|
|
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 2 3 Previous Next
|
CLT vs IMDB, sample size of ONE ... |
|
|
|
Author |
Message |
Registered: March 29, 2007 | Reputation: | Posts: 4,479 |
| Posted: | | | | Quoting Kathy: Quote: This is how I believe it should be, you type in exactly what you see. You see what you have learned to see. You see 0 as a letter in a word and as a numeral in a number. So in EVENEMENT (Event), you just see four E, and I see é, è, e, e because I have learned that this word is évènement and nothing else. You do not "see" electricity, or radio waves, does it mean that they do not exist ? The only problem is to enter correct data, and allow people knowing proper accentuation to correct false data. | | | Images from movies |
| Registered: May 20, 2007 | Reputation: | Posts: 2,934 |
| Posted: | | | | Quoting surfeur51: Quote: Quoting Kathy:
Quote: This is how I believe it should be, you type in exactly what you see. You see what you have learned to see. You see 0 as a letter in a word and as a numeral in a number. So in EVENEMENT (Event), you just see four E, and I see é, è, e, e because I have learned that this word is évènement and nothing else. You do not "see" electricity, or radio waves, does it mean that they do not exist ?
The only problem is to enter correct data, and allow people knowing proper accentuation to correct false data. There is only one way that this problem could even be remotely "Taken Care" of. The program would have to be rewritten, to allow one primary set of credits per movie. That primary set to be edited properly, Then that set be transferred to every profile of that movie. (a serious rewrite of the code) Otherwise, you or somebody would have to go through every profile with the affected names, and correct those names, in every locality. I seriously doubt that will ever happen, and as such will never solve this particular problem. I see this more as a program limitation, than a database limitation. I think Ken made the correct decision, within the confines of the current program set. A decision that can be applied to every profile, across all localities, without question. There is no insult. There is no mistreatment of actors, it is just data entry into a database. | | | Last edited: by CharlieM |
| Registered: March 29, 2007 | Reputation: | Posts: 4,479 |
| Posted: | | | | Quoting CharlieM: Quote: There is no insult. This is your point of view, with your culture. In my country, voluntary misspelling of names is an insult. ... and I'm not sure, when you write to your boss, that you add voluntarily spelling mistakes in his name to make him feel better with you... | | | Images from movies | | | Last edited: by surfeur51 |
| Registered: May 29, 2007 | Reputation: | Posts: 3,475 |
| Posted: | | | | Quoting surfeur51: Quote: Quoting Kathy:
Quote: This is how I believe it should be, you type in exactly what you see. You see what you have learned to see. You see 0 as a letter in a word and as a numeral in a number. So in EVENEMENT (Event), you just see four E, and I see é, è, e, e because I have learned that this word is évènement and nothing else. You do not "see" electricity, or radio waves, does it mean that they do not exist ?
The only problem is to enter correct data, and allow people knowing proper accentuation to correct false data. If I see an accent, I enter it. If there isn't one, I don't. Simple and can be done exactly the same way by everyone irregardless of language or culture. No one has to learn or understand anything this way. Edit: Of course I'm talking only about the online database. I am a firm believer that that needs to be as simple as possible. | | | Last edited: by Kathy |
| Registered: May 20, 2007 | Reputation: | Posts: 2,934 |
| Posted: | | | | Quoting surfeur51: Quote: Quoting CharlieM:
Quote: There is no insult. This is your point of view, with your culture. In my country, voluntary misspelling of names is an insult.
