Editing "Subject" information

When entering books, some have "Subject" information already in the subject field and some do not. I have 2 questions in this regard:

First is it possible to edit this information when it is there? I've noticed that at times it is incorrect especially with regard to Large Type Books. "Large Type Books" has come up in the subject field and my book is not a Large type book and I cannot seem to edit it out of the field. Is it possible to edit this field?

Second, when the "Subject" field is empty, is there anyway to insert information into the field?

Any help would be appreciated.

The Subject field is non-editable. It's generated from the Library of Congress subjects for your book. You may notice (since your library is private, I can't see for myself) that sometimes the text here is blue, and sometimes it's green. Blue text is from your own edition; books added from library sources often have this information available directly. Green text is a "best guess", which you'll always see for books you added from Amazon (and sometimes even from library sources). This is where the edition-specific subjects (like "Large Print") that are incorrect for your edition come in.

While subject is not user-editable -- and a good thing, too -- you will have better luck if you add books from library sources rather than Amazons. (If you've done so, I apologize, but I can't tell since your library is private, so I have to be very general in my suggestions.)

To add - many library sources import subjects into your LT record when you add a book. Not just the LoC. (But many U.S. libraries appear to use the same subjects as or from the LoC).

So - maybe the specific edition you added from a library was a large type book? I recommend checking for (more) info at the add books part on each edition to see what the subjects are before selecting the one you wish to add.

Subjects are blue if they were imported when you added a specific edition from a library source.

They're green if the source you used didn't have subjects; however the global work record of that book does have subjects from other members' imports. That's why a Harry Potter book with Green Subjects will have motion picture, CD etc in there as well as the expected subjects.

Either way, they're not an editable field. Or a retro-insertable field. :-(

The only current way to get the subjects is to re-enter the book. Sadly I have over 300 books that I entered in early LT days before I realised how important adding from a library source was to me.... so I know how you feel!

The only current way to get the subjects is to re-enter the book. Sadly I have over 300 books that I entered in early LT days before I realised how important adding from a library source was to me....

Unfortunately, that's no comfort for those of us who want all the other information to be correct, too, but have too damn many books to edit it all from library sources.

With all due respect, I must say that there are not only LARGE PRINT errors in SUBJETCS.

Sometimes the info coming in is simply bad, wrong, misleading, whatever. I would actually like to see a function that enables us to turn this off.

Take the EPIC OF GILGAMESH, for example.
Subjects claims that this is English poetry ... whereas the translation in question (of an ancient sacred text from Mesopotamia) is done in prose ...

Not funny ... so one better can't value this field too much.

I've got Google open in another tab in an attemt to find the an official request for editable and/or cleaned-up subject data. My copy of Jane Austen's Persuasion, for example, has some subjects (starting with four of the first five listed) that start off just fine before veering off into irrelevancy. Not to mention the last subject, "Upper class › @".

Edited to add: Even limiting the search to talk threads, I couldn't find an "official" thread. Of course, it would help if I could remember whether or not it was in the Recommend Site Improvements group.

6>

Wow, that one's...interesting. I can understand not wanting arbitrary editing, but maybe there could be some threshold, that if a subject appears fewer than N times it doesn't appear in the overall data? "Upper class > @" has got to be a weird error that on a book as popular as Persuasion could easily be screened out.

Just agreeing with others that the info in the subject field is often seriously wrong. In my opinion, the subject field is worthless.

Some books are better than others, and some headings are further afield (for lack of a better description) than others.

One of the things I suggested in Message 8, before it got eaten, is that we have a series of flags available to us. One might be for headings that are obviousle edition specific, like "Audiobook" or "Large Type Book". Others might be for heading duplications that vary only by comma placement* or language**.

----------
* As in "Joan of Arc, Saint, 1412-1431" and "Joan, of Arc, Saint, 1412-1431" here and here, for example.

** I've seen this in the Harry Potter series in particular -- including the pair "Sorciers › Romans, nouvelles, etc. pour la jeunesse" and "Wizards › Juvenile fiction" here. But it happens elsewhere, too: "Jeanne d'Arc, Saint, 1412-1431 › Drama" and "Joan, of Arc, Saint, 1412-1431 › Drama" here.

