Ben non, ils réinventent pas la roue! En clair, ce qui interesse actuellement les gens de Kde avec Tenor, c'est le contexte des informations, pas d'écrire un bête indexeur. Un peu comme Winfs sauf que pour l'instant Microsoft a l'air de penser que l'utilisateur n'aura que ca a foutre à faire des associations entres ses fichiers. Après ca me semble un projet très ambitieux mais on verra bien !
Well, I one thing that I would add ... a couple of things that are happening in the Appeal project that we didn't really get to discuss thus far taht I find really interesting are, one: the contextual link enginge. We were talking previously by email, you had asked is there a search engine sort of thing coming around. And search tools are a really hot topic right now. It seems every desktop environment is throwing up a search tool. In the KDE project I think in the last two months I've seen three or four different approaches from people that are working on it on their own. We think that's kind of just the first step and that what's a lot more compelling than saying, "Let's throw all the full text and all the meta data into a big glop database and let you full text search it" is the ability to record, mine and then go back and navigation and search through contextual relationships between information.
It's funny how much information we throw away on the desktop. Insane amounts of information. If you're doing research on the Internet, for instance.. An example I've used is say you're doing an essay on panda bears. You go on the 'net, you go to ilovepandabears.com, find an image you really want to use in your presentation, you download that image. When that image is downloaded to your disk, we no longer know where it came from. We've lost a bit of context for that image. Then you take that image and you put it in your word processing document. We have lost yet more of the context because we really no longer know that that image is actually the same image as the one that's on disk and that it actually came from the Internet, in fact it actually came from this website.
We're looking at capturing that sort of information and allowing the user to actually navigate and move through their information using that contextual web. Uou look at Google, that was how they broke through: they use their whole page ranking thing. The idea of context being more important even than the content of the actual information you're looking at. And if you examine how we use groupware or how we build word processing documents or slide presentations, media files, the whole bit there's a lot of context that were not allowing people to gather.
We don't allow people to annotate information freely and then collect those annotations. We have these sticky notes, KNotes or whatever, that you stick randomly in places on screen that aren't semantically connected to anything.
The contextual linking engine, which is called Tenor, allows us to start bridging that gap. It really starts allowing us to provide context-based, in other words informational or meaning based, searching and navigation on the desktop. And this is widely applicable from things like KControl and the relationships between the different panels to managing your files in a content manager to your address book: Who is Jane Doe andw hat is she associated with in your life?
[^] # Re: Kat
Posté par gnumdk (site web personnel) . En réponse à la dépêche Beagle : Un "Desktop Search" sous Linux. Évalué à 4.
Sinon pour beagle y'a ca :
http://www.kde-apps.org/content/show.php?content=28437(...)
Interview de Aaron Seigo :
Well, I one thing that I would add ... a couple of things that are happening in the Appeal project that we didn't really get to discuss thus far taht I find really interesting are, one: the contextual link enginge. We were talking previously by email, you had asked is there a search engine sort of thing coming around. And search tools are a really hot topic right now. It seems every desktop environment is throwing up a search tool. In the KDE project I think in the last two months I've seen three or four different approaches from people that are working on it on their own. We think that's kind of just the first step and that what's a lot more compelling than saying, "Let's throw all the full text and all the meta data into a big glop database and let you full text search it" is the ability to record, mine and then go back and navigation and search through contextual relationships between information.
It's funny how much information we throw away on the desktop. Insane amounts of information. If you're doing research on the Internet, for instance.. An example I've used is say you're doing an essay on panda bears. You go on the 'net, you go to ilovepandabears.com, find an image you really want to use in your presentation, you download that image. When that image is downloaded to your disk, we no longer know where it came from. We've lost a bit of context for that image. Then you take that image and you put it in your word processing document. We have lost yet more of the context because we really no longer know that that image is actually the same image as the one that's on disk and that it actually came from the Internet, in fact it actually came from this website.
We're looking at capturing that sort of information and allowing the user to actually navigate and move through their information using that contextual web. Uou look at Google, that was how they broke through: they use their whole page ranking thing. The idea of context being more important even than the content of the actual information you're looking at. And if you examine how we use groupware or how we build word processing documents or slide presentations, media files, the whole bit there's a lot of context that were not allowing people to gather.
We don't allow people to annotate information freely and then collect those annotations. We have these sticky notes, KNotes or whatever, that you stick randomly in places on screen that aren't semantically connected to anything.
The contextual linking engine, which is called Tenor, allows us to start bridging that gap. It really starts allowing us to provide context-based, in other words informational or meaning based, searching and navigation on the desktop. And this is widely applicable from things like KControl and the relationships between the different panels to managing your files in a content manager to your address book: Who is Jane Doe andw hat is she associated with in your life?