Bumpy desktop search experience (KDE 4.3.4)

Hi,

after wrestling with nepomuk to get it to work at all, I have both
“Enable Nepomuk Semantic Desktop” and “Enable Strigi Desktop File
Indexer” checked in the KDE Desktop Search preferences without any
errors. Interestingly, it seems the latter checkbox only works when the
former is checked…

Anyway, to me as an enduser it is not quite clear what the desktop
search frontend GUI is supposed to be. I can search in the krunner
(Alt+F2), but there is also strigiclient.

Now it seems that strigiclient starts up a strigidaemon which maintains
an index seperate from nepomuk? I’m asking because I get different
search results in krunner and strigiclient, and because strigidaemon
really only started using CPU when I used strigiclient, and it does so
in parallel to nepomukserverstub.

Also, so far I didn’t get it to index files on a CIFS mounted
fileserver, at least I can’t get any corresponding search results. Yes,
I do have the corresponding directories checked in “Advanced Settings”
of the Desktop Search System Settings.

Thanks for any hints,
Jörg

Hi,

As far as I know, Nepomuk is different from Strigi. Nepomuk is used with files that have metadata. It can be used in conjunction with Strigi to index files with metadata. You can read more about it here:

Discover Nepomuk As a User | Nepomuk

Strigi is the desktop search function that indexes your files and allows you to search for files on your computer from krunner.

Take Care,

Ian

Hi Ian,

thanks for the hint. So I guess I have to use strigiclient for
full-text-search, even though I wonder why the KDE preferences for
Nepomuk are called “Desktop Search” then, if it’s only about indexing
file meta-data.

Your link shows screenshots
<http://nepomuk.kde.org/discover/user?q=image/view/4/_original> from
some “Simple KNepomuk Search Client”, does anybody know whether this
thing exists somewhere to be downloaded and installed? I couldn’t find
it anywhere. Strigiclient isn’t exactly comfortable.

Whenever I start strigiclient, it doesn’t show any contents where it
should show directories being indexed. Also, upon starting strigiclient,
the strigidaemon starts off eating my CPU, although I had it running for
a couple of days now (so I gues it does index some directories, even
though it doesn’t show them). It seems to be reindexing where it
shouldn’t, and I remember having read this in some other forum as well.
Now when typing something in the search field, it takes minutes until
results are shown (all this on a quad core machine with 8GB RAM). Only
10 results are shown, without any way of seeing the rest. The results
never include anything from directories on a CIFS-mounted fileserver,
though I configured its directories.

To me all this seems premature, counter-intuitive, buggy and, in the
end, completely unusable.

I wonder what other people’s experiences with Desktop Search in KDE are?

Regards,
Jörg

On 12/11/2009 03:56 PM, ijbreakey wrote:
> Hi,
>
> As far as I know, Nepomuk is different from Strigi. Nepomuk is used
> with files that have metadata. It can be used in conjunction with Strigi
> to index files with metadata. You can read more about it here:
>
> ‘Discover Nepomuk As a User | Nepomuk’
> (http://nepomuk.kde.org/discover/user)
>
>
> Strigi is the desktop search function that indexes your files and
> allows you to search for files on your computer from krunner.
>
> Take Care,
>
> Ian
>
>
>

well, as far as i can see:
“To me all this seems premature, counter-intuitive, buggy and, in the
end, completely unusable.”
pretty much explains how i feel about it too. If it isn’t eating recources, then I don’t bother having this in background. but I noticed it consumes quite some RAM.

Hi,

It seems that Nepmuk does index any files, e.g. .java files, and it does
index their full contents. I can find some Java source files by their
contents in krunner now.

On 12/11/2009 03:56 PM, ijbreakey wrote:
> Hi,
>
> As far as I know, Nepomuk is different from Strigi. Nepomuk is used
> with files that have metadata. It can be used in conjunction with Strigi
> to index files with metadata. You can read more about it here:
>
> ‘Discover Nepomuk As a User | Nepomuk’
> (http://nepomuk.kde.org/discover/user)
>
>
> Strigi is the desktop search function that indexes your files and
> allows you to search for files on your computer from krunner.
>
> Take Care,
>
> Ian
>
>
>

I’m noticing the same thing over here. I’m new to KDE and using Kubuntu 9.10 (moving over from gnome).

Strigiclient wasn’t even installed by default! Clicking on the nepomuk system tray icon shows that strigi indexing is always idle, and there’s no way from the configuration panel to actually access strigi’s indexing database. And worse, when strigiclient never remembers the directories I add and remove! In fact, I can’t even get it to add directories anymore. And it took me a few minutes to figure out that the empty text box at the bottom is actually for searching! And why isn’t strigi integrated into Dolphin or Konqueror? It seems far more useful to most than nepomuk. A truly awful experience so far.

(I am trying to access strigiclient right now and it hangs!)