I happily use both IMDb and Profiler but for quite different ends. They are after all fundamentally different by design. One is designed from the beginning to handle movie data, mainly cast and crew but also many other film facts and trivia. The other is designed around profiling mainly the technical details of individual discs or rather releases. Discs can be the same for different releases but no action is taken by the program to link the content even when the discs are identical. Instead everything is left to the user. I think the increasing problems to maintain the integrity of the database have come from the fact that we are trying more and more to profile movies but we lack the efficient tools for it. We basically have a database without interrelations except for very rudimentary cast and crew linking. I rather liked the days when crew was limited to director only, at least it kept the problems to a minimum. And people would use the program for its intended purpose and not complain because it wasn't IMDb. 

Quoting MikaLove:

Quote:

Anyway this discussion is less about iMDB and more about finding a good and final solution to the problems with name variations in the database, which currently is one of the absolute greatest and bothersome flaws about DVDP.

It's not really a problem if you're not trying to use Profiler as a substitute for IMDb. I regard it as a "bonus" if it works, but it's not my main motivation for using the program. I try to work with what we have, that's the only part of it I can control. And what we have was not built to handle the things you demand of it. I'm afraid that's going to spell disappointment, but let's try and maintain a little realism, it's not all going to change anytime soon. If it was that easy, it would already have been done.

I don't feel that I'm losing anything, since it was never part of the deal. But you can get all this today, it just requires more work than a simple copy & paste job. Linking has actually improved lately since the local linking now seems to ignore name parsing. And the contribution filters have also helped in this matter to standardize some of the names. So there *IS* progress, but it's not going fast enough for everyone.

@Kinoniki:
Well what about all the work that has already been done, but which isn't correct?
And all the coming entries that won't be correct either.
People submit them all the time and they even slip through and get approved.
It's weird how easy it is to submit erroneous data and still both get all positive votes and a final approval.

@TheMadMartian:
Well, I wasn't only talking to you, here.

While the CLT may be flawed...it is flawed because of us, the users.

Far too many people submit IMDB or personalised data to the online database while claiming that it is from 'the credits'.
I've recently seen a couple of films where the contributor has stated: "Cast and crew added. Taken directly from film credits"...when, in fact, this information was taken from IMDB (how do I know this you ask?....well....it certainly didn't match the credits (which I was looking at) and when I checked IMDB it was identical).
The two people who contributed that data are now on my 'do not trust' list and I will always check their submission or never download their submission. Harsh, I know, but the only way to protect my database.

If everyone had copied the credits (and, unless I'm mistaken, the rules have always stated to use the credits...even when they were guidelines), we wouldn't have the problems we have.

So all we can do is use the CLT, identify and try to correct it's IMDB mistakes, and slowly get to where we want to be.

What I would like is for Ken to implement an automatic check (invisible to users) that would scan a contribution and compare it to the IMDB page...and automatically reject it if the data is identical (I have NEVER NEVER NEVER seen an IMDB page that matches the onscreen credits).
Or, if it doesn't reject it completely, puts it on hold until the user/contributor or another independent user can verify that the information IS from the credits.

I would also like users to stop copying cast/crew from other accepted profiles without checking them. I've seen far too many updates where the data has been copied but doesn't match the credits...this is just making a bad situation worse.

IMO, if you can't be bothered to check the credits then don't submit the data.

I don't want IMDB rubbish in my local db.

The CLT will work if all users work together to make it work.

Also...and lastly....Ken should reintroduce the contribution test we used to have that will stop new users making common mistakes when they start contributing.

P.S. When reviewing contributions I click on the IMDB link through Profiler and can see, at a glance, if the information being submitted is from IMDB or not.

The reason why I want linking rofiles in DVDP and not just in in IMDb is because DVDP represents what I own. I double-click on a name and it shows me that Person X was in movie Y and Z and also in the TV Show A, episode B - all of them in my collection.

But that only works if the linking works. And it can't work if the key to the linking (in this case the common name) is in constant flux. It's not just because people change their names, it's also because there's so much legacy data in the database (and also because of spelling issues like the comma before the junior).

Computers are dumb. They compare "A" to "A" and say it's identical and they compare "A" and "a" and they say they aren't - unless you teach them that it might be the same in this special case.

Even if one day all the "The Rock"s, the "Dwayne 'The Rock' Johnson"s and "Dwayne Johnson"s have one unique common name of "Dwayne Johnson" - in profiles splattered over X different DVDs in Y localities and Z different movies, the minute that guy marries Mrs. Gerald and calls himself "Dwayne Johnson-Gerald" and makes a ton of movies under this name the game starts anew. You have to touch all profiles in all localities all over again.

What you need is a constant (like Desmond ) and not a variable. It does not really matter what that constant is: roman numerals, japanese kanji or kryptonian symbols as long as it is A CONSTANT. A constant that links all the different names that belong the same person together.

Until then the system is broken and will always be. Regardless of how much effort you put into it.

Agreed on the constant.
Even though the only issue that will be solved with this is the possible shift of common names. It will not solve the problem of correct linking. Means you'd still have to find out if John A. Doe is a variant of John Doe or if he is a different person.
I'm not sure if we are ever going to find a satisfactory solution for the linking as such.

Quoting AESP_pres:

Quote:

3. the Internet adult film database,
4. the Internet movie database,

While I love iafd as well, they are far from perfect and I find more errors there the more classic porn I download. For a miniute I was submitting scene and cast lists to them, one showed up online then they stopped acknowledging me ever since.   I tried!

I also recently signed up for imdb and tried to correct a profile there that so FUBAR. Of course I forgot which one that was so I can't see if it helped or mattered, but it led to a fun thought....

I've noticed many imdb profiles are closer to the real credits lately, so lets all sign up there and when we submit a profile here, we can copy the info and submit it there too. We can take over imdb and make it right!