Correcting entries

ConversazioniManga!

Iscriviti a LibraryThing per pubblicare un messaggio.

Correcting entries

Questa conversazione è attualmente segnalata come "addormentata"—l'ultimo messaggio è più vecchio di 90 giorni. Puoi rianimarla postando una risposta.

1mikeneko
Ago 12, 2006, 6:53 pm

Many people who entered Japanese titles (including manga) before the last server crash have corrupted entries in their catalogs. It seems a lot of people missed the fix for this, so I'll quickly post it here.

Over on the Google Groups forum, there's a link to a screen for retrieving UTF-8/unicode information from the old database. Read the directions there, give it shot:

Unicode, Unicode, Unicode.

In my own case, the Summary data was usable, and I could copy/paste from it to correct the Author and Title fields. (Your mileage may vary, naturally.)

2mikeneko
Ago 12, 2006, 7:11 pm

While I'm at it . . .

There's a known bug that affects only data imported from Amazon.co.jp, namely, the ISBNs do not transfer into your catalog. You have to input them *manually.* (When will this be fixed? Dunno. You might ask on the Google Group . . .) So that's why covers for these imports won't show up in your catalog unless you force them.

When adding books:
1. Add books screen: Type in ISBN to search Amazon.co.jp. Results appear on the right sidebar.
2. *Before* you click on the proper result, highlight and copy the ISBN you just typed to your clipboard.
3. Click on the link to add the book to your catalog.
-----
4. Click on the book's pencil icon to go the Edit screen.
5. Scroll down to the (empty) ISBN box and paste the ISBN from your clipboard.
6. Click the Submit button.
If there's an Amazon cover, it should show up now.

For fixing manga that are already in your catalog, start at Step 4.

If you don't feel like dragging the book out again to find the ISBN, you should be able to copy/paste it from the manga's Information screen:
1. Click on the book's Information icon (the card).
2. Under "Editions, by Popularity" on the left sidebar, you should see the manga's ISBN. Highlight it and copy it to your clipboard.
3. On the right sidebar, click on the Social Information icon (a people head), then click on the pencil icon to go to the Edit screen.
4. Paste the ISBN into the proper box; click Submit.

3The_Holy_Terror
Ago 13, 2006, 1:23 am

I noticed today that you can combine authors' works if their names are spelled wrong or if it's written with the family name first on one volume and the given name first on another, etc. I don't know if this is something new or if I just never noticed it. This works especially well when Amazon screws up.

The thing that's kind of frustrating is you can't combine a name written in kanji with one in romaji. I've got the first volume of yotsubato! on order and I was going to combine it with the english version. But I noticed that I'd have to retype the manga-ka's name.

Now I don't have that many imports so it's not that much work, but I was looking at your collection mikeneko, and it'd be a pain to retype all of those names.

Oh well.

4trollsdotter
Ago 13, 2006, 1:44 am

I was very relieved to find the combine authors feature, especially since it is hard to tell sometimes which is the family name. The publishers are contributing to my confusion--Tokyopop has Hee-Joon Son's name differently on PhD and Id_Entity, and CMX started out using family name first and then changed. I've been using the Animenewsnetwork Encyclopedia to verify name order and hoping they are correct. Since other catalogers aren't as concerned, the combine authors is useful.

I'm adding english translations for the few Japanese language books that I have, since I don't read Japanese. That helps with searching and sorting.

5mikeneko
Modificato: Ago 13, 2006, 12:51 pm

Yup, combining is an option, though the rules can be a bit tricky. (The "Combiners" group obsesses over this particular topic.) Because U.S. Amazon usually provides *wrong* information about the writer and volumes with manga/manwha (e.g., misspelling names, giving the English translator or editor as the author), those really ought to be compared with your actual manga and corrected after you've imported them. Hardly anyone bothers to correct the bad information, so you'll lose some of your social hits after fixing your own entries.

Even so, when someone else has a wrong name listed in the "Author" slot manga, that entry still shouldn't be combined under the proper author's name -- it just has to be left orphaned. Here's an example of what I mean. Volume 1 of the English translation of the Korean manwha "Demon Diary" (TokyoPop) has been listed three separate ways by LT users.

