I've been playing with Goodguy's Name Variants program which identifies possibilities for name linking. I have around 1900 variants to look at.  As I went through my list, the most common discrepancies I see are accented names that are credited without accents and also the names in the sound categories.  Those credits are all over the place. And as I looked for proof, I found that most databases out there run low on info after the Big A stars. Even for big name directors and cinematographers, the resources are much less. And sound crew? Forget it.

But what should I do when I make a decision about resolving a name variation? Should I contribute it and let the voters decided if they agree with my decision? Or do I have to prove somehow that Buzz Knudson and Robert "Buzz" Knudson are the same person or that an accented name and a non-accented name are the same person? If I have to prove that, would I then have to prove which of his credits is his 'common name'? Or should I just keep these local. Or what?

It would be great after I go through this list to offer this work to the main db, but I don't know to what extent I want to research 1900 names.

To what extent does everyone want "resolved" name variations contributed?

Contribute

Contribute only with proof that the names are for the same person

Contribute only with proof that the names are for the same person and prove the common name

Don't contribute

Something else

I don't care

Quoting m.cellophane:

Quote:
I've been playing with Goodguy's Name Variants program which identifies possibilities for name linking. I have around 1900 variants to look at.  As I went through my list, the most common discrepancies I see are accented names that are credited without accents and also the names in the sound categories.  Those credits are all over the place. And as I looked for proof, I found that most databases out there run low on info after the Big A stars. Even for big name directors and cinematographers, the resources are much less. And sound crew? Forget it.

But what should I do when I make a decision about resolving a name variation? Should I contribute it and let the voters decided if they agree with my decision? Or do I have to prove somehow that Buzz Knudson and Robert "Buzz" Knudson are the same person or that an accented name and a non-accented name are the same person? If I have to prove that, would I then have to prove which of his credits is his 'common name'? Or should I just keep these local. Or what?

It would be great after I go through this list to offer this work to the main db, but I don't know to what extent I want to research 1900 names.

To what extent does everyone want "resolved" name variations contributed?

Contribute

Contribute only with proof that the names are for the same person

Contribute only with proof that the names are for the same person and prove the common name

Don't contribute

Something else

I don't care

As we get further and further into this mess, I become more and more tempted to solve the whole thing the same way the Gordian Knot was untied!  Most of the information out there seems to be 2nd rate at best ranging down to non-existant.  Given the different levels of expertise and tolerance of the users, I don't think we can expect a fair distribution of effort in verifying these, even the A-list names.

I therefore suggest we follow one of two options:  drop the common name thing altogether and go strictly 'as credited'; or, do as Ken suggested and let the program plug in the common name based on what percentage of profiles show that name. 

If 70% (arbitrary number) of the profiles listing Actor X are the same, then that becomes the common name.  If, over time, the as cred. name changes to Actor X-Y and exceeds the number of profiles with the old name, it gets changed by the program.

In other words, take the choice completely out of the hands of users.  That will prevent huge arguments over documentation, name parsing, accents, etc.

Contribute and explain your decision if it can't be proved.

Ideally, the online DB would give some feedback based on the "Credited As" name, presenting you the common names already in use for it.

Quoting antolod:

Quote:
Based on Ken's comments in the Rules Committee, I would say bring them here to the contribution forum for discussion, vote, then contribute.

I think we could come up with a general framework for how to deal with them.  That's my hope with this poll...to measure what the community thinks is the right method for most of these. Case by case if there are debates, they can be discussed here. But I alone have 1900 of these to figure out, so I don't think we should bring each of those here. Surely we can have a framework to get the bulk of them through the normal contribution process. Save the discussions for the disputed ones.

James, I was hoping that Ken would run his "comon name finder" on the main db and create some for us. With that as a starting point, maybe there won't be so many go through. My concern is if everyone starts adding common names, all names will become common names, and we will be back at square one.

Having said that, I'm open to hearing ideas. For diacriticals, I would think that in 99.9% of those cases, the names are the same person. I think if you do contribute them, you should at least give your reasoning as to why you believe the two names to be linked. If there is controversy (never )
then bring it here for discussion/poll.

Quoting m.cellophane:

Quote:
I think we could come up with a general framework for how to deal with them.  That's my hope with this poll...to measure what the community thinks is the right method for most of these. Case by case if there are debates, they can be discussed here. But I alone have 1900 of these to figure out, so I don't think we should bring each of those here. Surely we can have a framework to get the bulk of them through the normal contribution process. Save the discussions for the disputed ones.

The good thing is, that you have 1900, I had a couple of hundred when I ran the tool, and many others using the tool may get similar results. ...and there will be overlap, so you'd not be doing the work alone.

Quoting hal9g:

Quote:
It seems to me that we should wait for Ken to complete that work before we spend a lot of time entering "Common Names" and linking them to "as credited" just to find out that we used the wrong "Common Name".

I was under the impression, that it was only a suggestion of his so far. In that case, he is not yet building that tool. But, I guess depending the outcome of this poll and discussion, we could ask him to move in that direction...


I voted for "Contribute with proof [documentation of some sorts; tougher for crew...] that is is the same person". I think if we wait for all discussion to get concluded we will still not be using this tool until 2008. At least the well known cases should get into the system and it is happening already anyway. Some of the contributions I have seen I personally thought I would have reversed Common and Credited As names, or a few more Notes would have been nice, but people are eager to use this and if we keep shooting down these contributions "until further notice", we are going to loose a lot of contributors, which is highly undesirable.

Quoting ya_shin:

Quote:
Some of the contributions I have seen I personally thought I would have reversed Common and Credited As names, or a few more Notes would have been nice, but people are eager to use this and if we keep shooting down these contributions "until further notice", we are going to loose a lot of contributors, which is highly undesirable.

I've been voting yes on the vast majority of these that I've seen, even though I think we should hold off for a week or two and see if Ken is working on a db.  He said he expects some enhancements to the website shorlty, so maybe I was just hoping it would include this tool. 

Just voted but what is proof? If I have 2 films and I can see it's the same guy in both that's enough for me. Eg Ray Winstone in Quadrophenia. But how do I prove it. In this case IMDB agrees with me but I would never rely on that solely.