[mb-style] CSG: Classical "superworks" and "performance of" links

Frederic Da Vitoria davitofrg at gmail.com
Tue Jul 19 08:13:00 UTC 2011

2011/7/18, symphonick <symphonick at gmail.com>:
> On Mon, 18 Jul 2011 22:23:14 +0200, Frederic Da Vitoria
> <davitofrg at gmail.com> wrote:
>>> I agree. So what I'm trying to say is that it is - as always -
>>> preferable
>>> to be as specific as possible, if the data is available. & if we know
>>> exactly what 4 parts of a super-work is performed, IMO it's more
>>> accurate
>>> to link directly to those than to a super-work containing more than 4
>>> parts. If there's no different versions or added movements to consider
>>> (like in the OP), then I'd choose the most convenient option. No
>>> difference in accuracy what I can see.
>> And create 32 ARs for a global performance of the Goldberg Variations? It
>> seems a little overkill to me, but if you feel it is really important, I
>> won't fight it. It won't happen often. I still don't like saying that an
>> AR
>> to a super-work means fuzzy, I'd rather have an explicit "fuzzy"
>> attribute,
>> but if others agree your solution is better...
> I don't know if it's important or not, maybe the inaccurate cases are
> rare, so we don't have to worry about it. I'm hoping we can figure out
> what we want from works; what level of detail can we manage?

I am not sure they are rare, I even believe they are the majority :-(

> If we forget my "add a fuzzy work" suggestion, and try your fuzzy
> performance, would it be something like this:
> There are 2 super-works in the db:
> Sinfonie in g K. 550 (1. Fassung)
> Sinfonie in g K. 550 (2. Fassung) (assuming for now that this is the
> "default")
> and we have a performance where we don't know the version:
> Symphony No. 40 in G minor is a (assumed) performance of Sinfonie in g K.
> 550 (2. Fassung)

I guess we are misunderstanding each other. I was wrong writing about
a fuzzy attribute in my last post, this was misleading.

The most logical to me would be to add a generic or fuzzy work named
"Sinfonie in g K. 550 (fuzzy)" which would be used each time someone
would not know which specific version to use. From my point of view,
users would in a way be encouraged to use those fuzzy works because
I'd rather have unspecific ARs than bad specific ARs. Something like
"If you are unsure about which version was performed, please use the
fuzzy version".

This is where the "attribute" comes in: "fuzzy" works (and the ARs
using them) would somehow be easily extractable, so that interested
users could check those ARs and try to link to specific works instead
of the generic ones. Instead of an attribute, we could use a
normalized term in the title, for example "[unspecified]": "Sinfonie
in g K. 550 [unspecified]"

> (someone can find a better term)

I agree we need something better than fuzzy. Indeterminate? Uncertain?
Unspecific? Unspecified? Generic?

Frederic Da Vitoria

Membre de l'April - « promouvoir et défendre le logiciel libre » -

More information about the MusicBrainz-style mailing list