... and I'm not sure, when you write to your boss, that you add voluntarily spelling mistakes in his name to make him feel better with you... I know I am going to regret this but... Yves, what would you have us do? Even with the "common name" threads, we are not correcting the names that are known. There are thousands of names that do not link across the whole of the online db. There are a lot that are connected in personal DB's, but do not connect outside of that for the amount of research that would need to be done. Do you really expect the quality of the DB to get any better, with only a select few people trying to cover the vastness of the potential "problem" that you describe? Do you expect the DB to get any better by updated maybe 5000 of the profiles that need to be corrected (worldwide)? Do you even realize the scope of the correction that you are describing? Or do you expect a special rule for French Profiles, and a special rule for German profiles, and ....? How about the inult to the people of the countries that we cannot even put in the correct letter. That must be a real insult. How dare Ken insult those people. In all realism, the problem that you describe, will never be corrected to your satisfaction. There are too few people that are submitting profiles now. Most people that use the program do not even care about this discussion (not only in the US). This is not a US vs French issue. You may see it as an insult, there are others that may see it as an insult. If I were to address a letter or an email to an affected actor/ess, and misspelled their name, then they can be insulted (I would be). Again, There is no insult. It is Data into a database... |
| Registered: March 29, 2007 | Reputation: | Posts: 4,479 |
| Posted: | | | | Quoting CharlieM: Quote: It is Data into a database... The first thing to do when you enter data in a database is to give a proper format to this data, to allow linking, sorting and filtering, that are database functions. Invelos database is the only one I have ever seen that requests to enter voluntarily a unique name in more than one unique form, to avoid to have those database functions work correctly. I agree with you that now it is probably too late to try to correct everything. I wrote , when Ken made his clarification, and when, at that time, most of accented names were entered correctly, that it was a dramatic error to generate non existing variants. If once again I speak of this problem is because, once again, a thread was opened to criticize another database that resolved correctly this problem. We have no lesson to give to people who took the time to enter correctly names, when we do what we do. | | | Images from movies |
| Registered: May 20, 2007 | Reputation: | Posts: 2,934 |
| Posted: | | | | Quoting surfeur51: Quote:
If once again I speak of this problem is because, once again, a thread was opened to criticize another database that resolved correctly this problem. We have no lesson to give to people who took the time to enter correctly names, when we do what we do. 1. I don't remember a criticism, within this thread, except to the fact that there is still IMDB mined data within DVDProfiler. That primarily came from a time in which it was allowed. (though, i feel there are still people submitting such data) 2. IMDB is a different type of structure than DVDProfiler. IMDB only collects information from a movie for 1 listing (not tens/hundreds). Therefore, they do not have the same issues that DVDProfiler has. DVD Profiler was and is not structured as a movie DB (in the strictest since). Please note in one of my previous posts on how to solve this perceived problem. 3. I have been around since 2005, and I do not ever remember a point, where this was not an issue someplace. You may have had correct linking in your side of the profiler world, but never on this side of the world. 4. IMDB, still has not resolved many issues that I have with their DB, including but not limited to INCORRECT LINKING, improper crew rolls, improper cast rolls, improper listing of cast/crew members (use of middle initials where none exist, use of proper names where nicknames are used, etc) WOW, imagine the same problems we have with Profiler... I will say this again. IMDB has the same problem as DVD Profiler. Like it or not it is a HUMAN problem. IMDB has its uses and place within the movie information spectrum, just as many other tools do. We are not IMDB. If I wanted IMDB, there are other software packages that I could use (and have tried at least three others). I am relatively happy with profiler. It fits mine and many other peoples idea of a collectors DB. and for all of its shortcomings, it is still within the top of most rankings. (The biggest criticism is the lack of different media it can handle) I do understand your criticisms of profiler, for I have a few myself. In the end it works for you and it works for me. We perceive this differently and in our locals we make it ours. Still no criticism, no intended insult, just data in a DB | | | Last edited: by CharlieM |
| Registered: March 29, 2007 | Reputation: | Posts: 4,479 |
| Posted: | | | | Quoting CharlieM: Quote: I will say this again. IMDB has the same problem as DVD Profiler. Really, I cannot understand your point of view, as you probably cannot understand mine. I have quite never seen the same actor with two different names in IMDb, but I admit that perhaps some obscure cast or crew can be in this case. In Invelos, we have top actors , having played in more than 100 movies, that have non-linking variants generated by our rules. This is not a human problem, as rules request to enter errors, even if we know the correct name. And this is not because of multiple region entries for the same movie , but even inside the same region for different movies : even if we had only one profile per movie, Déborah François in Movie 1 would not link with Déborah FRANCOIS in movie 2 and with DEBORAH FRANCOIS in movie 3, because of our rules. I agree that roles in IMDb are often different than the role in credits, but for me this has strictly no importance as long as we can get for each actor the list of movies in which he played. So same problem for you, very different one for me, we'll probably never agree even on the diagnostic of the situation. | | | Images from movies | | | Last edited: by surfeur51 |
| Registered: May 20, 2007 | Reputation: | Posts: 2,934 |
| Posted: | | | | Quoting surfeur51: Quote:
The first thing to do when you enter data in a database is to give a proper format to this data, to allow linking, sorting and filtering, that are database functions. Invelos database is the only one I have ever seen that requests to enter voluntarily a unique name in more than one unique form, to avoid to have those database functions work correctly.
