Preferred Conventions Multiple People/Artists with same name ?

[Vicpa 573]

Hi All

 

What are people doing with people with the same names? Is there a preferred convention ? As I have been working on my metadata this has started to crop up. This also occurs across types. I came across an actor with the same name as an artist.

 

Thanks any help would be appreciated.

 

-Vicpa

[Koleckai Silvestri 1150]

It is all problematic because the system doesn't handle duplicate people names very well. People seem to be referenced by name, not a unique ID. Though they have a unique I'd in the data.

[Luke 37276]

It is all problematic because the system doesn't handle duplicate people names very well. People seem to be referenced by name, not a unique ID. Though they have a unique I'd in the data.

 

we're moving to being entirely id-based but it's not necessarily going to resolve this. in most cases we don't get external id's from metadata provides, just names.

[Vicpa 573]

In the future, as long as we can have multiples with the same names, and I can manually pick the right one or add one. It is a step in the right direction.

 

 

Thanks

[ebr 14960]

It is all problematic because the system doesn't handle duplicate people names very well. People seem to be referenced by name, not a unique ID. Though they have a unique I'd in the data.

 

That depends on what data you're talking about.  Most of the providers from which we get information just provide a name, not an ID for people.

[Koleckai Silvestri 1150]

 

While that isn't always true based on this TMDB API Call (http://docs.themoviedb.apiary.io/#get-%2F3%2Fmovie%2F%7Bid%7D%2Fcredits), the biggest issue is how MediaBrowser handles people once they are in the system. The problem with handling two people with the same name isn't based on how they are retrieved. It is based on how Media Browser stores the information. They are stored in the ImagesByName folder by name. Take Kevin Smith here:

http://www.imdb.com/name/nm0003620/?ref_=fn_nm_nm_10

 

Or maybe this one:

http://www.imdb.com/name/nm0808963/?ref_=fn_nm_nm_61

 

Yet, I can only have one Kevin Smith in the system because I can only store one XML and image with that name. Since I am a fan of the Clerks series and Xena: Warrior Princess, this gives me faulty data.

Edited May 19, 2014 by Wayne Luke

[Luke 37276]

without having a unique identifier from the providers it renders our storage a moot issue. the only way we could handle it on our side independent of provider data would be to take the xbmc approach and store people information + images locally within each series or movie. but then you have lots of duplicate images and you'd still run into the same problem if you wanted to cross-check and find out what other titles each person appears in.

[ebr 14960]

Well that API has evolved and our method for storing these items dates way back.  And changing that method will be quite painful.  I'm sure we'll get there some day.

 

But, as Luke said - we need an industry-standard ID system.  That ID you see in that result, I'm pretty sure is defined by tmdb and not any universally accepted party that everyone uses.  I could be wrong there because, as I said, this API has evolved over the past few years.  But, it doesn't look like an IMDb Id...

[Koleckai Silvestri 1150]

Nothing else uses universally accepted IDs. They are different from each provider. The only Data Provider for People right now is TMDB so I am not sure why others matter at the moment. 

 

Anyway, people need to be exposed better. So they can be edited easier and duplicates are allowed. 

[ebr 14960]

Actually the IMDb ID for a film is pretty universally accepted and we do use that.  You can lookup a movie on TMDb using an IMDb ID.  People are much more likely to have information gathered from multiple sources as well (images especially). Plus, we need to work across all the other products too.

 

If there isn't a universally accepted method of identifying people beyond their name then programs like us, metabrowser, mcm, xbmc, etc. cannot all work together to define a structure in which to save and access them.  e.g. If you use metabrowser to gather people images and we store our people by ID instead of name, then metabrowser needs to be able to know what that ID is too.  Repeat for every other system we want to remain compatible with.

 

I agree we need this but it doesn't exist now and it will take some coordination between a number of different systems to make it all work.