• LibraryThing
  • Book discussions
  • Your LibraryThing
  • Join to start using.

Combination Opportunities

New features

Join LibraryThing to post.

Jan 13, 3:15pm Top

I've made a new feature, called "Combination Opportunities" that serves up a "problem"--two or more works that are intertwined in some way and need help. At least half the time, the solution is obvious—combine them.

At present there are almost 100k of them. They were selected by an algorithm. The algorithm was:

1. An ISBN that is "split" between works.
2. At least one edition had the same title between the works.

The most common type are combinations between works with an author and works without (e.g, "A by B (ISBN C)" and "A (ISBN C)." But there are lots of other types.

Most are REALLY simple, so it's a relaxing activity (for weird people like many of us). If something is too complicated, click "skip to next." If something isn't actually a problem, click "Okay as is." If something was a problem, but you've solved it (e.g., by separation), click "Mark as done."

Here's the link!


PS: I'll put up the helper list soon. Helpers will of course get combiners medals, and I'll make a special medal for helping on this.

Jan 13, 3:48pm Top

I'm totally hitting "skip" on ones that will take too long to figure out, or for which I don't have the right skills.

Jan 13, 3:51pm Top

Thank you for acknowledging our "weirdness."

I'm skipping any in languages I don't know, e.g. Latvian, Romanian, Finnish, Russian, etc.

Jan 13, 3:56pm Top

Thanks for another time sink, Tim!

Jan 13, 3:58pm Top

This is fun, and I don't have time today! Sigh. Don't do it all right away, Combiners!

Jan 13, 3:59pm Top

Oh no! I just combined two things that shouldn't be combined. I have to go pick them apart.

There goes my speed. jbd1 is going to beat me.

Jan 13, 4:00pm Top

We DEFINITELY need a way to pick more than one edition out together.

Jan 13, 4:04pm Top

>7 timspalding: - yeah that would be handy. When there are just a couple, it's easy, but when it's lots and lots of different ones all knotted up, tricksy.

Jan 13, 4:09pm Top

>7 timspalding: Yes, please :) I just spent more time untangling a mistake than I would have spent dealing with a few of those....

Jan 13, 4:09pm Top

Just skipping the rats nests for now. Don't have the time or the energy. :-p

I figure this will last us about... three weeks?

BTW: a great opportunity to find authors that need combining as well, particularly foreign-language versions of author names. If you're not gonna combine, say, the kanji and romaji versions of a Japanese author's name, best to skip it in deference to someone who will.

Jan 13, 4:16pm Top

>10 davidgn: "Three weeks" as being the next step up from Tims "Two weeks"? Or based on some actual calculation?

>4 lilithcat: My exact thought when I saw this :-)

Jan 13, 4:25pm Top

>11 bnielsen: Two weeks for the easy 90%, and another week for the rat's nests. :-p

Jan 13, 4:26pm Top

>11 bnielsen: I dunno, I once spent three months on The Rubáiyát. Look's like it's a mess again, too.

>1 timspalding: Thank you for recognizing that some of us do this for fun and relaxation!

Jan 13, 4:41pm Top

>12 davidgn: Figures. I'm sure I have a short story somewhere with a title like "The Rats in the Walls". :-)

Jan 13, 6:16pm Top

What fun!

Edited: Jan 13, 6:34pm Top

>6 timspalding: Oh no! I just combined two things that shouldn't be combined. I have to go pick them apart.
>7 timspalding: We DEFINITELY need a way to pick more than one edition out together.

Yup. Users have known this since 2007: https://www.librarything.com/topic/5923

Could you at least make the (separate) link an actual link so it can be middle-clicked to open the confirmation in a new tab, allowing multiple separations without reloading the page? (There's a script that does that and much more - http://www.librarything.com/wiki/index.php/Hacking_LibraryThing#Work_Editions_Gr... - but it would be nice if it was written into the page.)

Jan 13, 7:04pm Top

>16 r.orrison:

I've known it for a long time too. It's not an easy task. I've been working on it since then, with a break for a meal, and I'm still just working on the pieces.

Edited: Jan 13, 7:29pm Top

Oh man, this totally beats my current time-filling/wasting method of just using "Folly" to find random books in my library and investigate their combinatory needs.

