.

Quoting Dr Pavlov:

Quote:
ERRORS based on hard data, not uer-interpreted data. What you THINK is an error is for you to deal with locally.

WSIWYT, Taro

Skip

I think we mean the same thing, but I'm not sure. What I meant with errors on hard data is that the credits themselves contain errors that have nothing to do with personal ideas of preference but are plain errors.
Fictive example:
At the end of Mission Impossible, the lead actor is credited as 'Tom Cruuse'. That is clearly and error, not because I think it is, but because it actually is an error when cross-referencing with other reliable sources. I would like to be able to display that he is credited as 'Tom Cruuse', but still have an option to show it is indeed 'Tom Cruise' as well as cross-linking correctly to that actor.

Here's a concrete example I've run into:
The Japanese actor Tooru Ookawa has many possible writings in the database, among which:
- Tôru Ôkawa
- Tooru Ookawa
- Toru Okawa
The first two are correct because they use one of the official romanization methods, as approved by the Japanese government and worldwide universities. The third one however doesn't match any of the official romanization methods and as such, is an error in the hard data, not because I think so but because it goes against international, official writing rules for romanized Japanese.

I agree that personal interpretation is very debatable, but as it is now there are also irrefutable errors in the hard data and current rules do not permit users to rectify these, even with strong evidence (like linking to an actor's personal website or even to the website of the Japanese government with the official spelling rules).
For example, for the sake of uniformity we could debate which of the two, Tôru Ôkawa or Tooru Ookawa is the best way, even though both are officially acceptable writings. That is indeed personal interpretation and a matter of internal consensus. However, Toru Okawa is wrong. That transcends my preference, yours or even the preference of the distributor. It is a matter of not complying with internationally agreed upon romanization rules. In short, I think that the following priority order applies:
our opinion < credits <<< official writing rules

An interlinked cast/crew database with the official names and linked to them the various aliases and 'credited as' names would solve these issues, I believe, while still retaining the hard data for those that wish to display it.

Quoting mdnitoil:

Quote:
Honestly, I can come up with a much more elegant process that completely maintains exact onscreen data, yet still provides universal linking.

In my mind, the need for a centralized cast/crew database is paramount.  In my world, everyone downloads the db and starts to build profiles with these core names.  Robert Downey, Jr.'s name is displayed on the credits differently?  No problem, simply edit the display name for that particular DVD profile.  By having everyone choose from the same core people, the linking becomes automatic regardless of spelling variations on individual profiles.

Now, further down the road we discover that Joe Shmoe and Joe T. Schmoe are the same person.  Document the fact, get it accepted into the cast/crew database.  Ken runs a backend process to adjust online profiles with these two names consolidating them into one.  The next morning, I find that I have a cast/crew database update and 10 profile updates automatically generated to create the new expanded linking.  The display name is still attached to the actual DVD profile, so exact onscreen credit integrity has been maintained.

Now it's a heck of a lot easier for me to type this than it is for Ken to code it, but this would be my idea.  If we're going to invest all this energy into discovering what all these common names are, wouldn't it be nice once the work was done to have all their profiles adjusted automatically?  The system we have now requires us to do all this work and then it still depends on folks in all regions to then update every single profile that's affected.  Herding cats would be an easier chore.

That's a very interesting idea and I believe a workable one at that too. It takes the current 'credited as' system basically and puts it online, where the actual name of a person is the key to link to. Very good suggestion, if you ask me, and well explained!

Quoting Taro:

Quote:
Quoting Dr Pavlov:

Quote:
ERRORS based on hard data, not uer-interpreted data. What you THINK is an error is for you to deal with locally.

WSIWYT, Taro

Skip

I think we mean the same thing, but I'm not sure. What I meant with errors on hard data is that the credits themselves contain errors that have nothing to do with personal ideas of preference but are plain errors.
Fictive example:
At the end of Mission Impossible, the lead actor is credited as 'Tom Cruuse'. That is clearly and error, not because I think it is, but because it actually is an error when cross-referencing with other reliable sources. I would like to be able to display that he is credited as 'Tom Cruuse', but still have an option to show it is indeed 'Tom Cruise' as well as cross-linking correctly to that actor.