Correct: Demon Diary 1 by Lee Chi Hyong (writer, vol. 1)
Incorrect: Demon Diary 1 by Kara (artist)
Incorrect: Demon Diary 1 by Jee-Hyung Lee (writer, vols. 2-7)

For the whole series, you'll find:
Demon Diary by Lee Chi Hyong
Demon Diary by Lee Yun Hee
Demon Diary by Jee-Hyung Lee
Demon Diary by Kara
Demon Diary by Shotaro Ishinomori

* Lee Chi Hyong and Lee Yun Hee CAN'T be combined, even though they were writers on the same series.
* Lee Yun Hee and Jee-Hyung Lee are the same person; they CAN be combined.
* Kara and Lee Yun Hee CAN'T be combined, even though they worked on the same series.
* Shotaro Ishinomori CAN'T be combined with anything. (I have no clue where this one comes from -- he's not listed anywhere in the publishing information for this manga.)

Wug! Anyway, LT preference for authors is this "LAST Name, First Name", which flips as "First LAST". However, Amazon.co.jp imports always list "First Name, LAST Name", which then flips to display as "LAST First" (i.e., the proper _Japanese_ name order). This can get a bit confusing . . . I've decided to leave the Amazon.co.jp imports as they are and give each author his/her own tag to keep them properly sorted. :P

Something that a few people are trying is to use the mangaka's name in romaji in the author slot followed by the kanji/kana in brackets or parentheses (). This does get you a combinable combination for social purposes, but I'm not certain how this interacts with the different options for list-view that flip the first and last names around the comma. (I haven't tried it myself.)

6trollsdotter
Ago 13, 2006, 2:00 pm

Interesting choice of example. I gave it a lot of thought and chose to use Kara as the author of the series and put "Jee-Hyung Lee, Chi Hyong Lee" under other authors. I chose this method because this is how I shelve the series. I may have reversed it if the series had been written by the same author--or maybe not.

"Correct: Demon Diary 1 by Lee Chi Hyong (writer, vol. 1)"

If I used the writer as author, I would have this as: Demon Diary, Vol. 1 by Chi Hyong Lee

7mikeneko
Ago 13, 2006, 3:10 pm

It's a problem, isn't it? I'd initially opted for Kara as author, but later decided to change it because LT in general gives preference to writers over artists whenever there's a combination (like with comics). Neither way is a win-win situation for this series; the same issue has been chewed over in the Comics group as most American comics have the writer/artist splits.

If I used the writer as author, I would have this as: Demon Diary, Vol. 1 by Chi Hyong Lee

My thinking is any way is fine, with author combining for all variations, as long as it's one and the same person.

"Lee Chi Hyong" is what you get when you place the comma after "Hyong" (as in "Hyong, Lee Chi"), which is the format TokyoPop uses in the actual volume. Depending on where that comma gets placed (or if a comma is used at all), you could get several variations. Where (or even whether) to split a name comes up with double-barreled names in English and Spanish also. Argh.

8chamekke
Ago 13, 2006, 4:00 pm

There's a known bug that affects only data imported from Amazon.co.jp, namely, the ISBNs do not transfer into your catalog. You have to input them *manually.* (When will this be fixed? Dunno. You might ask on the Google Group . . .) So that's why covers for these imports won't show up in your catalog unless you force them.

Yes, please do ask about this on the Google Group. I know it affects many people, but it always seems to be the same (small) number of people bringing it up. If more users mention that it's something they want fixed, it might move up higher on Tim's priority queue.

Also, re: the covers... I've been downloading them from Amazon.co.jp (where they exist) and uploading them as local copies. Is there a quicker/more reliable way?

Lastly, about the character corruption - as a backup, I've added a comment to each Japanese title that contains the link to the Amazon Japan listing. That way, if I need to retrieve any data from scratch, it's right there.

9lohengrin
Modificato: Ago 13, 2006, 4:29 pm

It's bizarre that TokyoPop would list it as Hyong, Lee Chi--Lee is one of the most common Korean family names, and I've never heard of it being used as a given name. Almost certainly, the correct name is Lee (family name) Chi-hyong. Amazon's dual listing of the book under Lee Chi Hyong and Chi-hyong Yi basically confirms it--Yi is an alternate spelling of Lee, and Chi-hyong is hyphenated as is most common for romanised South Korean names.

Korean names are actually among the easiest to figure out, as there are relatively few family names currently in use (250, according to Wikipedia, but with 45% of all Koreans having the family names Lee/Yi, Park/Pak or Kim ^^;; ), and it is most common that the given name is the two-character one (Chi-hyong, Jee-hyung, Ji-Tae, Myeong-min, etc.). So it's not as complicated as English and Spanish names, where there is a "middle name" involved. ^_^

10mikeneko
Ago 13, 2006, 8:01 pm

Re: ISBNs. I'm kinda, sorta hoping someone else will bring it up. I'd whined about the trashed UTF-8 so much that they're surely tired of hearing from me.