Jan 13, 7:32pm Top

I love this! However I am reluctant to do the foreign language ones unless they are exactly the same.

Edited: Jan 13, 9:35pm Top

I love this so much. I'm coming out of my hibernation to tell you that this is about 50% of my workday sometimes (I mean, on our database, not LT - though when I first started entry-level office work 8 years ago, it was also 50% of what I did at work but on LT not on our database) and we have a similar tool that throws out random potential duplicate donor records to examine. It is weirdly relaxing and satisfying, especially if a bit of research is necessary.

There are always ways to dig and find the records here on LT that I wouldn't have thought the tool necessary, but heck yeah it's great!

One question: I just came across "Svampar i naturen, kulturen, köket" by Olle Persson with the suggestion of "Svampar i naturen" by Bo Mossberg. But there were 4 suggestions. Once I combined two of the obvious matches (a 0-entry by ollepersson with a regular by Olle Persson), I'm taken away to another set to examine. Does that take the whole set out of the pile, or does it act like a "skip" on the remaining potentials?

(PS: I am fairly certain they should all be combined, because all the names are on the cover as different types of author, but I figured I'd use it as a test case to see what would happen with the multiples.)

Jan 13, 8:56pm Top

Does that take the whole set out of the pile, or does it act like a "skip" on the remaining potentials?

The combine does mark the opportunity as closed. But if another remains, it'll be added back next time I run the finding script.

Jan 13, 9:34pm Top

>21 timspalding: Good to know, thanks!

I'll make sure to do the multi-step things outside this tool, then, and only the final combination or "it's done" inside it, so any lingering potentials can be looked at by someone better. Or me, later, when I'm in the mood for a complicated tangle.

Jan 13, 9:45pm Top

Ideally, I'd like a way to separate and combine in one step--like moving pawns around on a chessboard. But, wow, there's some tricky logic that has to go on in these things. So some doing outside and in sequence is unavoidable for complex situations.

Jan 13, 10:14pm Top

I tried one where there was some simple combining to do. Then I noticed that the main entry had a mixture of vols. 1 and 2. Trying to separate them out and find them again became very confusing. At first it disappeared. Fortunately I remembered it and was able to go back to it. However it is still somewhat of a mess because the same work has multiple authors.

Jan 13, 10:26pm Top

>24 krazy4katz:

Yeah. Nightmare. If someone sees a good example that needs the separate of multiple editions, send it to me, or post it here. I'm struggling to test my code.

Jan 13, 10:28pm Top

24> Yes, I've come across several like that. The most complicated I think so far tonight is A History of Religious Ideas vols. 1-3, which may need further work by someone familiar with the series. It appears that the Germans decided to make it 5 vols and the French only 2? I'm not sure. I had to move several volumes around between three works.

For those, I've been doing the separate tab thing and making ample use of the Work Table, which is a marvelous invention for tracking them. I've been opening all the works in new tabs and not working from the combining tool page until I'm ready to mark as done.

Edited: Jan 13, 10:39pm Top

25> Try Pete Frame's Complete Rock Family Trees and Rock Family Trees v.1

eta: i just realized that i can use the back button on my browser to return to the ones i had skipped A+++ usability there

Jan 13, 10:41pm Top

>27 keristars:

Yeah, that helps.

Jan 13, 10:41pm Top

Ooh, this is a neat toy. Off to play!

Jan 14, 12:21am Top

Back in the top 5!

Edited: Jan 14, 2:45am Top

>25 timspalding: Here's a work that is a mess of volumes 1, 2, 3, 4 and the full set:


Jan 14, 7:28am Top

So my only issue with this "toy" is this:

If something takes more research than a simple click to combine, or there are different combines on one page, you can only do one before the system marks it as done. What I mean is say there are six different works listed on a page, because of the volume issue, and you need to combine groups of two so it reduces to just three works. The first time you hit combine, the system moves you to the next option, and you can't use the back button, because the screen blanks when you refresh to find the one you missed.

Jan 14, 8:24am Top

You can Ctrl+Click on the different "go to work" links to open them in new tabs and do whatever combination and separation needs doing, then come back to the Combination Opportunity tab to Mark as done.