I don't know how I missed this. The first thing in a db, is you allow consistent input in a consistent format to allow linking sorting and filtering, by the people that use the DB. You design it, so that it doesn't matter your background, culture or heritage, you can enter it into the DB. Quote: even if we had only one profile per movie, Déborah François in Movie 1 would not link with Déborah FRANCOIS in movie 2 and with DEBORAH FRANCOIS in movie 3, because of our rules. Our rules allow for this already. It is not the perceived misspellings that is the issue, it is the data entry that is a problem. As long as I know these are the same person, just a different entry for them, I can link them through the DB. Ken gave us this ability. And believe it or not, it doesn't matter whether I understand the idiosyncrasies of the french language, or any language for that matter. Unfortunately, it may not be the common name that you desire or think should be correct because of outside factors from DVD Profiler itself, but it would be a common name, that is consistent. Do not get me wrong, I think the way linking and common names is done, is a bit awkward. I have put forth idea on what I think it should be. It is up to Ken to decide what changes he would like to make. I only work within the environment that I have been given. If people would follow the rules, and enter it per the rules, then our CLT would be correct, and then we could allow proper and easy linking within the confines of our system. (one of the items that really needs work is original title, for the title inconsistencies are throwing things way off) When we have people entering data, that is obviously against the rules, then YES we have a HUMAN problem. Until you get beyond the fact that what you call an error or a misspelling, is just a format for entering data within guidelines, then our discussion and debate will never connect. There is a solution to the problem, and unfortunately it is in the Data Entry, not in the rules |
| Registered: March 29, 2007 | Reputation: | Posts: 4,479 |
| Posted: | | | | OK, you are right... common name is the best solution to link invented names. A pity that nobody asks why is it useful to invent names... And can you explain me, since you completely ignore that DEBORAH FRANCOIS is something else than Deborah Francois ( And believe it or not, it doesn't matter whether I understand the idiosyncrasies of the french language, or any language for that matter.), how you'll have the idea to look at CLT for Déborah Francois and Déborah François to know the common name? | | | Images from movies | | | Last edited: by surfeur51 |
| Registered: May 29, 2007 | Reputation: | Posts: 3,475 |
| Posted: | | | | Invented names? I don't understand what you are talking about.
I take the names from the credits and put them in exactly as I see them. I don't invent anything.
Edit: I came back to address the CLT question but CharlieM has explained things quite clearly - thank you. | | | Last edited: by Kathy |
| Registered: May 29, 2007 | Reputation: | Posts: 3,475 |
| Posted: | | | | Quoting surfeur51: Quote: Quoting Kathy:
Quote: This is how I believe it should be, you type in exactly what you see. You see what you have learned to see. You see 0 as a letter in a word and as a numeral in a number. So in EVENEMENT (Event), you just see four E, and I see é, è, e, e because I have learned that this word is évènement and nothing else. You do not "see" electricity, or radio waves, does it mean that they do not exist ?
