Author combination guidelines - rewriting?

On the "Author" HelpThing page, there is a bit at the end with some instructions on author combining. We probably need a whole page for this, but regardless, I think these could use some revision. I've been seeing some people repeatedly making basic combination errors. Think: J. Smith combined with John Smith, or John Smith and Jane Doe combined with John Smith.

Here are the current instructions. I think we should delete the bit I've struck out. I think that line is fundamentally misguided, and no different in principle from the "author with collaborator" "should not" case; it's also reduntant, now that we can add other authors and change primary authors. What do other people think? Can we delete that line? Are there other revisions needed?

More on Author Combining

LibraryThing allows combining of authors much in the same way that it allows combining of books, so that "J. R. R. Tolkien", "JRR Tolkien", and "John Ronald Reuel Tolkien" all point to the same person. The top part of this box allows you to see which variants have been combined into one author's page (like the list at the top), and to separate out any combinations you think are in error.

There may be some suggested authors to combine. These are based on authors whose works have been combined via ISBN. For instance, if one person has the book "Combining for Dummies" by Random Q. Author and another person has the book "Combining for Dummies" by author R. Q. Author, both copies with identical ISBNs, someone may have gone through the Edition page of either work and combined the books. If that's the case, then R. Q. Author will show up in the "Combine with..." section of Random Q. Author's author page.

If you suspect there are other variants of an author name out there, you can search for them - results will appear on the author search page, with options to combine them with the author page you came from.

Who Should/Shouldn't Get Combined

Should

-- Variants of spelling, middle initials, capitalization, etc.
-- Pen Names ("Samuel Clemens" and "Mark Twain"), UNLESS there is another author by the same of either name
-- Designations of role after the author name ("George Lucas (director)" and "George Lucas")
-- Ordering differences where multiple authors are listed as one ("Ingri & Edgar D'Aulaire" and "Edgar & Ingri D'Aulaire")
-- Authors where another non-author contributer (i.e. Illustrator) is listed in with the main author ("Julie Andrews Edwards; Emma Walton Hamilton (Illus.)" and "Julie Andrews Edwards")

Should NOT

-- A surname, or surname and initial, only with an author with that surname, even if it's the most popular. (Neither "Brown" nor "D.Brown" should be combined with "Dan Brown")
-- An individual with a collaboration ("Neil Gaiman" should NOT be combined with "Neil Gaiman & Terry Pratchett")
-- Anything you're not sure is right! When in doubt, don't combine, and ask a question in the Combiners! group.
-- A named individual with "Anonymous", "Various", etc.

If anyone appears on the "suggested combination" list that belongs to one of the "Should NOT" categories, you can "never" that author to prevent them from being combined in the future. Again, don't "Never" someone unless you're sure, and ask questions if you're not.

From here: http://www.librarything.com/wiki/index.php/HelpThing:Author#More_on_Author_Combi...

I think we could use a whole new page on author clean-up, with detailed instructions on combining, splitting, aliasing, etc. I also think we could use a page on "how to get my book on the right author's page," since that seems to be what prompts a lot of the incorrect combinations; there are several ways to do that now, and which one to use really depends on the situation (primary/other authors, work combination, author splitting/aliasing, author combination).

And maybe some more specific directions on how to treat this 'should' ---

" Ordering differences where multiple authors are listed as one ("Ingri & Edgar D'Aulaire" and "Edgar & Ingri D'Aulaire")"

Do we chase down all of the works page variations and edit from there?

Agreed that "Authors where another non-author contributer (i.e. Illustrator) is listed in with the main author" should now be corrected with the other authors tool.

But why not also correct all instances of multiple authors listed in the main line and not worry about combining "Ordering differences where multiple authors are listed as one" or "An individual with a collaboration."

"There may be some suggested authors to combine. These are based on authors whose works have been combined via ISBN." Co-authors entered with the other author tool are now the primary source of suggested author combinations that I see. However, this may be a bug and not something that should be written up in the instructions.

