No surprise that there would be cross-pollination between topics, given that content discovery depends on content publication...

http://talk.maemo.org/showpost.php?p...2&postcount=18

Exactly, which is why automatic contextual metadata is sooo handy. Time/date stamp is the obvious easy one, and location is for devices that have that feature... but to go further we get into the area where devices like the N900 really set themselves apart from traditional cameras, recorders, at all. And to explore that we come right back to profiles.

I don't know how many people fully configure their digital cameras before using them (I don't) but I think it's safe to say we personalize our phones to a large extent. All of the info so used is potential metadata for media sharing purposes.

And we could go further with automated tagging as well... weather conditions, temperature, and as I mentioned in the intelligent postcard example (linked above), local events. All of this and more could be pulled in using an active internet connection, and shared to recipients in useful contexts.

So, my daughter was watching Madagascar when I came home tonight. I wondered about the brain sizes and intelligence levels of the various animals. Which is smarter, a lion or a lemur? A hippo or a zebra?

A good content discovery system would let me find that out. A good content discovery system lets me follow a train of thought:

Madagascar (movie) - characters - animals - encyclopedia - bookmark intelligence - compare bookmarks.

Which brings us back to augmented reality, the ultimate (or maybe just penultimate) outcome of this train. True augmented reality will rely on a semantic web (thanks tswindell) which will rely on rich metadata sets which really on standardization (such a dirty word), and the crux of Cory Doctorow's distrust:

http://www.well.com/~doctorow/metacrap.htm

...and a shameless self-plug for my data cloud series:

http://tabulacrypticum.wordpress.com...days-for-data/

Right. I was thinking along the lines of building your profile gradually in the background using your content access and usage patterns together with metadata that should be almost trivially available - music titles & artists & genres, location tags, browsing history and the like. But obviously there's a limit for the generalizability (is that a word?) for all this.

-Otto

EDIT: As Texrat says above, automated contextual metadata is certainly important here. It could conceivably be produced by all kinds of apps for different types of content.

Well, it was fun participating in these preliminary discussions. For unexplained reasons I didn't get selected (have to admit to being surprised and disappointed) so whoever did, do us proud!

Texrat: Did you get a rejection letter or something?

Yes I did. No useful details though.

I am really, really stunned, too. But crap happens.

Time to change my flight plans...

I didn't even see a place to apply for the co-creation project.

Nah, just come and hang out in Amsterdam! It is a really nice city. A great museum, too.

There's a thread here somewhere with volunteer and nomination requests. I really thought my broad experience would get me in.

I'd love to come a day early but doubt I can afford a hotel room. Heck, I still don't think we have final word on the Summit rooms, and if I did come early I'd want to have the same room or at least hotel...