Quoting Prof. Kingsfield:

Quote:
but i also don't accept an undocumented autograph as any sort of proof

It's not about that. It's about the simple fact that she's mostly credited as "Zoë Saldana". It's lovely that that happens to be her "real" or "correct" name as well, but that really doesn't matter. I just included the pic because it was used earlier in the thread.

With a direct linking system, this problem could have been completely fixed in 2007 with a single entry linking Zoe Saldana and Zoë Saldana.

No fixing and contributing a correction for every on-line profile.  No downloading corrections (for everyone else except the contributor) for each and every profile containing this actress.

Correcting the actual entries in all of the profiles could have continued as they were gotten to, but in the meantime, everyone would have been enjoying actual linking of all of their profiles.

I hate to say it, but years after implementing the "Credited As" system, linking still sucks!

Quoting hal9g:

Quote:
With a direct linking system, this problem could have been completely fixed in 2007 with a single entry linking Zoe Saldana and Zoë Saldana.

No fixing and contributing a correction for every on-line profile.  No downloading corrections (for everyone else except the contributor) for each and every profile containing this actress.

Correcting the actual entries in all of the profiles could have continued as they were gotten to, but in the meantime, everyone would have been enjoying actual linking of all of their profiles.

I hate to say it, but years after implementing the "Credited As" system, linking still sucks!

I agree, hal, I still think simple association is the best method. I have seen nothing in the current system to change that opinion in the slightest, and i have tried.

A little suggestion: common name = IMdb name (except when roman numbers, for them still use birth year date, even fake if necessary). Sure it will improve quickly the linking, with a minimum of contributions to change existing data...

Quoting Prof. Kingsfield:

Quote:
I think the problem that you talk about is simply because people are not REQUIRED to actually look at the data

In part, perhaps, but I don't think Invelos can force this kind of research upon each and every user. I know you certainly wouldn't like to have that forced upon you, for instance. There is, however, a big difference between not being REQUIRED to drill down the CLT results to find the actual common name, and actively changing existing "Zoë Saldana [Zoe Saldana]" credits to "Zoe Saldana" credits - which was the cause for this thread back in 2007. That was a decidedly incorrect change, one that will now need to be changed back, and that pointless back-and-forth really wouldn't have been necessary if those contributors had actually looked at a few credits rather than taking the hugely flawed CLT numbers at face value. Again, not something I think we can require of everyone, because it's hard and tedious work, but you'd sure hope that the people who do take the trouble to tinker with such a cast entry, might look a tiny bit beyond just those numbers, or at least listen to those that have done the work, instead of blindly propagating the incorrect name variant for several years simply because the batch of IMDb-mined data in the database says so.

Quoting surfeur51:

Quote:
A little suggestion: common name = IMdb name (except when roman numbers, for them still use birth year date, even fake if necessary). Sure it will improve quickly the linking, with a minimum of contributions to change existing data...

My impression is that imdb usually goes for the longest name (all initals, middle names etcetera), even if the mostly credited name is different. However, we can hardly expect contributors to go hunting for absent name parts, so the credited name must be the starting point. CLT is an attempt to make linking work in case of variations, and with all its flaws, it does a good job in many cases.

Quoting surfeur51:

Quote:
Quoting Prof. Kingsfield:

Quote:
Ummm Yves, this is DVDProfiler not IMDb, we don't need to be a copy of them in ANY way, if we are then there is no need for the existence of Profiler. After all all of your precious IMDb data already exists...right.

dvdprofiler is a program, with no correct online data (I'm sure that many local data are of a high quality).
IMDb is a database, with no program to manage a personal collection.

To speak of dvdprofiler versus IMDb is totally inappropriate.

Not true, you may not be able to do it the way you would like, Yves, but you CAN track up to 10,000 titles there for free. But then you don't like profiler's standards, and you claim that our data is not correct, not correct to YOU, correct to the film's data then yes we are far more accurate and always will be. I don't know what the answer is for you, IMDb does not give you what you want in terms of management and since profiler has standards that do not allow for your version of "correct" then you don't like profiler. If you try another program you will find shortcomings there as well...and that will always be true. I don't know what to tell you, Yves...maybe try and write your own, then you will be happy because it will do what YOU wish it to. And if you sell it, you will, just as Ken does, have users that are not happy with the way YOU wnat to do things, and then you will have to listen to your own version of surfeur51, who simply wants things done his way and refuses to try and understand  anything.

Quoting T!M:

Quote:
Quoting hal9g:

Quote:
I hate to say it, but years after implementing the "Credited As" system, linking still sucks!

Well, in my local database, it's a work of art, but as far as the online is concerned, we're faced with thousands of situations exactly like this one, where people insist on propagating incorrect IMDb-names for years just because that's the name variant the CLT erroneously awards the highest number to. If we ever get the batch of IMDb-mined data in the database down to about 10 or 20%, then the CLT numbers might start to give somewhat of a hint towards the common name. Alas, as just the number of incorrect, brand-new 'Avatar'-profiles show, fresh IMDb-data is still dumped into the database on a daily basis... 

Well, it's a broken system when each and every user has to go through what you have to make it "a work of art".

Most are not going to go to that much trouble.  Nor should they have to.