We also need some new language to deal with the options of editing the primary author and dividing author pages. The following instructions could be better worded but I think cover the issues.

Only combine authors when both names refer unambiguously to the same individual or corporate author. Incomplete ("Brown" or "D.Brown" instead of "Dan Brown") or otherwise erroneous names should be corrected by editing the primary author using the Add other authors tool on the work page.

When two or more authors share the same name the author page should be divided. Do not combine a divided author with a more specific author (e.g. do not combine "Richard Smith" with "Richard C. Smith"). If books by an author appear on both pages use Alias divisions to other authors under edit the division on the shared name page to assign the appropriate author division to the author page with the more specific name (e.g. "Richard Smith (11)" aliased to "Richard C. Smith").

The last sentence needs better wording, yuck. These things are hard to describe briefly.

3, I was with you until "If books by an author appear on both pages use Alias divisions to other authors under edit the division on the shared name page to assign the appropriate author division to the author page with the more specific name..."

I say, "If books by an author appear on both pages, COMBINE the works. Do NOT use ALIAS to move works from one author page to another. If you need help combining works, visit the Combiners! group."

4 - Agreed that if they're the same work, they should be combined. But, I think what eromsted meant was different works, not the same works. So if you have works XYZ on John Smith (1) and works STU on John Adam Smith, you can alias John Smith (1) to John Adam Smith, so that you get S T U X Y Z all together.

I do think we need some instructions on when to use aliasing and other things. Aliasing is useful in certain specific cases. But I also dislike seeing it used as a kludge to move books around, when there are other methods that work better and involve less cleanup.

This is what I wrote on another thread, about why books end up on the wrong page, and what the options are to fix it. http://www.librarything.com/topic/131131#3175448

There are several reasons why a book might not show up on an author page (edit to add: or on the wrong page):
1) Caching. It can take some time for the author page to catch up with member catalogs.

2) Variations of author's/authors' name: Not everyone has a book cataloged under the same author, or with the same spelling. In a single-authored book, there might be 2 or more variations of the name in peoples' catalogs. In the case of this book, you can see those by looking at the editions page: there you'll see that members have this book cataloged with several variations of the author's name in their individual catalogs. Most have the book listed with " Rosenfeld, Jordan" as the author (no middle initial). Some, like you, have "Rosenfeld, Jordan E." One has "JordanRosenfeld" - all one word. And someone at some point had the author name listed first last instead of last first: "Jordan Rosenfeld."
Chances are, in this book's case, the book was until just now being listed on the Rosenfeld, Jordan author page, rather than the Rosenfeld, Jordan E. author page.
A book will show up on whatever author page "wins out" -- usually due to number of copies, although this is somewhat mysterious.

3) More than one author. This is a similar case. Where a book has more than one author, some people might have it listed under one author's name, while other members have it listed under another. Some - heaven forfend - might have it listed under an "author name" that is actually both authors. Again, the book ends up on one winning author page, by processes mysterious.

Now, what to do about it.

This depends, and not all options will always be possible. (edit to add: not all options are equally appropriate for each scenario. These are roughly presented in rank order. I definitely don't think people should be using option 3 when option 2 would work.)

Option 1: If possible/permissible, combine the author pages. If the author pages *definitely, absolutely, and with no ambiguity* refer to the same person, and there are books on both author pages, you can combine the authors. See the right hand side of the author page for this feature. HOWEVER, this should not be done lightly. In this case, I would not combine "Rosenfeld, Jordan" and "Rosenfeld, Jordan E." unless I was 99.9999% sure there are no other Jordan Rosenfelds out there with some other middle initial.

Option 2: Edit the primary and/or other authors for the work. If a book is co-authored, you can add the second author as a main author, and it will then show up on both author pages. If the winning author's name has a typo, you can fix it by changing the "primary author." Primary author changes should also not be done lightly, however, since they affect global-level settings for the book.

