Concept of "hashing" stings in Amarok (sih1)

Mathias Panzenböck grosser.meister.morti at gmx.net
Mon Jan 11 16:32:48 CET 2010


On 01/11/2010 02:44 PM, Jakob Kummerow wrote:
>> In fact, we've always over the years gotten emails asking for *more*
>> specificity, which is why we recently switched the database to be
>> case-sensitive. People *want* to know when they have both AC-DC and
>> AC/DC so that they can fix their tags.
> 
> I concur that fixing your tags is a more robust long-term solution.
> Especially when fetching covers/lyrics/Wikipedia articles or whatever,
> it doesn't help at all if Amarok internally knows that "AC/DC" and
> "acdc" are probably the same band; rather, when requesting stuff from
> external services, Amarok must provide the exact correct spelling
> (unless the other side also employs a "sih1"-like algorithm).
> 
> So, to make sih1 (or similar ideas) really useful, it would need to be
> established as an internet-wide standard - which is probably about as
> easy as just establishing correctly spelled tags as a standard ;-)
> 
> One point where I do see its usefulness is in a (standalone?) tagging
> application, where such an algorithm can be used to present the user a
> group of tracks that are *probably* performed by the same artist,
> giving the user the chance to correct the tags accordingly. But from
> my POV, it's at least debatable whether it would be a good idea to
> integrate that directly into Amarok.
> 

A GUI that provides a workflow to merge/normalize tags would be helpful.
Such a GUI would automatically detect groups of similar names and present them
to the user. The user can add/remove names from it and say to which name all
tags shall be changed.

	-panzi


More information about the Amarok-devel mailing list