Google Summer of Code 2008 : Mass tagging

[maxime rameau]

Hi,

here is my proposal for this summer of code concerning the mass
tagging. Is it possible to have some feedbacks on it please?

--------------------------
Amarok mass tagging


--Summary--
The goal of the project I propose is to improve the mass tagging
system of Amarok. Right now, the only way to tag file is to do it file
by file and it is just unthinkable to have a big collection well
tagged with this system.


--Project--
As an user, I enjoy the fact that Amarok is really easy to use. In
this way, the mass tagging system has to be user friendly too. I tried
several mass tagging programs and I discover that most of them are too
many difficult to use at the first look.

In order to keep the system simple, I think that the mass tagging
system should first operate in a  basic mode in order to propose the
user some results easily.
Firstly, a possible way to find tags is to calculate the id from all
the files of the album and to interrogate a web database like Freedb.
Secondly, the user could give some piece of informations thought a
textbox without specify their type (like the author name following by
the album name) and the program interrogate a web database.

If the basic mode failed to find good tags, then the user could switch
in advanced mode and use three new ways :
1°) Firstly, the web based search. The user, thought a form, could ask
a web database (Musicbrain, Freedb, ...) in order to get some results.
I saw a very good feature in program that allows the user to put wild
cards in the queries and I think it has to be implement.
2°) In certain cases, it could be possible to extract some piece of
informations directly from the filename. In this way, the user has to
define the separator and the field order.
3°) At least, if all previous methods failed, the user should be able
to tag manually the files. But, the user interface has to be simple
and quick to use. For instance, fields like « author » or « album »
should be fill only one time and it should be easy to switch quickly
between files.

With well tagged files, it is possible to order them correctly to have
a clean library. For instance, all the albums of an author could be
group in one folder. Each folder could be named with the same format
(for instance, the name of the album, following by the year of
release) and this feature is applicable for file too.

Other improvements that may be considered for this project :
with well tagged files, it's easy to search for duplicates and propose
to the user to delete them.
a good way to tag an album is to perform this action while ripping it
by calculate the disc ID. I don't propose to implement a ripping
system but it could be interesting to think this project in
consideration of this idea.
As soon as a file is tagged, it possible to search for his lyrics and
to download it.

--Road Map--
I could spend 20 hours a week in this project until the end of June
because I will be in school. After that, it will be my summer
vacations and I will be able to spend as much time needed by the
project (probably 40 to 50 hours a week).