Edited: Jan 14, 9:08am Top

>34 r.orrison: Yeah, this is what I've been doing. I untangled a real mess of volumes with Analytical Key to the Old Testament last night. How am I going to move up the leaderboard if I spend 30 minutes on one opportunity!?

Edited: Jan 14, 10:26am Top

>35 Stevil2001: My sympathies. But you are working for the greater good, eh?

On the weekend, I’m handicapped having only my phone to play the game; so I’m skipping some of the enticing complex messes.

I am separating some of the wrongly placed individual entries that suggest an incorrect combination and thinking that will simplify the combining when it comes around again. Like removing the book ‘Checkers’ from the game.

Jan 14, 11:15am Top

Asimov’s Guide to the Bible needs some separation. Volumes 1 and 2 are sometimes mixed in with the complete set.

Jan 14, 11:32am Top

>35 Stevil2001: Right after I posted that I spent 30 minutes unentangling The History of Sexuality's three volumes. Yikes! But this is good fun.

Edited: Jan 14, 12:51pm Top

My issue was with Thomas Fitzpatrick (https://www.librarything.com/author/fitzpatrickthomasb). He coauthored some of these works with other authors, including Klaus Wolffe. So some are under Wolffe, some under Fitzpatrick and a mixture of volumes. Was trying to work on Dermatology in General Medicine. I straightened out some, but might have lost a few to who knows where. This is great in principle, but I will have to be careful not to get in too deep.

Edited: Jan 14, 1:00pm Top

This is fun. I've been free with the "skip" option, and mostly resolved ISBN splits where one or two records failed to enter an author for what is obviously the same book. Respect to those who tackle the thorny multi-volume morasses!

Jan 14, 2:34pm Top

I am ready for my badge now, Mr. DeMille.

Jan 14, 3:14pm Top

>40 paradoxosalpha:

Yep, The Hobbit came up for me - as it was matching an audio dramatisation against the main work (which seems to contain video and audio dramatisations). At over 65K copies it wasn't going to be one that I would touch.

Edited: Jan 14, 4:23pm Top

>42 andyl:
Same for The Three Musketeers and Twenty Thousand Leagues under the Sea: someone with time on their hands can take a jab at those.

Edit: and Hamlet. Just... no.

Jan 14, 5:40pm Top

The Bloom whatever are NOT the main work please (untangling from All quiet on the Western Front at the moment)

Jan 14, 5:57pm Top

I just got this one:
I will skip it and let the spam-fighters figure it out.

Jan 14, 6:04pm Top

>45 Collectorator: Huh. Looks like it might have been an algorithmic catch made in error?

Jan 14, 8:04pm Top

Here’s an oddity:

I was taken to https://www.librarything.com/work/5630995 and there was nothing proposed with which to combine it.

Jan 14, 8:12pm Top

>47 lilithcat:

I had one of those last night - went back into the helper log and found that someone had combined it earlier in the day (apparently outside of the new page mechanism so the opportunity did not get cleared out). Maybe something like this happened here as well?

Jan 14, 8:14pm Top

>48 AnnieMod:

That sounds likely.

Jan 14, 11:31pm Top

No way I’m doing Dracula.

Jan 15, 12:16am Top

I just skipped Frankenstein.

Jan 15, 12:27am Top

>51 Collectorator: Just saw and did it, but the ISBN highlighted for me was just a few stray Penguin Easy Readers, there are probably worse tangles in there.

Jan 15, 4:15am Top

What to do with combination candidates that have multiple authors where they are each assigned to the different authors? They should be combined, but do you need to do more? Other Authors?

Jan 15, 5:32am Top

I've been doing other authoring when it made sense - but then I had been falling down the rabbit hole a lot with these combinations. :)

Jan 15, 8:31am Top

This has not been a free exercise for me. I’ve noted a dozen or more titles I’ll need to order from the library to at least skim.

Edited: Jan 15, 9:49am Top

One way to prevent this happening all the time is not to allow a book to be added without an author. Even if the author is an organisation i.e. Readers Digest it has to be added. I assume most of these additions are manual.

e.g. How did this ever happen?

The Firegator by Debbie Leland (4 copies; go to work)

The Firegator / Debbie Leland / (ISBN 0966708628) (separate)

The Firegator (1 copy; go to work)

The Firegator / (ISBN 0966708628) (separate)

Jan 15, 9:40am Top

>53 JerryMmm: Yeah, I usually go to the combined work and make sure both authors are assigned.

