It's a little frustrating that it is so easy to correct casts for television episodes and so hard to correct them for movies. I submitted 3 cast corrections for Harriet to match the onscreen credits and 1 to make an uncredited member follow the format of the credited casts. The submission was 200422-193844-431000.
First, Antonio J Bell is credited as "Antonio J. Bell" with the period as you can see below. This one baffles me because I am easily able to add this attribute for him in television episodes.
The next one is a complete obmission. William L. Thomas is credited as "Senator Seward" in position 39 (which there is currently no actor with that order number). Also, in the picture below LondonRose Sellars is credited as "London Rose Sellars" with a space between "London" and "Rose".
Finally, there is an uncredited cast member who's credit looks like this:
DeVaughn, Arykah Naomi Angerine 2yo/3yo (uncredited)
As you can see above, this movie lists ages just in parenthesis so it should look like:
DeVaughn, Arykah Naomi Angerine (2/3) (uncredited)
First, Antonio J Bell is credited as "Antonio J. Bell" with the period as you can see below. This one baffles me because I am easily able to add this attribute for him in television episodes.
The next one is a complete obmission. William L. Thomas is credited as "Senator Seward" in position 39 (which there is currently no actor with that order number). Also, in the picture below LondonRose Sellars is credited as "London Rose Sellars" with a space between "London" and "Rose".
Finally, there is an uncredited cast member who's credit looks like this:
DeVaughn, Arykah Naomi Angerine 2yo/3yo (uncredited)
As you can see above, this movie lists ages just in parenthesis so it should look like:
DeVaughn, Arykah Naomi Angerine (2/3) (uncredited)
Adrian, Champion
Posted 3 months ago
Michelle, Official Rep
Phil G
If this was just about parentheses, I wouldn't bother making a fuss. But the slash is much more important: getting that wrong means actively corrupting the database. Perhaps I should have explained this more clearly last week.
From the guidelines:
This is not just a vague convention that can be followed or ignored as contributors feel like. The slash has a very clear and significant meaning to IMDb, and the software processes credits differently based on that. When a slash appears in a character name, the software splits the credit into distinct characters, and processes and stores each character separately.
When a slash is included for any other reason, it creates two (or more) distinct characters where there should only be one: not the intended result, and this data corruption causes various further problems.
The issue manifests in many ways across the site, some more subtle than others, but is perhaps most clearly demonstrated in the datasets. For example, in title.cast.tsv.gz, the line for the credit discussed above (pre-update) looked like this:
Notice that the credit is treated and stored as two distinct characters: anyone using the datasets to query how many characters this actress played immediately gets the wrong answer. And the implications of this data corruption go much further than just the datasets. Try adding a quote involving this character and because of this bad data, you'll have to jump through extra hoops to get the character name included as intended. That's just for starters; I won't try to list all the problems that can be caused by this.
Adrian is quite happy with the way this credit currently displays on the site, but I maintain that this is by luck, not by design. If IMDb decides to change the display logic for multiple roles, then this credit could easily be messed up. For example, in future, characters could be listed in alphabetical order; or each character on a separate line rather than side-by-side. How would this credit look then? Including a slash in the character name, tells the software that these are distinct characters that can be handled separately. Therefore, the software has no obligation to always display them side by side, in the same order, with a slash between them. It's free to display one or both, together or separately, in any order, with anything or nothing between them, or even in different colours, etc. When the underlying data is wrong, correct display cannot be guaranteed.
Adrian, Champion