I know this has been asked many times before, but I'd really like to question how somebody can begin to use Nepomuk within their workflow in a practical way.

For example, I could start tagging all my files. But I have a lot of files, and it's likely to end up a huge mess and a waste of time compared to when I use it. Especially as I don't see a good way to bulk add/manage tags. Perhaps tags should be "guessed" while indexing?

Star ratings are cool, but I don't see it used outside very large and varied media collections where you might want to only check out the cream of the crop of your collection - actually, I don't see anybody trying to search for anything below 5 stars "hey, I want to hear that reaaaly **** song again". I really don't see my documents in terms of /5 ratings.

Adding comments to files? Perhaps, but definitely not mass scale. I can see it being used once in a blue moon.

As for searching with Nepomuk, for example when I do alt-f2 I need to do "hasTag:foo" to find files tagged with foo. Why I have to prefix it with "hasTag" and why I had to search around forums to find this out does not represent the potential usability that Nepomuk is meant to offer.

Not trying to shoot down anything, but I really want to see Nepomuk become what it is capable of becoming - I would really appreciate any information on this.

You can also use "tag:" instead of "hasTag:". By the way, I'm working on a small PyKDE4 application to tag files and directories recursively. Nepomuk's API is nice enough to make tagging programmatically a breeze. Right now it's limited to certain files that are downloaded from certain sites (because they have the tags in their file name), but I guess it could be extended.
Also, pointers on how to use Nepomuk are also on Mandriva's documentation page. Notice that part of these examples require nepomuk-kde, which is in the playground area of KDE SVN and AFAIK hasn't been packaged by any distro but Mandriva.

Actually, just searching for the tag name (without the prefix) in KRunner will bring the tag up. You can also tag multiple files in Dolphin by selecting them and setting their tags as if they were an individual file.
Nepomuk will also index not just file names, but file contents as well (a la Strigi). The search bar in Dolphin or searching with KRunner has, for me at least, made a hell of a difference, though there's still lots of room for improvement.
You might not want to search for content that's only got 3 stars, but you might want to be able to sort content (e.g. Amarok's collection) by rating. I know I would
Also, Nepomuk is still in its rather early stages - though it's far more noticeable now then in KDE 4.0, there hasn't been as big a development rush in Nepomuk as there has been in Plasma, for example.

Still planned are the likes of contact aggregation and search (using Akonadi), searching for files by where they came from (an E-mail from Joe User, perhaps? Might also use Akonadi) and several other niceties. One thing I would particularly like to see is tight integration with Konqueror: visiting a known/popular website, like youtube, twitter, facebook etc., would give the ability to add contact information to Akonadi, with Strigi associating several methods of contacting someone with a single, "person" (something like Kopete currently does), so searching for, "Tim Baker" in KRunner would show, "Tim Baker: view more details" or some such, and selecting that would pop-up a new window with his last tweet, facebook status, facebook news, new Youtube videos, his E-mail address and instant messaging address, related events etc., and each would provide a type of link or button to take the user to one of the various associations with Tim Baker, like his Facebook wall or Youtube channel. Yeah, that would be awesome.

Just typing in a tag name into KRunner brings up the tag, yes, but not the files with that tag. Personally I think it's rather silly to bring up the tag, which for me opens up Dolphin with a search for files with that tag.

I'll play around with file tagging later (is it true you cannot delete tags?), but right now I'd like to question Strigi. Whenever I enable Strigi it indexes my files alright. But the problem is that it doesn't stop indexing them. Not when I'm trying to use my computer, not even when I right click the systray icon and check "suspend". Htop reports higher and higher CPU % values even hitting the high 90%. This on my Acer really pushes the temperature to unsafe heights. Strigi should do it when I'm idle - and if not, it should at least suspend when I tell it to.

I am using Gentoo with KDE 4.3.1. Is anybody else experiencing this?

Moult wrote:I'll play around with file tagging later (is it true you cannot delete tags?), but right now I'd

You can delete tags. In fact, I just did (KDE 4.3.1) because when testing my tagger something screwed up...

Yes, you are right you can delete tags. To try it out I tagged a directory full of photos, here are some issues I noticed:

1) When searching for photos, I would like to see previews of them through krunner, it really makes it much more convenient. It's not much use searching for a tag if it gives you a list of random filenames like DSC0573.
2) I didn't figure out how to do a search that combined two or more tags. How do I do this? EDIT: tag:foo tag:bar works

EDIT: added some more user-end issues:
3) Shouldn't Amarok pass on its star ratings? (this goes for gwenview, digikam, etc as well)
4) Option to add artist, genre, album, year, song title as tag to music files.
5) When adding a tag, it should dropdown a list while you're typing of already existing tags to prevent duplication, eg when I type "p" it should drop down "photos" "peter" "paul", etc. If not I could write "Photo" when "photos" already exists.
6) What about case insensitivity or plural-sensitivity when searching with tags?

I was thinking of what would be a great way for somebody completely new to nepomuk to start using it as part of their daily work: an application that uses strigi indexing to build up a list of your files, and at the same time checks whether or not they have been tagged. It creates a list of files where you can sort by date or type etc of all your untagged files. You can create exclude lists for example for dotfiles which I see no sense in tagging (but you can if you want to) and it shows up in the systray with a number of how many files you've got left to tag. This allows you to start bulk tagging files until you know you've got a system that knows what's inside it.

Whenever a new file is added and is untagged it will get added to the list. Whenever a file is updated, you have the option to retag it but this will show up in a separate list with it's own exclusion filters.

What do you think of that idea?