Subject: Grenoble SYMM f2f and INRIA visit trip report Date: Fri, 30 Jun 2000 10:15:32 +0200 From: Lloyd Rutledge Hello all, I was just in Grenoble, France, main city of the French Alps, and site of the 1968 Winter Olympics. I was there on Monday and Tuesday for the SYMM face-to-face meeting. I was there on Wednesday to visit our buddies at INRIA. I was there on Sunday to screw around. The biggest news in the SYMM f2f is that Aaron and Lynda are due to have a second child in January. That makes December a pretty firm deadline for getting the next version out (not that that's why it's the biggest news). The second biggest news is that it was finally decided to drop "SMIL Boston" and officially make it "SMIL 2.0". This switch will be made in all documents. The SYMM f2f was mostly about wrapping things up as we go to last call. "No new features" was the motto. Take a look at the minutes at http://www.w3.org/AudioVideo/Group/2000/f2f-grenoble/minutes.html Months of efforts for agreeing on base media types to "strongly recommend" for SMIL players resulted in recommending JPEG and PNG for images and Audio Basic for sound, and that's it. Nothing could be agreed upon for text because Real will not commit to implementing HTML because it would be a monumental task. We pestered them about it at previous f2f's, but who can blame them -- they're primarily in the streaming server business with multimedia as a tangential market development hobby. I don't know why nothing was recommended for video. The issue of needing desired features to be implemented came up often, of course, since this is just before the last call. One such contested feature is third party temporal markers, such as video timestamps on editing lists that annotate the video. Microsoft and Oratrix agreed to be the two vendors that each implement a simple case of external file temporal markers for interop testing. Perhaps we could help have in GRiNS shortly a tool that handles edit lists of some proto-beta-pre MPEG-7 that allows us to run useful examples of annotated video. Can anyone thing of useful characteristics of a preliminary external temporal marker format to be implemented in GRiNS (and IE 5.5+, for that matter)? July 15 is the date for going to last call. December 15 is the aim date for a SMIL 2.0 recommendation. The next f2f is mid-late September in Yorktown, New York, USA, 45 minutes up the Hudson from New York City. It is hosted by Michelle Kim of IBM, from MPEG-4 fame. I gave a talk at INRIA on Wednesday about our work with rhetorics, inter-dimensional hypermedia communicative devices and the constraints and resolution strategies that define them. Afterwards I spoke with them about what they are up to. They are doing some work that we had been thinking of heading towards. These discussions gave me a sense of deja vu, reminding me of our visit from Elisabeth André, and the discovery of topic overlap. It induced both the anxiety that work we'd been heading towards is already being done, and the intrigue that their work is on a different level and so their work complements ours well without invalidating it. Furthermore, INRIA's work is on a level closer to presentation than ours, whereas DFKI is on a level more abstracted from presentation than ours. For example, they have just submitted a paper on comparing the performance of different constraint systems and algorithms on handling multimedia problems. This makes our hope for a journal paper on comparing constraint requirements for the different communicative devices a bit redundant. But then, we would be comparing requirements, not resolvers. We would be comparing problems, whereas INRIA is comparing solutions. The folks at INRIA said they were looking for example applications to validate their tools with. We have something to contribute to them there. We have, and continue to develop, an understanding of user needs and application areas. INRIA understands constraints well, but doesn't really know what to do with them. So perhaps a basis for collaboration is that we provide problems that they, or CWI or all of us, test and validate and compare their solutions on. Perhaps we could get involved in an EU-funded project with them. This was discussed vaguely during the meeting. The lower key option would be to start a thematic network, like SIMOS was, were CWI and INRIA and other select groups meet several times a year for three years to discuss a topic in common. The topic could be "hypermedia authoring abstractions" or something like that. CWI and INRIA could be in. Perhaps Franca as well, bringing Italy in. Perhaps Oratrix would be interested in joining such a project and bringing to it industrial involvement?. Franca, if still interested, would be able to get some industrial partners, especially from publishing. Perhaps Maja would be interested (are you listening, Maja?), bringing in both industry and another country ... not to mention Maja. Perhaps DFKI or other members of their cartel would be interested as well: both research and industry. This could be a good way to maintain contacts and communication with these people. Doing an EU project would bring more money for us to grow our empire with, of course. But it would be a lot more paperwork. And, of course, we'd have to figure out what exactly we were supposed to do with all those involved. I don't think we are at that stage, but perhaps we could lead up to it, and a thematic network would be a good environment for that. How much new project money do we need, and how soon? An imaginable tangible project focus could be like how the CHAMELEON project was started, but updated. The project would develop a prototype for yet-more-abstract hypermedia authoring, with the idea that the results could/should be spun off into economic activity. It doesn't have to mean us starting a new Oratrix, with I wouldn't want. But someone else could. Or the prototype could be shipped off to Oratrix or some other existing company. I'm not proposing this particular idea, it's just a half-baked idea that is one example of the many possible half-baked ideas of which one could develop to the good idea that forms that basis of a project. Anyway, other new work INRIA is doing is that Nabil is investigating structured video: making temporal and spatial markers. I suggested he might be interested with working with Oratrix (and Microsoft) on developing this prototype third file marker format that enables the interop testing that enables temporal attributes to refer to timestamps in external files that are separate from the media files being annotated. Maybe us working with him on this would be a good initial collaboration in the short run. I have a copy of the paper he submitted on the topic to MMM00. Frank N, are you interesting in being involved in this. -Lloyd