Re: covers. If Amazon doesn't provide, www.jpqueen.com has nice cover scans (and romaji search). I open up two tabs in Firefox (LT, jpqueen), then right click on the image for Copy Image Location, then paste that URL into the http upload box. If neither of them has the cover . . . uh, I've scanned a lot of manga. ^^;

Re: Korean. I don't know anything about Korean, to be honest. That's the only manwha I have, and I'd flipped it open to peer at the publishing info when I'd typed that. So now I'm wondering what's best to do . . . *ponder*

11bunnygirl
Ago 13, 2006, 11:37 pm

To make the Demon Diary issue even more confusing, there actually is a Library of Congress entry for it, and this is how they list the author: "Kara, illustrator." The Other Authors entry has "Yi, Chi-hyæong" listed.

Looking up the LT entry for Demon Diary by Kara (http://www.librarything.com/catalog/859079) shows that one user did pull the LoC entry for the book.

An aside: A lot of the first volumes of older Tokyopop books will have data somewhere at the LoC. For some reason, that's not the case with the newer ones. The only other major manga/manhwa publisher I've managed to pull LoC data on is Del Rey, which isn't so surprising. NBM and Drawn and Quarterly seem to have some of their comics on file, but they don't publish a lot of Asian material. None of the Viz stuff I own is in there.

12lohengrin
Ago 13, 2006, 11:46 pm

Yi, as I mentioned, is a variant spelling of Lee, so the LoC entry makes a lot of sense, and got the family name correct. ^_^ The difference in Chi-hyong and Chi-hyæong is probably an accented character showing up oddly?

13bunnygirl
Ago 14, 2006, 12:01 am

It most likely is; I'm pulling all my LoC stuff out of their Z39.50 gateway and all their encoding is apparently ISO-8859-1, not Unicode. It's just to show how amazingly messy all the data that LT pulls in is. :)

As for manual corrections of dodgy English Amazon entries on authors of Asian comics: Ice Kunion's spines list the authors in Korean-correct family name, personal name order. I've had to fix the imported Amazon data on this so many times it's ridiculous. Tokyopop's Snow Drop also uses the Korean naming order, which is backed up by the LoC data, with the correct author being Kyung-ah Choi, but the Amazon data screws up the family name. I should probably relist my Snow Drop v.1 out of the LoC data, except that's also using the non-Unicode, odd-letter romanization. The older CMX titles also went with a listing of family name, personal name and now have reverted to the Western order on the redesigned covers. I've had to fix dodgy data there too. I've also found Chinese one-word pen names misspelled on Amazon.

(Why, yes, I do read a lot of manhwa....)

14The_Holy_Terror
Ago 14, 2006, 1:38 am

Questo messaggio è stato cancellato dall'autore.

15mikeneko
Ago 15, 2006, 12:33 am

Still thinking about this . . .

I did possible author combining searches on Chi-hyong and Chi-hyæong with Lee and Yi and got no matches, so apparently no one's using those. So far so good.

LT -wants- LAST, First in the Author box; this in turn flips to First LAST in the default list display. But! Amazon.co.jp provides First, LAST; this flips to LAST First in the default display. Because the latter is actually the "correct" way to read a Japanese name, I don't bother to change it to LT's preference.

With the Korean names, it's sort of the same. To get LAST First in the default display, you have to use First, LAST in the Author entry. This looks peculiar in English, and -technically- it's not correct by the LT "Western" standard.

So, well, I guess it's up to the individual user's preference for finessing the default author display. As long as "Lee, Chi Hyong" and "Hyong, Lee Chi" both get combined for the same author on LT's author page . . .

I suppose this'll be especially pertinent if/when the author names in unicode characters are made available for combining.

16mikeneko
Ago 17, 2006, 1:23 am

OK! So I posted a request to the Improvement group to consider fixing that Amazon.co.jp bug. I posted it here.

No one else bothered to chime in, and it died instantly. Looks like I was the only one interested in seeing this fixed after all. (Yes, I know someone -did- add to it -- by rambling on about a completely unrelated topic . . .)

That's it for me. If anyone else is interested in taking up this particular cause, douzo.

17chamekke
Modificato: Ago 17, 2006, 2:49 am

Whoa... I only just saw it this evening (my first time on LibraryThing since last night).

I added a "me too" - for what it's worth! I really want to see this fixed; and as far as I can see, it's really not an unreasonable request.

Part of the problem is that as you've mentioned, mikeneko, it's not clear to many of us where to post a request for a bug fix (or improvement) ... and unless you're lucky enough to score a personal reply, it's often impossible to tell whether anyone on the LT side has even "heard" the request.