I'm not sure that foreign language subjects are a problem. What do you expect would happen when you combine an English book with its foreign translation or vice versa?

Its not just libraries you get green data from Amazon too, though not so much for out of print books which have empty fields.

Looking at some in my library, I have a few Large Type also (which seem to be the only glaring issues - though I am not checking everything).

Perhaps the data is developed for all the books of that title and not a specific edition (though its hardly consistent) ?

I can't see why we can't edit it, if the data isn't used by LT for something it shouldn't matter. Is there a future development that is planned to use this data ?

I can't see why we can't edit it, if the data isn't used by LT for something it shouldn't matter. Is there a future development that is planned to use this data ?

You can't edit it because it's work-level data, with a controlled vocabulary.

Yet, some libraries are obviously using it for edition-level data (Large format, etc.).

14>

Right, but on LT it's work-level (not book-level) data, since they don't have an edition level. The "controlled vocabulary" is at least as important, though. I can see screening some subjects, but not allowing free-form editing. (And, though I forgot about this earlier, the data is used for some purposes -- on tag searches and tagmashes, you'll see "related subjects" in a green box on the right.)

I'm not unwilling to see a book-level subject field, but I'm not sure yet how it should work. Let me wake up fully and then I might be able to come up with something.

I just wanted to add (though it's been a long time) that huge numbers of books now have no subject at all. The LCC doesn't shelve books submitted to it that are self-published, and is frankly behind on the ones that are trad-published. In terms of the usability of the 'Subject' field it's now pretty much not at all usable. Even for libraries. Some method to allow people to enter a subject, even if they need to be vetted in some way first and restricted in vocabulary, would probably make LibraryThing more useful than the LCC at this point.

>17 WingedWolf: Honestly, just use tags. Not even larger developers are going to revamp and police subject headings, let alone LT's small team

I don't think subject headings are work-level. If they were, it wouldn't matter what source people used to enter their books. If I were to enter an older book via Overcat, and someone else were to enter that same work via Amazon, then if subject headings were work-level, we would have the exact same ones in both of our libraries, because they are the same work. Maybe the green subjects are work-level, but if there is any field that is different on my book than it is on someone else's, but both books link back to the same work, then by the very way the feature works, it can't be work-level data.

Personally, I would really prefer that this was something that could be edited. I own many books that are not found in any library LT sources from, and have no subject headings at all.

>19 GraceCollection: I've tried to sanitize some of my books by deleting them and adding manually. But Subjects pop up again with character issues so I've given up. Most (/All?) of it is just encoding gone wrong, i.e.

Bible > Antiquities > Fiction
Brennan, Temperance (Fictitious character) > Fiction
Israel > Fiction
Montr�eal (Qu�ebec) > Fiction
Mystery fiction
North Carolina > fiction
Women forensic anthropologists > Fiction

>20 bnielsen: Depending on how popular the book is, I wonder if you could use Overcat to try to deliberately find a copy with no subjects? I know I've seen that once or twice but I don't know exactly how common it is.

If you add books with no subjects, you get the green ones.

>21 GraceCollection: The popular ones are exactly my problem, since they populate the Subject field no matter what I do. And I like getting the book combined with other editions, so it is impossible to avoid getting the Subject field as an unwanted add-on. (Most of the time I like getting the Subject field, but I'm not fond of encoding errors.) My current workaround is to add a line in Comments and use a script to filter the export file. I'd like a way to fix the data instead.

>22 MarthaJeanne: I should add that I'm mostly interested in the contents of the Subject field in the export file.
Alas what you see in LT and what you see in the export file is not always the same. I.e.

Book_Id 22821953 displays this in LT (green text)
...
Ecoles > Romans, nouvelles, etc. pour la jeunesse
...

but when I look in the export file, it says:
...
�Ecoles > Romans, nouvelles, etc. pour la jeunesse
...

So the LT "Your books" view seems to hide some of the weird data.

>22 MarthaJeanne: >23 bnielsen: I see. I had never deliberately tried to get no subjects, so I thought Amazon got the green subjects, but a library which had left subjects blank might not. I have plenty of blanks which I wish weren't so.

>23 bnielsen: Haven't tried exporting anything (yet) but I certainly have found weird data in LT, so I wish they at least did a better job of hiding it.