Jan 15, 9:48am Top

>56 bergs47:

And for books where no author is given?

Jan 15, 9:50am Top

>45 Collectorator: Collectorator: Huh. Looks like it might have been an algorithmic catch made in error?

Post the combine page next time; it has a code in the URL.

It's quite possible you hit one someone else had just done. Maybe they did it outside of the system, so it never got cleared.

Jan 15, 9:51am Top

>51 Collectorator: Collectorator: Just saw and did it, but the ISBN highlighted for me was just a few stray Penguin Easy Readers, there are probably worse tangles in there.

Yeah, I'd only go for the tangle is specifically identifies. Incidentally, Dracula may occur again in another tangle--it's about ISBN overlaps, not works, so a work can occur twice.

Edited: Jan 15, 9:57am Top

Take a look at this, BUT DON'T FIX IT, PLEASE.


I take this to be unfixable--ratty data. Someone cut and pasted. The ISBN and the description clash.

Exception: These two should be removed from the Volume II work and moved to the Volume I work.

Luke: That You May Know the Truth, Volume I (Preaching the Word) / R. Kent Hughes / (ISBN 1581340281) / 30468543 (separate)
Luke: Vol 1 (Preaching the Word) / R.Kent Hughes / (ISBN 1581340281) / 19666626 (separate)
Luke: 1 (Preaching the Word) / R.Kent Hughes / (ISBN 1581340281) / 34657843 (separate)

The others are unfixable clashes between title and ISBN.

Jan 15, 9:59am Top

>47 lilithcat: Where two works need combining they may share more than one ISBN. In this case combining under one opportunity will make the other redundant but it will still be in the list.

Jan 15, 10:05am Top

>61 timspalding:

Yep there are a few I feel are just c+p issues where there are multiple volumes. Or even worse with yearly issued books.

Jan 15, 12:16pm Top

>61 timspalding:
The system has thrown several of those at me. Or cases where the isbn for an omnibus of novels has been used for all individual entries.

Jan 15, 12:58pm Top

>64 Petroglyph:

There are also some collected sets that really do have the same ISBN for all titles. Nightmares.

Jan 15, 3:45pm Top

Does anyone else feel that these opportunities are disproportionately skewed toward religious texts? I feel like I see a lot of those, and also academic books.

Jan 15, 3:49pm Top

>66 Stevil2001:

Yep it generally comes down to either
a) a religious text
b) an academic or professional book with multiple authors / editors
c) board books for kids
d) art / photo / museum catalogue books
e) misc. non-English text.

Jan 15, 3:57pm Top

>66 Stevil2001:, >67 andyl:

The only trend I've noted is towards multi-authored non-fiction, which is surely to be expected.

Jan 15, 4:26pm Top

>66 Stevil2001:
Even on random combining (i.e. via new additions to LT or via the "random book/author" feature), many religious authors tend to crop up, because LT is used often by US church libraries.

>68 AndreasJ:
I'm also seeing non-fiction from multiple authors (or variants of the author's name with/without prefixes such as MD). That and authorless singletons.

Jan 15, 4:46pm Top

>69 Petroglyph: Even on random combining (i.e. via new additions to LT or via the "random book/author" feature), many religious authors tend to crop up, because LT is used often by US church libraries.

Mm, whereas in my usual combining, I will almost never encounter a religious work b/c my library is sci-fi, comics, and Victorian lit!

Edited: Jan 15, 5:35pm Top

>66 Stevil2001: >69 Petroglyph: Yeah...

If someone has a lot of time to spare there's a Treasury of David that's a jungle of multi-volume editions, some with the same ISBN for all the set. Joy.
Hacked at it a bit but still plenty more to do.

Jan 15, 9:17pm Top

>55 2wonderY: Yeah, same here. :) The perils of combining, right?

Jan 15, 10:44pm Top

OMG. Don’t ever let me near Molière again.

Jan 15, 11:32pm Top

If anyone has an example of a page that needs multiple editions separated out, let me know. I need to test my script.

