I think everyone who is at least slightly interested in the world of software technology in the open source camp has heard about the new KDE branch number 4. This version of KDE has brought a lot of new products, which have already been discussed on Habré. This is
Solid , and
Phonon , and
Plasma , and
Decibel . Did you know that KDE4 has one more thing that was not in previous versions? This thing is called
NEPOMUK-KDE .
[NEPOMUK]Honestly, for the Russian ear this word sounds somehow, well, quite unusual. But this is not a word at all, it is an abbreviation:
Network Environment for Personalized Knowledge Management .
The main idea of ​​NEPOMUK is metadata.
')
[metadata]What is metadata?
Meta - label. It turns out that metadata is a label to the data?
In fact, metadata is information to information. For example, I have a music file
Within Temptation - Angels.mp3 . If you open this file in any audio editor or audio player, we will see the file metadata - information about the author, the title of the song, the year of release of the song or album, genre, and more.
Metadata can contain almost anything. Another question is how to distinguish metadata from data.
Metadata is also data. This means that metadata can have its own metadata, and those, in turn, have their own metadata.
[metadata types]Metadata are of three types:
- Metadata that can be found in files stored on the hard disk. This data such as tags in audio files, file access time, or simply indexed text. This is the simplest metadata that can be indexed at any time. And it is these data that are the main goal of such projects as Beagle and Strigi ;
- Metadata created by the user. In the simplest cases, this can be a comment to a file or the merging of several files into one group (tags);
- And the most interesting and at the same time the most complex type of metadata, which is very difficult to collect, because they are not stored for long, and they are not created manually by the user. An example of such metadata is the URL from which the file was downloaded from the Internet. After the file is downloaded, the URL information is completely lost. Or another example. The user has received an email with an attachment. As soon as the user has saved the attachment to the hard disk, the attachment’s connection with the letter in which it came is lost. This means that we will no longer know where the attachment came from, from which letter. These are just the simplest examples.
With NEPOMUK, the goal is to be able to manage all three types of metadata listed above.
[Why do we need metadata if we lived well without them? ]But see for yourself. Just an example on electrobooks and electrophotographs.
Personally, I have a good collection of electronic books sorted by directories. I have a guide to iptables (packet filter in Linux). The meaning of this guide is suitable in the category of OS / Linux, and in the category of IT Security. Where to put the document? Where to look for him then?
I have a lot of photos of different people and places. All photos have a name like this:
45DE1.jpgFriends come to me and ask me to show the amazing nature of Sakhalin. Uh-uh ... how to find the pictures I need? Scroll through the entire collection? And if it consists of thousands of shots? Sort in advance photos with nature in a separate directory? And if the pictures are not only nature, but there are people? How then to look for photos with people, if they lie in the same heap with nature?
This is where metadata will help us. One of the commonly used types of metadata is tags. In order not to go far, I will tell you that you can see tags on Habré (they are also tags). Very convenient, isn't it, click on the Google tag and get all the topics related to the topic of a popular search engine.
By the way, and after all it is possible to expose ratings to files. For example, audio files, like / dislike. Well, or any other principle, choose yourself. =)
In the case of a book, I can put
linux tags
, it security . In the case of photos, I can put the tags
Sakhalin, Peter, Masha .
Then I don’t have to scour the file system, I’ll just click on the desired tag and get a list of files.
This raises the question: why now we need to sort the files, rename them, if we can hang tags and get the desired file, regardless of its name and location in the file system? I have no answer to this question.
[problem of metadata and person]Unfortunately, man is a lazy creature. If he is too lazy to give the imputed name of the file, then why did we think that it would not “break off” to expose tags to files?
I do not know the answer to this question. Therefore, we are waiting for an advanced automatic tagging system.
[ in custody ]Today NEPOMUK is able, of course, not all. But this is still a young project, the rapid development of which is still ahead.
Finally a couple of screenshots.
NEPOMUK integrated into the KDE4 Dolphin file manager
(note the right column) :

NEPOMUK & Strigi:
