Welcome to the Invelos forums. Please read the forum rules before posting.

Read access to our public forums is open to everyone. To post messages, a free registration is required.

If you have an Invelos account, sign in to post.

    Invelos Forums->DVD Profiler: Contribution Discussion Page: 1... 9 10 11 12 13 ...15  Previous   Next
Credit Name Parsing
Author Message
DVD Profiler Desktop and Mobile RegistrantStar ContributorSH84
Registered: March 13, 2007
Reputation: High Rating
Germany Posts: 922
Posted:
PM this userView this user's DVD collectionDirect link to this postReply with quote
Quoting marcelb7:
Quote:
Quoting Kulju:
Quote:
d) Movie based credits instead of DVD based


Only with an option for localization. Some movies have localized credits (especially animation movies). Perhaps it could be tied it to the audio tracks (if you have an English track, show the English credits and if you have a German track, show the German credits, if you have both, show both and so on).

Deutsches DVD Profiler Forum: www.dvdprofiler-forum.de
DVD Profiler Unlimited RegistrantStar ContributorKulju
Registered: March 14, 2007
Finland Posts: 2,337
Posted:
PM this userView this user's DVD collectionDirect link to this postReply with quote
Quoting SH84:
Quote:

Only with an option for localization. Some movies have localized credits (especially animation movies). Perhaps it could be tied it to the audio tracks (if you have an English track, show the English credits and if you have a German track, show the German credits, if you have both, show both and so on).


I quote myself from another thread:

----clip-----
It just doesn't make any sence to do full Cast & Crew audit on Matrix when after that work there are still 80 releases of the same movie on db where your work doesn't effect. Of course I understand that it wouldn't be 100% perfect, but compared to currect situatiation I couldn't give a rat's arse if someone is actually credited Dave instead of David on a release I own. If I loose my sleep about that I could always do a local change and set needed locks.

It would be easy to deal with regional additions on Cast & Crew (animated films voices for example). There should be four tables for Cast & Crew.
1. Shared Cast (Type original credits here, shared with all other same film releases in db)
2. Shared Crew (Type original credits here, shared with all other same film releases in db)
3. "local" Cast (Type regional credits here, shared with other user but effects only to that release, like present Cast. )
4. "local" Crew (Type regional credits here, shared with other user but effects only to that release, like present Crew. )

User should be able to choose if he/she would like to have shared or local or both visible for him/her.

-----------clip-------------
DVD Profiler Unlimited RegistrantStar ContributorKulju
Registered: March 14, 2007
Finland Posts: 2,337
Posted:
PM this userView this user's DVD collectionDirect link to this postReply with quote
I most definately don't wan't to derail this thread, but in my opinion if we want better Cast & Crew db, many things should be fixed before we (and Ken) should put all the effort to actually start fixing Cast & Crew linking/parsing. To get the best end result I think:

First:
We should be able to make boxsets from releases that have more than one film on the same side of the disc. What would we get out of this in Cast & Credit wise? Cast & Crew would be separated to different profiles, not separated with dividers. This would lead us to second part.....

Second:
Movie based credits instead of DVD based. With first step done first we wouldn't have any problems with discs having more than one movie on one side of the disc. TV-series' Cast & Crew still had to be thought throughly.

After these two steps done Cast & Crew would be in completely different level. Sooooo much better....

Third:
After steps one and two, Cast & Crew should get unique IDs

At this point we would be near perfection
DVD Profiler Desktop and Mobile RegistrantStar Contributorhal9g
Who is John Galt?
Registered: March 13, 2007
Reputation: High Rating
United States Posts: 6,635
Posted:
PM this userEmail this userView this user's DVD collectionDirect link to this postReply with quote
Ken,

I  know that one of your reservations about using a "simple association linking system" has to do with linking people who use the same name.  Although the BY was a good concept in theory, it turned out to be quite unworkable in reality because of a lack of available BYs for many lesser known actors/crew.

I  had another thought that might be a viable solution for creating a unique key for cast/crew who use the same name but are in fact different people.

What if instead of using BYs to differentiate people, we used the title of the first entry in their filmography as listed on IMDb.  For all of IMDb's faults, filmographies seem to be quite accurate, and this could in no way be considered "scraping" data, so not a violation of their TOS.

This method would even work for a name like "Alan Smithee", where this alias was used by numerous different people.  For instance, Sidney Lumet used this alias when he directed 12 Angry Men. So the entry for this profile would be 'Alan Smithee (Crime Photographer)'.

Then a link would be created between 'Sidney Lumet' and 'Alan Smithee (Crime Photographer)'.

Arthur Hiller used 'Alan Smithee' when he directed Love Story, so for this profile, we would enter 'Alan Smithee (Matinee Theater), and a link would need to be created between the two.

This would link each instance of 'Alan Smithee' only to the actual person who used that alias for that film.

This still allows contributors to enter exactly what they see in the film credits without needing any "inside information" about the different ways people have been credited.  For those who don't know about other variations of the person's name, they just enter what they see.  Others, who recognize that the name should be differentiated and/or linked would be able to enter them initially and fix profiles that were entered by others.

At the time of implementing this change into the program, you could leave BYs in the database, and allow users to change them to 'Titles' as they edit/update profiles.  It is not a system that would need to be "fixed" wholesale upon implementation, but could be gradually changed over time.

Just something to consider.
Hal
DVD Profiler Unlimited RegistrantStar ContributorAddicted2DVD
Registered: March 13, 2007
Reputation: Highest Rating
United States Posts: 17,318
Posted:
PM this userEmail this userView this user's DVD collectionDirect link to this postReply with quote
I see how this would definitely be a better option then birth years. But I personally think we need to get away from using a year all together. Just for the plain reason that it is possible for the 2 same name people could possibly have the same year. We figured just how likely is this with the birth year and I seen it come up in the forums a few times since we had the birth year option. And even as little as it happens it still causes frustration for the people discovering it does happen.

So I feel we either need to get away from a year all together... or have a secondary way to differentiate between the two people that can be used when needed.
Pete
DVD Profiler Desktop and Mobile RegistrantStar Contributorhal9g
Who is John Galt?
Registered: March 13, 2007
Reputation: High Rating
United States Posts: 6,635
Posted:
PM this userEmail this userView this user's DVD collectionDirect link to this postReply with quote
Quoting Addicted2DVD:
Quote:
I see how this would definitely be a better option then birth years. But I personally think we need to get away from using a year all together. Just for the plain reason that it is possible for the 2 same name people could possibly have the same year. We figured just how likely is this with the birth year and I seen it come up in the forums a few times since we had the birth year option. And even as little as it happens it still causes frustration for the people discovering it does happen.

So I feel we either need to get away from a year all together... or have a secondary way to differentiate between the two people that can be used when needed.


Perhaps I wasn't entirely clear.

What I am proposing does get away from using a "year" altogether. 

It seems astronomically improbable that two same named people would have the same 'Title' as the first entry in their filmography.
Hal
 Last edited: by hal9g
DVD Profiler Unlimited RegistrantStar ContributorAddicted2DVD
Registered: March 13, 2007
Reputation: Highest Rating
United States Posts: 17,318
Posted:
PM this userEmail this userView this user's DVD collectionDirect link to this postReply with quote
Oh you said title!?! I  somehow read it to say year of title. Oh sorry about that!   
Pete
DVD Profiler Unlimited RegistrantStar ContributorKulju
Registered: March 14, 2007
Finland Posts: 2,337
Posted:
PM this userView this user's DVD collectionDirect link to this postReply with quote
Quoting hal9g:
Quote:
What if instead of using BYs to differentiate people, we used the title of the first entry in their filmography as listed on IMDb.  For all of IMDb's faults, filmographies seem to be quite accurate, and this could in no way be considered "scraping" data, so not a violation of their TOS.

You get my respect for out-of-the-box thinking, but I'm not sure if Ken is willing to base his DB's unique IDs to data from another web-site. I wouldn't. And I don't mean any legal issues. In my opinion Unique IDs should be based on something that you have control and cannot change. Info taken from any web site doesn't fall into that cathegory. I have nothing against IMDB eather.
DVD Profiler Unlimited RegistrantStar ContributorKulju
Registered: March 14, 2007
Finland Posts: 2,337
Posted:
PM this userView this user's DVD collectionDirect link to this postReply with quote
Quoting Kulju:
Quote:
In my opinion Unique IDs should be based on something that you have control and cannot change. Info taken from any web site doesn't fall into that cathegory. I have nothing against IMDB eather.

There are also a LOT of actors that IMDB doesn't recognize at all, even if they have been credited in many films for small roles.
DVD Profiler Unlimited RegistrantStar ContributorAddicted2DVD
Registered: March 13, 2007
Reputation: Highest Rating
United States Posts: 17,318
Posted:
PM this userEmail this userView this user's DVD collectionDirect link to this postReply with quote
Ken... you been pretty active here today... how about a little clue on how the new version will be handling cast/crew?

A lot of us are hoping for 1 database for both cast and crew together... so when we click on a name we will see both roles behind the camera and in front of the camera... chances?

What about the linking... anything fairly solid you can talk about yet?


Yeah... I know... curiosity killed the cat... but what can I say... it is my birthday on Tuesday.... and seeing some wrapped gifts around me is making me extra curios!   
Pete
DVD Profiler Unlimited RegistrantStar Contributorbizarre_eye
eye see you...
Registered: May 1, 2008
United Kingdom Posts: 503
Posted:
PM this userView this user's DVD collectionDirect link to this postReply with quote
*Please delete.*
DVD
Blu-Ray
LastFM
 Last edited: by bizarre_eye
DVD Profiler Desktop and Mobile RegistrantStar ContributorTheMadMartian
Alien with an attitude
Registered: March 13, 2007
Reputation: Highest Rating
United States Posts: 13,199
Posted:
PM this userEmail this userView this user's DVD collectionDirect link to this postReply with quote
Quoting Kulju:
Quote:
You get my respect for out-of-the-box thinking, but I'm not sure if Ken is willing to base his DB's unique IDs to data from another web-site. I wouldn't. And I don't mean any legal issues. In my opinion Unique IDs should be based on something that you have control and cannot change. Info taken from any web site doesn't fall into that cathegory. I have nothing against IMDB eather.

I was actually thinking the same thing.  Ken needs to create his own unique identifier.  Say we do use IMDb then, for whatever reason, they go dark.  Where would that leave us?
No dictator, no invader can hold an imprisoned population by force of arms forever.
There is no greater power in the universe than the need for freedom.
Against this power, governments and tyrants and armies cannot stand.
The Centauri learned this lesson once.
We will teach it to them again.
Though it take a thousand years, we will be free.
- Citizen G'Kar
DVD Profiler Unlimited RegistrantStar ContributorThe Movieman
DVDP User Since 2007
Registered: March 18, 2007
Reputation: Highest Rating
United States Posts: 2,544
Posted:
PM this userVisit this user's homepageView this user's DVD collectionDirect link to this postReply with quote
Kind of off-topic, but I could've sworn that when you look at a contribution for cast and/or crew that the names would be a different color between First, Middle and Last Names. But I just looked and it's all the same color (blue). I know within DVD Profiler itself it is color coded but I thought it was same online as well. Am I mistaken? Did I miss something in this or another thread?

Edit, here's a screen cap I took:


Like I said, I thought it was color coded especially since it was helpful when voting to see how somebody parsed a name...
 Last edited: by The Movieman
DVD Profiler Desktop and Mobile RegistrantStar ContributorTheMadMartian
Alien with an attitude
Registered: March 13, 2007
Reputation: Highest Rating
United States Posts: 13,199
Posted:
PM this userEmail this userView this user's DVD collectionDirect link to this postReply with quote
Quoting TheMovieman:
Quote:
Kind of off-topic, but I could've sworn that when you look at a contribution for cast and/or crew that the names would be a different color between First, Middle and Last Names. But I just looked and it's all the same color (blue). I know within DVD Profiler itself it is color coded but I thought it was same online as well. Am I mistaken? Did I miss something in this or another thread?

Indeed you did.  Start with Ken's post here.
No dictator, no invader can hold an imprisoned population by force of arms forever.
There is no greater power in the universe than the need for freedom.
Against this power, governments and tyrants and armies cannot stand.
The Centauri learned this lesson once.
We will teach it to them again.
Though it take a thousand years, we will be free.
- Citizen G'Kar
DVD Profiler Unlimited RegistrantStar ContributorThe Movieman
DVDP User Since 2007
Registered: March 18, 2007
Reputation: Highest Rating
United States Posts: 2,544
Posted:
PM this userVisit this user's homepageView this user's DVD collectionDirect link to this postReply with quote
Quoting TheMadMartian:
Quote:
Quoting TheMovieman:
Quote:
Kind of off-topic, but I could've sworn that when you look at a contribution for cast and/or crew that the names would be a different color between First, Middle and Last Names. But I just looked and it's all the same color (blue). I know within DVD Profiler itself it is color coded but I thought it was same online as well. Am I mistaken? Did I miss something in this or another thread?

Indeed you did.  Start with Ken's post here.


Ah, ok. Thanks. 
DVD Profiler Unlimited RegistrantStar ContributorCubbyUps
Registered: March 14, 2007
Reputation: Highest Rating
United States Posts: 4,245
Posted:
PM this userView this user's DVD collectionDirect link to this postReply with quote
Just adding my idea for creating unique ID's.


David Arnold

Basic ID would be 0401220904011814151104

Of course that's likely too long, so let's shorten it by adding up all the numbers. Which would equal 49. Now that would be too short and likely other names would total that amount.

So let's take 40 and add it to the first letters of both the first and last names.

So either something like D49A or 044901.

Middle initials or "Nicknames" would be ignored.

Now let's say that there are multiple David Arnold's in the database.

We could take an already Accepted BY and add it to the end of the ID.

For example an actor by that name has an Accepted BY 1949,
so we would have two David Arnold's in the database.

One with an ID of D49A or 044901.
Now a second with an ID of D49A1949 or 0449011949.

Now let's say that there is a third David Arnold and he has no known BY.

We could either use multiple sources to find their earliest role and add that in place of the BY at the end, or if not listed on those sites than use what is currently in the main database via the CLT, of course excluding the other David Arnold's credits.

So let's say that there are two more David Arnold's in the database.

One has a verifiable first role in a film made in 1967 and the second has a verifiable first role in a TV series made in 1978.

Those years would be added to basic ID for David Arnold.
1. David Arnold D49A1967 or 0449011967
2. David Arnold D49A1978 or 0449011978

In the rare case where a person has no known BY but has the same year of first role then perhaps something can be added to the end.

Perhaps something like
1. David Arnold D49A1978-A or 0449011978-A
2. David Arnold #2 D49A1978-B or 0449011978-B


Like I said just throwing this out there, it's very likely that this could be improved or made easier.
    Invelos Forums->DVD Profiler: Contribution Discussion Page: 1... 9 10 11 12 13 ...15  Previous   Next