Edited: Jan 15, 11:47pm Top

>74 timspalding:

Here is one: https://www.librarything.com/work/322088

All the editions starting with "Houghton Mifflin Reading" should be pulled out

Edited: Jan 16, 12:01am Top

>75 AnnieMod:

Holy shit—I WORKED on those books at HM. What's the rationale? Socially, it's the same book, isn't it?

Edited: Jan 16, 12:26am Top

>76 timspalding:

They are for different grades though - so they need to be separate in my book. But even IF they were to stay together later, they need to be out from "Jamaica Tag-Along" - "Houghton Mifflin Reading: Teach Edi Theme 4 Level l K 2008" have nothing to do with it for example. Or "Houghton Mifflin Reading: Student Anthology Grade 4 Traditions 2006". Or "Houghton Mifflin Reading: Student Edition Grade 3.2 Horizons 2001"

Jan 16, 12:27am Top

Okay, the new editions-separation code is in place. It is slower if you are removing one edition.

>78 AnnieMod:

Hmmm. I'm not sure. I suspect they're exactly the same book, but marked different for some marketing reason. Either way, I'm not getting into the weeds. So whatever you and others think is right.

Jan 16, 12:30am Top

>79 timspalding:

Hm... I guess they will need a second look again.

Jan 16, 12:33am Top

I just did some awesome work on Anais Nin. Much easier with multiple work.

FYI, although indeed this was requested, it took me at least eight hours to pull together. So, yes, it would have been nice to have, but I hope you can understand why it languished.

Edited: Jan 16, 2:05am Top

>79 timspalding: Do you mean you have set it up so that one item from a group can be easily separated away from the group? I can't get it do that for
Trying to separate
Maigret's Memoirs / Georges Simenon / (ISBN 0905712234) (separate)
since it looks like it belongs better with those omnibusses.

EDIT: now I see. I was trying to separate with middle click, but that does not work. We have to regular click and get the new fancy page. Thanks, Tim!!

Jan 16, 2:14am Top

>79 timspalding:

Awesome! Just untangled a mixed volume 1/2 case that usually would have taken a few minutes and about 20 tabs. :)

Jan 16, 4:34am Top

>81 timspalding: timspalding:

I don't think Henry Miller would be pleased

Jan 16, 5:42am Top


It seems the Wind in the Willows is going to need work, as people have combined and ignored notices, but I don't have the time or patience to work through it at the moment. All I did was remove the movies from the book so far.

Jan 16, 7:46am Top

>79 timspalding: This almost makes things TOO easy.

Jan 16, 9:42am Top

>79 timspalding:

I have wanted this for over a decade. It deserves more of an announcement than to be buried eighty posts deep in an announcement of a cute new toy (which is also cool, but don't get me wrong. But something we've wanted since before Talk existed? That deserves a post of its own.)

Jan 16, 10:21am Top

>87 lorax:

Will do. Busy morning.

Jan 16, 11:04am Top

Working from my amazon fire. Big fingers, small screen. Accidentally hit separate. Editions to separate box came up on page, ignored it and went on with combining. Several combinations later the box is still there!

Jan 16, 12:42pm Top

>89 ringman:

The first time you add a new element for separation, the old entries get cleaned up from the list. Hit that a few times last night and was wondering what was happening for awhile. It needs fixing though :)

Jan 16, 1:36pm Top

Obviously I'm doing something wrong because today every time I try to combine the page refreshes and is almost blank except for this message:

Combine works

Combine the following works?

No works

Have I forgotten how to combine?

Jan 16, 1:51pm Top

>91 Taphophile13:

Had you updated your browser lately? Any new plugins?

The IDs of the works you are combining are coming via the cookies so it sounds like they are getting lost somewhere along the way.

Jan 16, 2:00pm Top

>92 AnnieMod: Thanks. I don't know exactly what's going on but disabling Greasemonkey seems to have done the trick.
And I'm blown away by your numbers. My wrist starts to hurt after a while.

Jan 16, 2:04pm Top

>93 Taphophile13:

My knee was bothering me over the last few days so had a lot of sitting time - and a few podcasts to catch up on. :) Decided to get some combinations done - and got a bit carried away I guess :)

Jan 16, 3:26pm Top

I’m not seeing anything different. What am I missing?

