Edit: After typing all that below, I neglected to clarify that a release which is literally empty is kind of silly, indeed, and there are better ways to handle it even with our limited current tool-set. The following is the original post I made mostly in response to the discussion of whether media with musical portions have a place in the database:
I hard-agree with @aerozol on this issue, fwiw. I understand wanting to restrict the scope of MusicBrainz so that it doesn’t get out of hand, but it seems very counter to the database’s mission statement to not allow inclusion of music ‘by-portion.’
I think the disagreement here is philosophical, and based in the idea of completeness. For instance, must a Release consist of nothing but itself, and nothing other than what we can represent in MusicBrainz about it? There are cases where there’s “more to” a piece of art than its musical portion, far beyond just games and movies, and to deny those entry in the database would, I feel, be to shoot the DB in the foot. And drive away a lot of like-minded users whose hearts and heads are otherwise in the right place.
Currently, to the best of my knowledge, we don’t represent Releases as “unordered collections” or anything, but a single Recording that represents the entire audio portion of a piece of media not a farfetched thing to store in the database. We already sit squarely at the top of this slippery slope by storing songs that are typed as “music videos.” Even if you want to argue that something doesn’t belong as a release, allowing Recordings to exist in the database for its aural representation isn’t exactly a difficult accommodation to make.
Furthermore, take the case of games with music rooms, which provide a means to play the music in an ordered fashion, often with at least some basic metadata presented. In this case, the game becomes an entire medium and you’re looking at a sort of nested-media release. The means by which you play it back isn’t conventional or traditional, but when you get down to brass tacks, the release of something which contains a soundtrack is a release of that soundtrack.
And, one last assertion of principle. The database exists to facilitate the mission statement. The mission statement is to be the ultimate music database. Music is contained in things. This needs to be handled, and if the database schema is not up to par to accomplish that task, then the database needs to change to accommodate it, rather than users being restricted by the schema. That’s putting the cart in front of the horse.