The only problem is to enter correct data, and allow people knowing proper accentuation to correct false data. Once again you are applying language and culture attributes to data. What is "correct data" to you might not be the "correct data" to others. Here is a perfect example: Which is "correct data"? Color or Colour? The answer is simple - they both are. So, what is the simple solution? Type in the data as you see it. |
| Registered: May 20, 2007 | Reputation: | Posts: 2,934 |
| Posted: | | | | Quoting surfeur51: Quote: OK, you are right... common name is the best solution to link invented names. A pity that nobody asks why is it useful to invent names...
And can you explain me, since you completely ignore that DEBORAH FRANCOIS is something else than Deborah Francois (And believe it or not, it doesn't matter whether I understand the idiosyncrasies of the french language, or any language for that matter.), how you'll have the idea to look at CLT for Déborah Francois and Déborah François to know the common name? Exactly the same way that I know George Zephyr is the same as Georgi Zlatarev which is the same as George Zlatarev, which is the same as Georgy Zlatarev, which is the same as Joro Zlatarev. I use other tools at my disposal. I still follow the rules and make DEBORAH FRANCOIS, Deborah Francois. | | | Last edited: by CharlieM |
| Registered: May 20, 2007 | Reputation: | Posts: 2,934 |
| Posted: | | | | Quoting surfeur51: Quote: OK, you are right... common name is the best solution to link invented names. A pity that nobody asks why is it useful to invent names...
And can you explain me, since you completely ignore that DEBORAH FRANCOIS is something else than Deborah Francois (And believe it or not, it doesn't matter whether I understand the idiosyncrasies of the french language, or any language for that matter.), how you'll have the idea to look at CLT for Déborah Francois and Déborah François to know the common name? I have tried to be polite, and you are coming off as sarcastic. Obviously you are not reading or comprehending what I am saying. I never said that common name was the best solution for any linking issue, as a matter of fact, I said quite different. I also do not "invent" names. I follow the rules and guidelines for the DB that I work in. I work in many Databases, and am forced to follow DB guidelines in all of them. I format my data, per the rules that are given to me (work with what I am given). I have, multiple times, explained myself, as others have. Do I think that the data that is being entered is necessarily "Proper" as you define, which means correct spelling according to culture and nationality? No i do not. Do I claim to know the proper spelling of any actor born outside the US? no I do not. What I do claim, is that the data I enter follows the format and guidelines given to me by the program and the rules. What I do know, is that if everybody did this exact same thing (from all parts of the world), our online DB would be in much better shape (It may not be proper spelling, but it would be tons better) If I were expected to know the ins and outs of every nationality of every actor that I deal with within my db, I would quit using this program entirely. I would need to know French, German, Greek, Chinese, Japanese, etc. etc. and it would take the enjoyment out of what I do. And to top it off, because of the data structure associated with DVD Profiler, there is no way for you to effectively correct even just the French cast and crew of my personal DB. I am sorry that you seem to take it as a personal affront, that this program and rules do not conform to your ideas of correctness. Charlie |
| Registered: March 29, 2007 | Reputation: | Posts: 4,479 |
| Posted: | | | | Quoting Kathy: Quote: Here is a perfect example: Which is "correct data"? Color or Colour? The answer is simple - they both are. Is that true for both English and American pupils during a dictation at school ? Colore and Farbe are also correct data, but in Italian and German, not in an English text. The problem with names, is that we are not supposed to translate them, and the spelling of the concerned persons' language applies. | | | Images from movies | | | Last edited: by surfeur51 |
| Registered: March 20, 2007 | Reputation: | Posts: 2,851 |
| Posted: | | | | Quoting surfeur51: Quote: The problem with names, is that we are not supposed to translate them, and the spelling of the concerned persons' language applies. The problem with names is that some people become obsessed with them. --------------- |
|
|
Invelos Forums->DVD Profiler: Contribution Discussion |
Page:
1 2 3 Previous Next
|
|
|
|
|
|
|
|
|