Edited: Jan 16, 3:40pm Top

>95 lilithcat:

Click on "separate" inside of a combination page (normal left click, not a right one) and instead of going for a separation, it will add it a running list at the top so you can separate more than one together.

PS: You may need to disable some plugins if you cannot see it - they may be interfering.

Jan 16, 3:52pm Top

I don’t do left/right clicks. I’m on a Mac. (In fact, right now, as I’m traveling, I’m on an iPad.)

So I’m not sure how to get to that.

Jan 16, 3:55pm Top

>97 lilithcat:

Just try to separate on the line you are trying to separate the same way you had always done separations. Instead of it asking you if you want to separate (as it used to work until now), it will add it to the list at the top of the page for you and you can then separate multiple editions together. If you are doing a single separation, it will be one more click than usual.

Jan 16, 3:59pm Top

Oh now THAT is a sweeeet refinement! Was able to do multiple separations on the phone even.
Thank you, thank you, thank you!!

Jan 16, 4:00pm Top

>97 lilithcat:

Just a regular click. Same way you've always done it.

Jan 16, 6:44pm Top

I'v found the X which removes the separation box|

Jan 16, 10:58pm Top

But if you're on the regular editions page you don't get the multiple separation option? I has a sad. I could have used that ten minutes ago.

Jan 17, 4:07am Top

Love the new multi-separate tool! One request - after clicking separate to add an edition to the list, can you highlight or dim or somehow identify the editions that have have been added to the list? It can be difficult to keep track of which you've done, especially as the page moves when they're added up at the top.

Jan 17, 5:35am Top

A question. I just combined


with the warning that the author is not the same. I went to the authors' pages and I added to their CK the other form of the name. Did I do right?

Jan 17, 6:10am Top

>102 PhaedraB:. Wha?!!!
My joy balloon just went pop.

Jan 17, 6:18am Top

>102 PhaedraB: >105 2wonderY:
Clearly it's a beta feature that's gonna get rolled out everywhere once we finish testing it by completing these.

Right, Tim? ;-)

Jan 17, 6:37am Top

>25 timspalding:

Tim, is the multi-separate tool usable now? I just got https://www.librarything.com/work/3930281 which put together a number of different volumes.

Jan 17, 6:44am Top

>104 .mau.:

That's fine in this case although I guess a number of people won't always do this, and to be honest I am not hugely bothered about the practice. I don't add an alternate name when some work records have M.D. prepended (or appended) to the author's name for example.

However in some cases you have to be careful as there may be more than one person with the name so you should use author division and put your alternate name on the right CK.

Jan 17, 7:43am Top

>108 andyl:

I understand that there are a number of issues with people's names, but in this case I was confident they were the same person :-)

My only concern is that there are a lot of things in LT that I am not aware of, so I always fear to have broken something...

Jan 17, 8:45am Top

>102 PhaedraB:

Well, last night, that was a good thing. Because the tool wasn’t working right. It seems to be fixed now, but having that alternative was a godsend.

Jan 17, 9:05am Top

>109 .mau.:

You can't really break anything badly by adding an 'Other Name' record. At worst you can only make it a little bit more confusing for people temporarily.

Jan 17, 9:51am Top

>104 .mau.: You didn't do anything wrong, but it wasn't necessary for you to do that in order to combine the books. That warning is a reminder to double-check and make sure the books you're combining should actually be combined, but won't prevent you from combining them.

Jan 17, 4:49pm Top

Love the new multi-separate tool! One request - after clicking separate to add an edition to the list, can you highlight or dim or somehow identify the editions that have have been added to the list? It can be difficult to keep track of which you've done, especially as the page moves when they're added up at the top.

That's a good idea. But I hesitate at the work in this case. I'll take a look.

But if you're on the regular editions page you don't get the multiple separation option? I has a sad. I could have used that ten minutes ago.

Blech. Yeah, I guess it should go there too. Any counter-arguments?

Tim, is the multi-separate tool usable now? I just got https://www.librarything.com/work/3930281 which put together a number of different volumes.

I'm unclear what you mean. Yes? Are you perhaps thinking that when you separate, it will make a separate work for each edition, as before?

Edited: Jan 17, 4:59pm Top

>102 PhaedraB: >105 2wonderY: >106 davidgn: >113 timspalding: The multi-separate tool is available in the workbench too, so we can just stick a work there to dissect it already even if it's not part of the "Combination Opportunities" project.

