Spending a rainy day exploring my new KDE4 installation, I find I have something called Nepomuk available and running. Googling for this, it seems it's something to do with a "semantic desktop"(?) but apart from a lot of management-style gobbledegook (in the style I heard more than enough of in my many years with British Telecom) I can't find much about how (or more importantly, WHY) I should use it. Apparently it integrates into Dolphin, which I don't use any more than I used Konqueror, as I can do the sort of things these applications do most effectively and quickly at the command line.

So what benefits does Nepomuk offer me and how do I use it? I don't like to dismiss new things out-of-had, but I'm struggling to find any point to it.

Nepomuk is an investment in the future. You can think of it as a database, but instead of saving tables, it saves relations or properties of objects. Objects might be files, contacts from the address book, webpages, etc. An application can save and retrieve information from this database through DBUS, e.g. which contact sent you the file as an attachment, or ratings and tags on files. It is extremely flexible, as you can define practically everything as relations. Applications need to use it, though. The first examples can be seen in Dolphin, Gwenview and the file dialogue, but it will take some time to get integrated in more locations.

Andre wrote:Nepomuk is an investment in the future. You can think of it as a database, but instead of saving tables, it saves relations or properties of objects. Objects might be files, contacts from the address book, webpages, etc. An application can save and retrieve information from this database through DBUS, e.g. which contact sent you the file as an attachment, or ratings and tags on files. It is extremely flexible, as you can define practically everything as relations. Applications need to use it, though. The first examples can be seen in Dolphin, Gwenview and the file dialogue, but it will take some time to get integrated in more locations.

Maybe it's something to do with the way I work, but I can't honestly think of a single occasion when any of your examples would have been useful. That's not to say there might not be something it would be useful for, but it's eluding me.

At the moment, it seems rather to be a solution in search of a problem. When I see statements like

empowering individual knowledge workers to better exploit their personal information space and to maintain fruitful communication and exchange within social networks across organizational boundaries.

(nepomuk.semanticdesktop.org) I feel like I'm playing Buzzword Bingo with a pointy-haired boss.

Then how about this? Plasma, Context and Nepomuk

Maybe you don't like this particular feature, but it'll hopefully give you a general idea of how Nepomuk can be used.

Nepomuk is really amazing and, for me, the biggest leap in KDE 4. I testing it in KDE 4.2 beta 1 with my TV shows and movie collection and work's great.

For example, I have a tag "Not viewed" and a lot of tags for persons and gender. I'm tagging my collection and now I can do, in konqueror, dolphin or krunner, amazing things like:

nepomuksearch:/ tag:"takeshi" and tag="dorama" and tag="not viewed"

and I can show all my not viewed japan TV shows that have a tag that contains "Takeshi.

There is a lot of work to do with this technology but now, in KDE 4.2, nepomuk can be used and is useful.

Meh. I still think it's a buzzword. :undecided:

You need it!

When Strigi starts working ( Sesame2 backend crashes every time I try and start it currently ) it will also provide file searching, integrated with KRunner as well.

bcooksley wrote:When Strigi starts working ( Sesame2 backend crashes every time I try and start it currently ) it will also provide file searching, integrated with KRunner as well.

After I installed KDE4, I noticed vast amounts of CPU being used by Strigi for no apparent reason; fortunately I found out how to disable it fairly quickly.

When I need to find out where a particular file is, "locate" does the job quickly, with only a small amount of system resource being used when the index is re-built. And on the odd occasion where I need to find a file containing any particular string, "grep" doesn't take much CPU or time unless I'm recursing through a large directory tree.

So with grep and locate I can find what I need without any of Nepomuk and Strigi's system-hammering and ****.

Of course "locate" and "grep" are powerful tools to find files, but Nepomuk/Strigi will do lots more when they get usable and performant. The case that Ignacio Serantes described yesterday is exactly what I am waiting for since I heard of Nepomuk the first time in 2007. And I wouldn't call it "****" when the performance is still suboptimal - after all it's still KDE4 and therefore still not perfect in all details...

Locate and grep are about the names of the files. Nepomuk and similar are meant to classify the content.

einar wrote:Locate and grep are about the names of the files.

Not if you grep within a file

Eventually we'll get KDE5, and then everything will be working. IN THE MEANTIME we're stuck with version 4. I don't use my desktop for playing, I use it for work. Nepomuk is getting in the way, so I shut it off. If it's not ready it shouldn't be the default. Suggestion to the core developers: Stop using our desktops as your playground!

Grumble, grumble.

beermad: grep is useless with binary files. Can you told me how can I search for and actor in my movies or for a compositor in audio files using grep?.

Brandybuck: Nepomuk is mature enough to be enabled by default in KDE 4.2 and is needed for KDE applications, like Dolphin and Gwenview actually and Amarok and Digikam in the future, to tagging and scoring. On the other side, strigi is good for nothing, eats a big lot of resources and slows down system. It will be a big mistake if it's enabled by default. Finally, if you need great stability KDE 5.0 probably won't be a solution for you and KDE 4.4 or 4.5 will be .

There's a nice article on Technology Review about Nepomuk. Gives some clues as to what the semantic desktop paradigm could mean: http://www.technologyreview.com/computing/21840/?a=f