I see: >> >>> Plasma is KDE's new desktop shell which provides new tools >>> to start applications, presents the main KDE user interface, and >>> offers new ways of interacting with your desktop. >>> >> Which somehow urges me to show you the diploma/master's thesis of my recent >> student: (link removed ) >> >> Although the evaluation was not done as good as intended, his prototypical >> implementation of the Context-sensitive Dashboard rocks. The implementation >> is written in pure Java plus minimal Windows hacks to get events of the >> user's typed and copy/pasted texts. >> -- ____________________________________________________ DI Leo Sauermann http://www.dfki.de/~sauermann Deutsches Forschungszentrum fuer Kuenstliche Intelligenz DFKI GmbH Trippstadter Strasse 122 P.O. Box 2080 Fon: +49 631 20575-116 D-67663 Kaiserslautern Fax: +49 631 20575-102 Germany Mail: [EMAIL PROTECTED] Geschaeftsfuehrung: Prof.Dr.Dr.h.c.mult. Wolfgang Wahlster (Vorsitzender) Dr. Walter Olthoff Vorsitzender des Aufsichtsrats: Prof. Dr. h.c. Hans A. Aukes Amtsgericht Kaiserslautern, HRB 2313 ____________________________________________________ _______________________________________________ nepomuk-kde mailing list nepomuk-kde@semanticdesktop.org http://lists.semanticdesktop.org/mailman/listinfo/nepomuk-kde
