[Data-modeling] Proposed changes for music composition and instrument modeling

[brendan]

for musical instruments, the complexity and hierarchical structure can  
be comparable to automobiles:

maker
     series
         model
             model year
                 actual instance in the real world (with a Serial#)


Though it is probably overkill, all of this detail is of much interest  
to aficionados

maker: Gibson
series: Les Paul
model: Les Paul Standard
year: 1959
instance: Keith Richards' '59 Les Paul Standard played 1964-1966

the high order bit here is: a '59 Les Paul Standard is *not* even  
close to the same thing as an '08 in terms of the quality/value of the  
instrument. On the other hand, Gibson does identify them both as being  
of the same model.

For hand built instruments, like a Stradavarius violin, I doubt the  
"series" (or even "model" for that matter) is talked about since they  
didn't have marketing departments back then :)

So, how to reconcile, simplify?  Perhaps:

maker (Gibson)
     model (Les Paul Standard)
          instance (KR's Les Paul)

maker (Stradivarius)
     model (Cello)  ("model" for lack of a better term here)
         instance (Davidov Stradivarius)

So, that brings me pretty close to what you propose, except, as and  
example, "Gibson Les Paul" is not a model the "Les Paul Standard" is.   
So we need a separate type:

series (Les Paul)
with properties:
"model",  pointing to the models (Les Paul Standard, Les Paul Deluxe,  
Les Paul Goldtop)
"maker " pointing back to maker (Gibson)

models can have a "years produced" property  (maybe even just a date  
range)

and the instance can specify the year and SN of the actual specimen.

So, a little denormed but collector types can create some of the  
details they're interested in while the schema remains pretty simple.

thoughts?

brendan











On Nov 7, 2008, at 12:04 PM, Christopher R. Maden wrote:

> A few changes have been proposed to modeling of musical compositions  
> and
> musical instruments.  I think they’re pretty good, but I’d like to get
> some wider feedback before implementing them.  For the most part,  
> these
> are new types and properties, so nothing will break, but it would be
> nice to have some confidence that we got the model right before
> implementing it.
>
> First, for compositions:
>
> A new type is proposed, called something like “type of composition.”
> Instances would be symphony, opera, string quartet, pop song.
> Compositions would connect to type of composition.
>
> The types of compositions would be a phylogeny pattern; a string  
> quartet
> is a kind of chamber music, a rock opera is a kind of opera and a kind
> of rock composition.
>
> Types of compositions would have typical instrumentations, using a CVT
> to get a number and a performance role: a string quartet typically has
> two vioins, a viola, and a cello.  A brass quintet typically has two
> trumpets, a horn, a trombone, and a tuba.
>
> Compositions would get a new property for instrumentation as well,  
> using
> a similar CVT.
>
> As for musical instruments:
>
> We currently have a Musical Instrument Company with no properties.  We
> would change this to Musical Instrument Maker.  We would grab the
> current Guitar Maker property from the Guitar type and move it over to
> Musical Instrument, and reciprocate it from Musical Instrument Maker  
> as
> Models Produced.  We would also add a property to Musical Instrument
> Maker for types of instruments made.  That would allow us to express
> that Steinway makes pianos, and also that Steinway makes the Steinway
> Model L.  Gibson makes acoustic guitars and electric guitars, and
> specifically makes the Gibson Les Paul and the Gibson SG.
>
> Thoughts, comments?
>
> ~Chris
> -- 
> Christopher R. Maden
> Data Architect
> Freebase.com: <URL: http://www.freebase.com/ >
> Metaweb Technologes, Inc. <URL: http://www.metaweb.com/ >
> _______________________________________________
> Data-modeling mailing list
> Data-modeling at freebase.com
> http://lists.freebase.com/mailman/listinfo/data-modeling