Edited: Jan 17, 5:18pm Top

>113 timspalding:

If you do not want to add it to the editions page itself, a link that opens a new "combination page" containing only the work may also work.

Less development I suspect... Or just add a note for people to use the workbench

PS: And >114 Jarandel: beat me to it while I was not pressing send around here.

Jan 17, 5:46pm Top

>114 Jarandel: Unfortunately, if you don't use the workbench to separate, you never see the possibility.

Jan 17, 7:25pm Top

When you separate items from one of the 'Potential" works using the multi-separate tool, the items "disappear" and it can be hard to combine them back with the correct "Potential" work. There isn't the chance to add them to the workbench, and if you use back-arrow to return to the previous page (and re-load it) they don't show up, even if the have the isbn in question.

I've found that by going to my Share page I can go to these separated items to get the opportunity to combine them back into where they belong.

Jan 17, 7:29pm Top

I just open the newly separated work in a new tab and use back to come back into the "opportunity". They do not show up because the opportunities have the IDs attached to them recorded somewhere - which is why if you combine one of the found ones and it changes ID, it disappears as well.

Jan 17, 8:29pm Top

>118 AnnieMod:


In https://www.librarything.com/combine.php?combination=method1:9750452 there were 4 Vol II mixed in with the original "The Apple Barn Cookbook" which Amazon shows me are distinct works, volume 1 not being labelled as such because a second volume wasn't planned at the outset. The original work is ISBN 0961150823 which the Combination Opportunity page is about.

If I click on the "Separate These Editions »" I can return to
Original works
The Apple Barn Cookbook (editions)
Separated work
The Apple Barn cookbook Vol.II (editions)
Apple Barn Cider Mill & General Store (author combine/separate page)
Combination Opportunity

I actually want to go back to the previous page, as I now want to combine the residual items. I can right click on the Apple Barn cookbook Vol.II (editions) to open in a new tab but the newly separated works don't necessarily show up in the editions page, even if I do a recalculate. I can find them using my Share page.

It would be nice if the multi-separate tool gave me the option of going back to the Opportunity I was working on.

Jan 17, 9:59pm Top

>119 rodneyvc:

I open the new page and if the combination is not on the editions page, I add it to the workbench, go back to the opportunities page and open whatever else needs to be combined with that one (and add them to the workbench as well).

I do not disagree that it will be useful - just mentioning that there is a workaround.

Jan 18, 6:28am Top

>113 timspalding:

There are at least four different books in that work, and I don't know if there is a possibility to select a number of editions and put them in a new work.

Jan 18, 9:43am Top

It feels as if the percentage of difficult tangles and books I know nothing about and where the titles and ISBNs are of little help is increasing. Is it all the Skip to next's that are being rotated back into the pool?

Jan 18, 9:55am Top

>122 anglemark:

I expect so. I think that some of us are knocking off the easy ones first. I know that I’m skipping a lot that look as though they need a lot of investigation to figure out what’s what.

Jan 18, 10:07am Top

And I'm wandering off to try to fix some that are better served from the editions page or the author page.

Jan 18, 1:24pm Top

>124 2wonderY:

I went down a "author nevering" rabbit hole yesterday for almost an author - which contained a ton of combination issues, stranded links and what's not :) So yeah - I think that a lot more is being fixed around these opportunities than the opportunities themselves.

Jan 18, 1:25pm Top

Although the "opportunities" aren't about author combination per se, they've routinely pointed me at author combinations.

Jan 18, 3:45pm Top


I've been out of the programming game too long. C and C++ are two different languages, in essence, right? So they shouldn't be combined as the same? If so, this combo is a cluster I fear digging into...

Jan 18, 3:47pm Top

>127 gilroy: Yeah, I remember passing on that one too.

Edited: Jan 18, 3:57pm Top

Since C isn't an object-oriented language the title "C Primer Plus: Teach Yourself Object-Oriented Programming" doesn't make any sense. I suspect that the "++" got dropped by accident and those books from the C group that match the books with the title "C++ Primer Plus: Teach Yourself Object-Oriented Programming" in the second group should be reassigned.