Option 3: Splitting and aliasing. When two author pages can't be combined, because, say, there's more than one author named Jane Black, but you want to make sure that some of the books on the Black, Jane page are properly attributed to Black, Jane Q. You can split the "Black, Jane" page and then alias the appropriate split to the "Black, Jane Q." page. That will make the book show up on both author pages.

Option 4: Wait, if it's just a caching issue.

5, I see.

Now I say, "If the same books by an author appear on both pages, COMBINE the works. Do NOT use ALIAS to move works from one author page to another. If you need help combining works, visit the Combiners! group. If different books by an author appear on both pages, use ADD OTHER AUTHORS to move the books to the desired author page. Do NOT use ALIAS to move works from one author page to another."

6, I was with you until, "When two author pages can't be combined, because, say, there's more than one author named Jane Black, but you want to make sure that some of the books on the Black, Jane page are properly attributed to Black, Jane Q. You can split the "Black, Jane" page and then alias the appropriate split to the "Black, Jane Q." page."

I say, "When two author pages can't be combined, because, say, there's more than one author named Jane Black, but you want to make sure that some of the books on the Black, Jane page are properly attributed to Black, Jane Q. You can use ADD OTHER AUTHORS to move the books to the desired author page. Do NOT use ALIAS to move works from one author page to another."

>8
So do you think there is any use for the alias function now that we can edit primary authors?

8 - You basically seem to be saying not to use aliasing at all, ever.

I agree that it is used too often for the wrong things, and that in many cases it is not the appropriate first solution. But I disagree that every book that Jane Q Black published simply under the name Jane Black should have the primary author changed. IF there is an author page that refers to more than one person, then splitting and aliasing is appropriate.

Sometimes people split single author pages into one split, just for the purposes of aliasing. I don't think this is a good method for anything, and I suspect you don't either.
If there's only one author on the author page, then it can probably be combined with the other page. If it can't be combined with the author page because the other author page is split, then the *other* author page can be aliased. If it can't be combined with the other author page for some other reason (it's a last name only page or a first initial last name page, or just a *wrong* author for the books listed underneath it), then in those cases it makes sense to change the primary authors: e.g. last name only authors, or first initial last name authors.

9, no, but in practice, there will be situations where aliasing will probably "have to be" used because of the number of works being transferred. One example involves Golden Books and Walt Disney. On the author page for Golden Books, there are many works that should be re-authored to Walt Disney. But there are too many, unless a group effort was undertaken. I have thought about aliasing the whole lot over to Disney, but just didn't. I may move them one by one eventually, but that's pretty boring and I'd have to be desperate for something to do.

So my answer is no, but I understand it has to be yes when there are big numbers involved. But one or two or five? I would say please do not alias!! It leaves empty links at the top of the author pages.

edit grammo

10, I got a little confused with the last part of your post. Could you use real examples?

http://www.librarything.com/author/reynoldswilliamj without aliasing will be a fine, fine mess... :) For example.

Aliasing makes this kind of pages properly ordered.

Adding another author because of the variation of the name is a misuse of the other authors feature in this case - it is not another author, it is the same one. It is just that under the full name it is just one of them but when you use an initial, you have two of them.

This is what aliasing is all about.

13, done.

>14

Now move the photo to the correct page, don't leave the work half done.

And I still call this misuse of a feature. Because the name is VALID for this author and works will keep popping up..

Editing the primary author is a very powerful tool and I would like guidance from the staff on whether we should use it for improving organization not just fixing errors.

>16

Agree. And kind of the reason I don't use it for cases such as the above one.

The full name is NOT the name usually used by this author so defaulting to it is not the correct action in my opinion... So... we are down to someone deciding what is the correct name for everyone? That is just... ugly. Next time we see, books that are written under pseudonym will not just show up on the correct page but will also be "fixed" to show with the legal name...

>17 I am with you on this, AnnieMod. I like to see books under the name of the author as s/he wanted it on the front cover/spine, so H G Wells and P G Wodehouse but John Buchan and Evelyn Waugh. Fine, of course for the well known and/or prolific but harder to justify for J Jones, author of Swansea Nights.

Yeah...

