I posted this here, but I decided to repost it here:

Nepomuk: How do I actually USE it?

I can tag and rate files right in Dolphin. What benefit does that give me? What does it do if I rate some folder with 4 stars while other folder gets only 1 star? Can I use Nepomuk to search files (like using Spotlihgt in OS X)? If so, where?

Nepomuk is supposed to show me connections between different files (who sent what files as attachments etc.). Does that work yet?

Is there a "Nepomuk for utter dummies"-document somewhere? It's surely needed because I keep on reading comments about it, but I never see any documents on how to actually USE it.

I keep on reading snippets about Nepomuk, about Nepomuk-backends, tagging, searching... But nowhere do I hear anything about actually USING it in real life. Please: Educate me. I love the idea of Nepomuk, but I'm utterly lost when it comes to actually using it.

Yeah, there isn't exactly any graphical frontends that take full advantage of it yet (at least in stable KDE releases), but I'd assume in the future that more apps will come to use Neopmuk for data presentation, such as Dolphin and such.

Nepomuk was one of the feature that I loved most in the pre 4.0 era... And I still remember how happy I was when dolphin started to remember my tags and ratings
But indeed - it's still quite useless for me as a user. When there is a desktop search that can search for tags and rating and comments, it will surely be a revolution for how to work with folders, files and the computer itself...

I think nepomuk based search ( powered by strigi ) will be present in KDE 4.2, at least I know there was a nepomuk search KIO slave that went into trunk.

Nepomuk is first and foremost desktop infrastructure, i.e. something that provides services to applications.

Currently most applications which actually use such services only use storing/retrieving tags and annotations/comments.
Some of the might even allow to search/filter based on such information from within the interfaces.

The IO slave is a nice facility to present a search to applications not yet directly capable of using Nepomuk, however I am not sure if there is a generic GUI to create queries for it.
Generic searching can easily become a programming-like task, application-domain specific search needs can be presented much more userfriendly.

Cheers,
_

It's starting to appear in several applications besides Dolphin now... take a look at Gwenview, for example.

Yes, that all needs some time.

The nepomuk search krunner is now in trunk and will ship with KDE 4.2. So you will be able to search for files and tags with ALT+F2.

Gwenview was already mentioned. And with the new kio slave you can define virtual folders. (For example to use with a folder view plasma applet). Although a good GUI for that is still missing.

anda_skoa wrote:Nepomuk is first and foremost desktop infrastructure, i.e. something that provides services to applications.

Currently most applications which actually use such services only use storing/retrieving tags and annotations/comments.
Some of the might even allow to search/filter based on such information from within the interfaces.

The IO slave is a nice facility to present a search to applications not yet directly capable of using Nepomuk, however I am not sure if there is a generic GUI to create queries for it.
Generic searching can easily become a programming-like task, application-domain specific search needs can be presented much more userfriendly.

Cheers,
_

Why we need special GUI to create queries? If we go to root of ex. search:/ we have tags inside that folder. On all folder(tags) we have other tags and content folder. We can also go to search:/by-name and create ex. photo folder. If we go to search:/by-name/photo/content, we have any *photo* files in our HOME directory. Of course any tags have special prefix(tag-). We can also go to search:/by-name/photo/tag-love and have any *photo* with tag love files. We can also have search:/by-date, etc. And inside search:/by-name/photo we have by-date folder. One bad thing is we have to search by date. We can ask how input correct query? We can make predefined query inside all folders or some folder examples in root of search:/ . We can copy folder from examples and modify name. Of course search:/ will remember our queries.

Lachu:

The tags:/ kioslave is doing something similar but only for tags (and it is still in playgroung).

And also nepomuksearch works quite similar. nepomuksearch:/tag:Test for example.

The problem with your approach is, that it limits the possible querys. I can not image how you make "OR" querys with it. Or how to exclude (AND NOT) something. And Nepomuk will allow even complexer querys (for example "show me all files i received by someone who i have send mails to the last year").

Or well when you want to search for resources which are not files (kontacts, mails, websites (history/bookmarks), i can not image that you will do everything of that a kio slave.

For complex querys and for quick querying without knowing propertys and a syntax a nice gui is really needed.

And in addition to your idea: I would like to see a KIO slave which i can configure to create virtual filesystem structurs based on information in Nepomuk.

For example a virtual Folder structure "Music" were i can configure to show me first for a folder for every artist i have music of and in that folder wit the name "year - album" for ever album and in that the songs together with a png file for the cover.

And this kio slave should get configured to create folder structures based on every metadata. And well it could ship a default setting to do what you are proposing here.

DanielW

Are there already some prototypes or ideas for GUIs? I guess GUIs for creating Triples would be just as interesting as GUIs for querying the RDF base? Or does nepomuk use a fixed set of relations?