Robert Jan van der Waals
I need some input from you for the whole "how do we format a person's name" in Data Crow.
Currently, as you can see, the names are now "Lastname, firstname" and persons have a checkbox called "Company / Organization" to allow for "unreversed" names (Firstname Lastname). This has caused quite a number of emails were persons have found the band names to be reversed as well. I do understand this is unwanted behavior but unfortunately an online service (Amazon, Musicbrainz, etc) does not supply informatie on whether the artist is a band or a person. Possible solutions:
- Leave the all artist name unreversed by default (setting the check "company/organization" (which will be renamed to band). Sort on the full name (whch is potentially the firstname in case we are dealing with an actual person).
- <your solution goes here, please>
So as you can see this also has to do with the way these are sorted….. this is how this whole mess started to begin with :)
Your humble programmer,
I don't have any new suggestion for you, unfortunately, but I just wanted to share my two cents on the subject by saying that I've never been one for using lastnames first or writing "Beatles, The" - so for me, the best thing would be to avoid reversing anything in the default setting. For me, both people's names and band names should be written in the order that they were originally meant to be written (i.e. "Firstname Lastname" and "The Beatles"). I know that many people would disagree with me on this, but that's how I think things are best represented: in the order they were first intended to be presented.
Forgot to add (and I don't seem to be able to edit the post):
The same thing occurs with actors names, but on a much larger scale. On the music side, what I did as a stop-gap was to export the Firstname,Lastname and company fields as tab delimited to a spreadsheet, change the Company field to read all True, create a column called Name, and created a calculation cell adding Firstname and Lastname together, and then imported them back into Data Crow. For my music collection this worked well as the collection was not very large. However, trying the same approach for my movie collection is a HUGE job involving about 40,000 actors.
If there is an easier fix, I'd sure like to know.
The more I think about this issue - I have re-read your post many times now - the more I believe that maintaining a good solution for people who want their names in reversed order is going to be extremely hard to do.
In thinking of a solution for those who like "reverse order" listings of names, I have thought of different approaches, but none of them seem very good. For example, if the user wants names in reverse order the program could:
A) Treat every word except the first word as belonging to the surname (last name). But that won't work because many people have multiple given names (first names).
B) Default to taking only the last word as the surname, and add a list of "special words" that also should be treated as belonging to the surname, so that both "van" and "der" could be placed together with "Waals" as the surname if one's name was "Robert Jan van der Waals". However, maintaining such a list of special words that should be treated as belonging to the surname would be impossible to do, since you'll have people with dual last names (sometimes without hyphens), such as "Carl Johan Andersson Smith."
If no info is given at the source, and all you have is "Carl Johan Andersson Smith" - as one string - you could certainly TRY to figure out that this is a person and not a group, and you could TRY to guess that "Andersson Smith" make out the surname, but it will only ever be just that - a guess.
Again, for me, the most natural choice would be to not reverse anything, so you get firstname first and lastname last.
Also, I had to see the setting in Datacrow, and read this post of yours, several times before I understood that "company/organization" had something to do with not reversing names. It does not feel intuitive to use something called "company/organization", (or "band" as you were planning to re-name this setting to) - when I'm looking for a way to "not reverse names!" Instead, displaying all names of people and bands in their correct, unreversed order as the default and then have a checkbox to choose "Reverse names" - for those who want that - would be a much better way to go, in my opinion. (even if you still would like to have reversed names as the default - it would be an improvement to change the checkbox name in this way, I think)
Perhaps this can be defined as a check-box or a text field in the Artists list (if we're talking about the audio module);
Let say that a check-box would mean "reverse name" so datacrow will automatically check the box if there is a word such as "The" in which band by the name of "The Beatles" will default to "Beatles, The".
Since this is a database, there need to be some input from the user so maybe at the beginning the database will put names as they are (unless there are keywords such as "The" that the user define) but once you get the first name setup, whenever datacrow will import that name, it'll search the artist field and check if the user prefer it to be reversed or not (if nothing found, an entry is made and the user can change it by the check box or even a text field that says "display this artists name as…")
I hope I explained myself OK and it's not too confusing - if so, I'll gladly try to explain again.
Robert Jan van der Waals
1 - In the next upgrade where you will be greated with a message to change the default behavior for each person module.
2 - I will remove the "is company" checkbox to avoid confusion.
3 - The setting can be changed in the way the title rewriter works for movies, software, etc. It will also allow you to exclude certain keywords (The, a, etc) from reversal. This way you will be ablt to, at once, rename / reverse all artists/actors/etc names.
This is what I have come up with after recess of a bout 2 months :-)
What do you think? I would like to start and release this within 1-2 weeks..
I apologize for my late reply.
I just wanted to say that I support the solution you described above.
This way the wording of the option (to have reversed or unreversed names) might become more intuitive, and users can easily choose between having all persons' names reversed or unreversed.
For me, who will probably always want to have people's names unreversed, I believe everything will work flawlessly the way you described. However, if the source of the information that holds a person's name does not clearly mark what is the first name and what is the last name, I'm sure it won't always be easy for those who would like names to be reversed. For example, the made-up name I mentioned above, "Carl Johan Andersson Smith"… - I guess it'll be impossible for any program to figure out that "Carl Johan" make out the first name(s) and "Andersson Smith" the last name(s). Not much to do about that, I believe.
Two things in closure:
1) THANK YOU so much for sharing this great program! It might not be the prettiest looking program, and it is quite slow - but I love the many features it holds!!! Hopefully, with the speed-increase you mentioned for version 4 (using the new Java version), the program will become even better and more usable on slower machines (like a HTPC). I'm excited to hear that a new version is in the works!
2) If you would like any help trying things out, I would be glad to try and help out as a sounding board by trying things and giving feedback. *I guess I could just keep an eye out for when new betas are released and provide feedback on them…* And if you want to contact me about anything that I could look at, please feel free to do so.
first many thanks for sharing this program. I love it for all the possibities to customize to my needs.
I agree with Carl, that no algorithm can do the job to always guess correct what part of name a word belongs to.
So I think your suggestion is the best choice to fit for most needs.