PS: The other author above is http://www.librarything.com/author/reynoldswilliamjense
The problem is that he uses William J. Reynolds on the covers - same as the mystery writer. So... we have a more beautiful page but not what people will expect to see -- at least with a split and an alias, people could see the connection. The way it is now? They will simply decide that we do not have any of the books recorded in LT... Which just jibes me as wrong.

... and, presumably, any new copies of hymnody works will appear on William J.'s page and stay there until Collectorator or some other assiduous combiner notices them and moves them.

If they can find the other page - there is no link between the two just now. Which with the split was not exactly like that -- they needed to be assigned/combined again but at least noone had to do all the research again to see where they fit or guess that there is a place where they fit.

And I really do not like someone to force an author name that the author does not use on most of his covers (if not on all)...

#21 Actually covers aren't the standard - the title page is. However in 95% of the time, they match. And I do have to agree that aliasing in the example in #13 is the way to go, especially if it isn't ratty Amazon data. In those Amazon cases as well as lazy author entries, I believe that correcting the author is the right thing to do.

>22

Agree on the title page but in most cases they will match nicely enough anyway:) However not the Copyright notice from it (seen that done a few times so just mentioning it) - it can go to a different name at any time.

The whole idea of allowing is to "fix" names was to correct errors - ratty data, people that misspell names, things like that. Not to beautify the data as we see fit... except that obviously not everyone agrees and see what happens... That is why I had split and aliased -- instead of "fixing" a non-existent error... That is why I called it a misuse of the feature.

>18 yeah, but then there is the problem where a different form of a name is used on works published in another country, perhaps because there already is another author publishing under that name there...which means when the author name is split, the total number of works end up aliased to the author's alternate name choice, with only those winning out as the majority name after work combining on the other page.

At least Canadian author L. R. Wright #1 ( http://www.librarything.com/author/wrightlr-1 ) is assured of having a proper author page under the new system, which is ok I guess, better than the alternative of using only the US version of the name.

Splitting and aliasing is a much better solution for these books.

It takes me about three weeks to write an impromptu speech - Mark Twain

I'm sorry I haven't jumped in here before now; I wanted to see where the conversation went, and what issues around this particular situation cropped up. And I wanted to be able to chat with Tim about all this, which we haven't been able to do until recently. It's a tricky issue, and one with all sorts of annoying little edge-cases (as you all know extremely well).

Recognizing that there may well be certain exceptions to this general principle, where we're ultimately coming down on this is that in cases like the (made-up) example of Jane Q. Smith (who, let's say published works under both "Jane Q. Smith" and "Jane Smith), there would likely be author pages for both "Jane Smith-1" and "Jane Q. Smith," with different works appearing on each according to how the name appeared on the books. In this case, the alias feature should be used, rather than setting the primary author of works with the author listed as "Jane Smith" to "Jane Q. Smith," if that is not the way the author's name appears.

The main reason for this is to preserve the linkages between these different variants of author names, as pointed out in #20-21. Changing the primary author in these cases often seems to be resulting in difficulties later, as new copies of those works are added to the system, &c. Also, think Mark Twain/Samuel Clemens, or Anne Rice/A.N. Roquelaure; the work-level authors are as they display on the books.

We know this isn't the consensus answer on this, and there may well be specific cases where it becomes necessary to change the primary author for one reason or another (bad/incomplete data, misspellings, &c. &c.). Generally speaking, though, the "primary author" feature is not designed to remove the need for splitting and aliasing in cases where the data calls for it.

When those weird edge cases come up, I think this group is perfectly suited to discuss and handle them, but I'm always more than happy to help out when/if I can provide guidance.

Let me be very clear that changes made to this point were not made wrongly in any sense. We hadn't developed fully the general guidelines here (as with many aspects of the site, particularly with new features, it proved useful to see how things shook out a bit), and we don't want anyone to think that they've been editing things inappropriately at all. Going forward, we hope this will help.

Questions, &c., I'm here! And again, my apologies for taking so long to get this posted.