Nepomuk Metadata Extractor moved to KDE Review

[Jörg Ehrichs]

Hi,

>> First:
>> If the servcie is started the first time (which happens automatically when
>> it is installed) It will disable itself and also alter the
>> restart-on-reboot nepomuk config file.
>
> Not sure I get it, what's the use of this? You assume "user installs it -> is
> fine with requesting metadata for files from online sources", I think. I'm not
> sure what changes between the first and the second run, however.
>

The reason this is in is the following setup:
* User installs it / or it might be installed by default on a distro
* User thinks he can just use it to search for metadata on its own
(Dolphin ServiceMenu integration)
* While in the background the Nepomuk2:Service queries all
documents/videos/music files automatically

As there is no notice about this happening and Nepomuk starts all
service automatically.
The mentioned change above disables the service (and adds the
necessary config entry so it will not
be started on next restart of kde)

This will now prevent a user from accidental exposing data to the
internet unless he explicitly starts
the service on his own.

>> This means, if the user really wants to have automatic fetching, he needs to
>> enable it in the config on its own and no one is forced to expose their
>> private data/file on their disk if they do not want it (or if they are not
>> aware of it)
>
> Maybe after first installation, it makes sense to actually ask the user. I'm
> usually not a huge fan for this, but in the case of possible inadvertend
> privacy breach, it's warranted to be proactive here. I could imagine:
>
> At first run popup comes up, "You've installed online metadata extraction. It
> will collect additional metadata from the web (such as ... and ...), but it
> can reveal files on your disk to third parties. Only enable this for file
> types that do not affect your privacy. If in doubt, disable it. (Links to
> config, metadata extraction should probably have a separate page in the
> Nepomuk KCM.)
>

I don't like pop-up dialogs to inform the user that something "that
might be harmful" was started,
So as mentioned above, a disabled by default service seems to be the
way better solution here.


>> In the default configuration, the document lookup is disabled.
>
> Makes sense, lookup for music and video is enabled by default, I suppose?
>

yes music/video is enabled by default, once the service is started by the user.

>> Third:
>> It is now possible to search for the preferred plugin only. Beforehand, all
>> availabe plugins would be used, this is not the default case anymore, but
>> can be enabled again.
>
> What's the preferred plugin? Is that a plugin for a specific service, a
> specific resource type, or ...?
>

The preferred plugin is a user selected plugin for each resource type
(music, movies, tvshows, documents).
Each plugin (written in python/javascript/ruby) represents one
website, where data can be searched/extracted from.

Currently the defaults for this:
Documents: Microsoft Search Academics
Music: Musicbrainz
TvShows: thetvdb
Movies imdb

>> This should hopefully be enough privacy settings and don't leave any user
>> with a system that does not do what he expects.
>
> Is there any UI for this?
>

Yes there are ui parts for all the above mentioned settings in the KCM.
I'll blog about the latest changes in the next days.

> I think this is a very cool feature, and has the possibility to make Nepomuk
> really shine. Your work on it is much appreciated!
>

Thanks I really hope this stuff will make it a lot easier to fill the
Nepomuk DB with more information so all
programs can make proper use of it (Bangarang already does this quite nicely)

Cheers,
Jörg