Last September (yup, I've been here a year!), there was the choice of e-mailing Tim or posting a comment (private OR public) to his profile. I tended to do the latter, and Tim usually replied pretty fast. Of course, things have got somewhat busier, and a lot more complicated, since then.

AFAIK, these are now our choices:

- E-mail Tim
- E-mail Abby
- E-mail one of the other newer staffmembers (Chris, Christopher...?)
- Leave a message on Tim's LibraryThing profile
- Post a message to the LibraryThing Blog (note: it will have to be a reply to an existing topic)
- Post a new topic to the Google LibraryThing group
- Post to an existing topic on the Google LibraryThing group
- Post a message to the LibraryThing group Site Improvements
- Post a message to the LibraryThing group 0101010101 - alt. binaries.

So here's the thing:

- Which of these options is "best" or most appropriate for the purpose of reporting bugs?
- Are bugs and site improvements are to be reported in the same place?
- Which place is likeliest to get a response?
- Are all requests being considered, irrespective of posting venue?
- Because everyone is looking at different groups and forums, and posting about bugs and desired site improvements all over the place, is there any way to gauge how widespread the desire for a particular bug fix may be?

(One thing I have noticed, however, is that replying on the "late" end of a blog thread is pretty futile!)

Recently Tim posted to 0101010101 - alt. binaries to propose that it be combined with Site Improvements. He said: "Now we'd like to create a single "Site talk" standing group, for features, bugs, help and etc. We contemplated having separate groups for all of these, but decided against it. The lines are too blurry. To get there, we'd like to combine 0101010101 - alt.binaries and Recommend Site Improvements..."

So my guess is, the Site Improvements group may now be Tim's preferred locus for all bug reports, and maybe that's where we should put our efforts in asking for bug fixes and site improvements.

(Maybe. ;-)

18trollsdotter
Ago 17, 2006, 9:30 am

lohengrin: Thanks for the information on Korean names. I had intuited that was how it worked, but confirmation is nice. I wish the publishers would indicate somewhere in the text of the book or publication data page which is the family name. Both Amazon and Ingram data entery personnel obviously cannot tell either.

19mikeneko
Ago 19, 2006, 5:23 am

I agree that this communication gap (chasm?) is a problem that affects everyone, not just us. I think you should post this where it will be seen by more than this group. As for where that would be . . . I have no idea. For all the reasons you've listed.

I can understand why no one noticed that post on "Recommend Site Improvements" -- I had trouble locating it myself a day later. At least the Google group could be searched for key phrases such as "unicode" or "Japanese," but posts there are growing gray hairs waiting for responses now.

More to the point for here, I was as astonished as anyone when my begging about the unicode got a reaction. No one on the LT side seemed to be aware that this was a fundamental, deal-breaker issue for some users, despite all the past posts about it.

Similarly, I'm not anticipating that posts about any of the other broken parts of the unicode ghetto (i.e., the corrupted information pages, the ISBNs, the bad matches, the lack of linking, etc.) will have any impact. This simply isn't important to anyone but a minority of users, so the best we can do is learn to live with and work around the problems.

20simside
Feb 11, 2007, 10:22 pm

Sorry for bringing this topic back from the dead, I had a question about this, and I thought about posting a new topic, but since this isn't a terribly active community, it would have only been a few posts up on the main page. I'm not sure of the etiquette of this type of thing here.

So I've been importing my collection title by title from Amazon/Amazon.jp, and I've been editing each title as I import it. I noticed today that it seems to combine every volume of most series as a single edition and makes a unique page for the series just for my entries AFTER I edit them. The separate/combine function is a bit confusing for me too since it seems like the existing data for the book is sometimes a bit messy (as mentioned above) and sometimes unnumbered, though to be fair I've only tried it on a series which imported from Amazon largely without volume numbers.

Does it have to do with the way I retitle the books with the series name and the volume in parenthesis? Or is it my author data that's throwing everything off? I realized after I'd done about two hundred entries that I was putting the name format in wrong (first last name instead of last, first which left me often with two tags for the same author) and since most of the volumes at that point had imported without an author name, I just went back through and changed everything to first last name.

21trollsdotter
Feb 11, 2007, 11:48 pm

I've never been exactly sure what causes LT to combine manga volumes into one. It may be an inability to properly process the japanese characters or the volume numbers in parentheses. I find myself separating volumes like this on a regular basis, so it's probably not you.

22mvrdrk
Feb 12, 2007, 12:03 pm

LT ignores certain things in titles. From what I've been able to gather, the rules are
- anything in parens or brackets is ignored
- anything after 20 character bytes is ignored, remember that Japanese is a double byte character set which means you only get 10 characters