Of course you don't , Ace. I wouldn't expect you to.

Quoting Ace_of_Sevens:

Quote:
I don't know what this means. Why can't the linking system be derived from who the person actually is? That's sort of the point of being able to do these variant linkings. We are tracking people, not names. We could somehow tie the credited name to this fixed identity. This is how most every database I'm aware of works. It's how social security works. It's how registration in every college I've attended works and it's how IMDB works.

Actually, the way the system was set up, we are tracking names.  This started out as collection software.  As with all collection software, it was designed to track the individual traits of the items being catalogued.  As the userbase grew, so did the wants and desires.  Ken, people pleaser that he seems to be, has tried to accommodate those wants and desires.  Unfortunately, because the foundation was collection software, the result isn't what those people wanted...and IMDb style program for DVDs.

I don't know if we will ever get a satisfactory linking system without a complete program overhaul.  Only Ken knows the answer.

Quoting Prof. Kingsfield:

Quote:
The linking system must be derived from the ACTUAL credits, not the other way around, at least relative to the Online. The only question there is the best way, and i think most of us agree that the current system is not very workable, but there are answers. Will Ken change it to something else or leave us with what we have...I have no idea, he's the Programmer and it's his Program. I can say what I would do, but he does not have to go that way. Although I will say what I have always said, the method he chose makes no sense to me, but c'est la vie.

LoL. You made me think of Matrix Revolutions. You know, the Trainman? Ken=Trainman. Hehe.
"Down here, I make the rules [...] Down here, I am God!" 
------

Anyway, people. I think I came up with another Idea. It has also been suggested, in a form, just above my post.
My idea

Quoting Ace_of_Sevens:

Quote:
Quoting TheMadMartian:

Quote:
Actually, the way the system was set up, we are tracking names.  This started out as collection software.  As with all collection software, it was designed to track the individual traits of the items being catalogued.  As the userbase grew, so did the wants and desires.  Ken, people pleaser that he seems to be, has tried to accommodate those wants and desires.  Unfortunately, because the foundation was collection software, the result isn't what those people wanted...and IMDb style program for DVDs.

I don't know if we will ever get a satisfactory linking system without a complete program overhaul.  Only Ken knows the answer.

Right, I meant in theory we are tracking people, not names. Names are a proxy for people. They aren't a very good proxy, which is what causes our current problems.

                     

Quoting Staid S Barr:

Quote:
Recently I saw a suggestion (can't find it now) to use the birth date in addition to the BY.
However, I already found a case where even that doesn't help, not even in combination with initials, or gender (also suggested).

We have:
Robert M. Young, born November 22, 1924, in New York (mostly as director, but many other functions),
and
Robert M. Young, born November 22, 1924, in Edinburgh, Scotland (mostly as writer).

That is genuinely one in a million.

Sorry, I'm late to the party. I also voted for a better system and here is my suggestion of how I believe it can be improved.

- In the online, each actor/crew has his or her own unique ID
- Contributors enter cast and crew exactly as credited
- When submitting online, the online shows the closest matches existing in the online and asks you for each name that isn't an exact match to either:
    - select the name that matches that person (so they can be linked), in which case it automatically attributes that unique ID to that name
    - select the option new name, in which case it is attributed a new unique ID
    - the user can manually type in an existing name to make them link (useful for people who changed their stage names for example or wildly different romanization of Asian actors)

Locally then, we download those ID's from the online as well as the name that is used for that ID per profile/credit.
That way, we already have perfect linking with the names displayed as credited. Cast and Crew pictures are attributed to the unique ID, so if you add a picture for one of the alternate names, it's immediately set for all.

To give more flexibility to users locally, I can imagine a program update where a user can locally select a 'standard name' for a crew/cast ID. That becomes the local 'real name'. An option can be set with a radio button so that users can choose how to locally display those names:
- only the credited name
- only the real name (standard the user defined locally)
- both, credited name with real name between brackets


I think a system as described above allows for both the flexibility needed as well as the necessary local linking. The need for a unique ID is that because sometimes, names can vary so wildly that an automated linking system (which finds the closest matches and links them automatically) is unable to determine that for names that vary considerably (for example, stage names vs real names). Therefor, I think the best way is to have one unique ID per actor/crew and attribute to that the various names that person goes by as well as which name was used per profile or credit.

I'd add the following. We should do a system with unique identifiers for each person that are not their name, but also separate the movie and DVD databases. For instance, if I were entering the Terminator 2 Skynet Edition Blu-ray, I would enter that it contains Terminator 2: Judgement Day (theatrical) and Terminator 2: Judgement Day (extended) and Terminator 2: Judgement Day (Special edition). I would enter the running times and it would check them against known running times for those editions and question if it was off by more than a few seconds (we'd have to make allowances for PAL, of course) and just pull cast and crew lists based on that. We'd need a good system to keep different cuts separated, but this would be a lot less trouble than copying corrections to 100+ profiles.

Ideally, we would also enter all the docs on the disc this way and just flag them as special features.

Quoting Ace_of_Sevens:

Quote:
I'd add the following. We should do a system with unique identifiers for each person that are not their name, but also separate the movie and DVD databases. For instance, if I were entering the Terminator 2 Skynet Edition Blu-ray, I would enter that it contains Terminator 2: Judgement Day (theatrical) and Terminator 2: Judgement Day (extended) and Terminator 2: Judgement Day (Special edition). I would enter the running times and it would check them against known running times for those editions and question if it was off by more than a few seconds (we'd have to make allowances for PAL, of course) and just pull cast and crew lists based on that. We'd need a good system to keep different cuts separated, but this would be a lot less trouble than copying corrections to 100+ profiles.

Ideally, we would also enter all the docs on the disc this way and just flag them as special features.

That could indeed save out a lot of editing and correcting work for contributors, to have a basic list of a movie's various cuts to start with. On rare occasions I can imagine that the cast will need to be updated specifically for a certain profile (e.g. I've seen Blu-rays that have added crew for the HD remastering) but it's definitely a sound idea to build a base that can be reused.