Here's a concrete example I've run into:
The Japanese actor Tooru Ookawa has many possible writings in the database, among which:
- Tôru Ôkawa
- Tooru Ookawa
- Toru Okawa
The first two are correct because they use one of the official romanization methods, as approved by the Japanese government and worldwide universities. The third one however doesn't match any of the official romanization methods and as such, is an error in the hard data, not because I think so but because it goes against international, official writing rules for romanized Japanese.

I give you another example (real):
The actress Eva Reuber-Staier is credited Eva Rueber-Staier in 3 James Bond films and in Space 1999 and is credited with her correct name in 3 title.
The CLT results are:
"Eva Reuber-Staier" is credited in 5 titles (12 profiles)
"Eva Rueber-Staier" is credited in 28 titles (208 profiles)

So the "common name" (most credited name) is the wrong name Eva Rueber-Staier

In the 3 titles where she is credited with her correct name we use Eva Rueber-Staier [Eva Reuber-Staier].

We don't worry about the real name, but only about what we see on the screen.

The problems with the asian name are about the romanization if they are on the screen in non-standard characters, and about the parsing.

Quoting Dr Pavlov:

Quote:
I see three major problems with the whole Common Name theory.

1) NameA=NameB, assumptions are going to cause problems, n users are using ONLY the Name Variant Tool and if names are similar that is close enough, that is until someone discovers that close wasn't close enough.
2) Users are not bothering to look p[ast the CLT numbers, well now what does that mean. That means very simply, we know that the CLT is a wreck, how much crossover is there, that crossover is a problem and the CLT can't be considered useable until any given crossover is either verified as correct or corrected, not assumed that it is wrong, like some are doing, but ACTUALLY checking to make sure it is indeed WRONG and then correcting.
3)Finally deciding which part of the CLT we are using to determine Common Name, there is no consistency, I think most of us are using the number of titles, But I know some users are the number of Profiles and some are using whichever one is higher. We all have to be on the same page.

Skip

I understand where you are coming from here, but we have those exact same problems today.  Moving to a centralized cast/crew database doesn't magically absolve all these similar names, but when we do finally get one nailed down, we stand a much better chance of being able to make large-scale changes with the online data rather than the current micro changes of one profile at a time.  I wouldn't even necessarily mind the micro approach, but we need a whole mountain of users across all regions to cooperate in order for it to work.  Practice has shown that this won't happen.

If person A does indeed equal person B, consolidate the data once in the central DB and be done with it.  And let's face it, who really cares what is used as the common name just so long as the onscreen credits are maintained?

Quoting Taro:

Quote:
I agree with Skip that we need to come up with a system that as much as humanly possible prevents errors being entered in the database.

Prohibit contributions.

Human failure is by far the most common source of error.

Quoting bbbbb:

Quote:
Prohibit contributions.

Human failure is by far the most common source of error.

Nice one!    

Quoting xradman:

Quote:
You have got to be kidding.  I think just about anything would be better than our current system.

In his first sentence he says "there really is no easy way to do it," not that there isn't a better way to do it.  The problem with all these suggestions is that they require a lot of work on Ken's part which, as he said in another thread, he doesn't have the time to dedicate.

OK, time to put this out there...

I don't think anybody is trying to subvert the rules.  We are all trying our best to work within them, but the rules are not perfect and to continue to defend these rules as sacrosanct when they don't work is not going to get us any closer to a solution.

We are all trying to find a way to make the data we capture with this application usable and helpful.  If the rules inhibit our ability to capture useful data, then maybe, just maybe, it isn't the members that have the problem, maybe the rules need to change.

I think this is what many here on this board have been trying to say for many years.