There are also several books in the C group with C++ in their title that also ought to be reassigned to the C++ group.

ETA: In fact I'll do just that

Edited: Jan 18, 4:10pm Top

For a long time, LibraryThing was incorrectly handling "++" in book titles (here's one related bug), so there are quite a few books on C++ cataloged with C in the title. This one of the rare cases where I pay attention to the ISBN over the title.

Edit to add: I think the bugs are fixed now, but the bad data lingers.

Jan 18, 4:17pm Top

I untangled a few of those c/c++ books yesterday and added Susan big notes - explaining the issue with the LT/amazon handling of plus in the past and a note not to combine.

Jan 18, 6:18pm Top

Tim, I'm wondering if, after all 100,000 Combination Opportunities have been settled, you could provide us with some statistics. Like, how many combinations / splits / author combinations / author nevers (etc) were done during the duration of this project? How many more than during a typical week? (Or however long this is going to take -- we're burning through them quite quickly, I'd say!) Were more people involved than usual? Did non-combiners or infrequent combiners log more combinations/splits etc? What proportion of a member's Helping efforts are made up of the Combination Opportunities?

Is there a measure of how "combined", or disorganised LT's databases are? Do you expect our combining efforts for this script to make a sizeable difference to that measure? Is there even a way of estimating that? Is there a way to compare the subset of the database that was captured by the script before and after? Like, how many individual Works and Authors before, and how many after, and how many splits/combinations etc. to take us from one to the other?

On a related note, I suppose that there'll be plenty Opportunities again if you run this script again, say, a year from now, or something similar that captures a different set of Opportunities. Will this be a yearly or semi-regular occurrence? Are there more than just the 100k selected by this algorithm?

Jan 18, 6:56pm Top

With all the changes in works is this slowing down recommendations. I seem to be getting a lot fewer since this project started.

Edited: Jan 18, 7:07pm Top

>133 jjwilson61: Been getting a few new ones everyday as usual, most important factor seems to be if I've been adding works to my library that were conducive to generating recommendations.

That is, not so completely obscure that they're probably below some threshold for ability to generate something meaningful, and not so typical to the rest of the library that all or most likely candidates were probably already suggested earlier.

Edited: Jan 18, 7:48pm Top

>133 jjwilson61:

No. Recommendations have actually gotten a lot more frequent recently. In the last few months we've made a concerted push there, and no recommendation is more than 3 months old. I'll post about it later, if I can.

Jan 18, 9:23pm Top

>135 timspalding:

That explains why I've been seeing a lot of new recommendations lately without having added many new books. Sadly most of them aren't very good - lots of raw popularity, it looks like.

Jan 19, 12:41am Top

>133 jjwilson61:, >134 Jarandel:, >135 timspalding:, >136 lorax:

Yeah, I too have seen a flood of recommendations lately. Almost too many actually - some days there's been more new recs than fits in my recommendation module.

Edited: Jan 19, 7:32am Top

FYI, here's my internal chart of recommendations speed. The lines are months. "Now" is the gray line on the right. (The two low red bars after that are ones that haven't been done, and ones that are much older, for obscure reasons.) Basically, everything has been regenerated since the beginning of November. Speeding the process up, regenerating, and having it not sink the servers has been a major concern since November.

Jan 19, 8:05am Top

>138 timspalding:

Since you've been catching up, does that mean that the rate of new recs will decline somewhat again?

Jan 19, 8:06am Top

Is there a measure of how "combined", or disorganised LT's databases are? Do you expect our combining efforts for this script to make a sizeable difference to that measure? Is there even a way of estimating that? Is there a way to compare the subset of the database that was captured by the script before and after? Like, how many individual Works and Authors before, and how many after, and how many splits/combinations etc. to take us from one to the other?

I don't think there can be any absolute measure. You can't really know what books should be combined until you look. Think a book that was republished under another name—nothing in the bare data will tell you that.

The 100k I've made, however, are a good metric, and represent the most fertile source of combination opportunities. They represent almost every case where an ISBN crossed works.* Most of those opportunities are real--there really is some combination work to do. A minority are user or publisher error.

When this is done, I'll probably run it again. All the movement will have created new opportunities--but only a few thousand, I bet. Time and new cataloging will produce new ones of this type too, but on a very manageable scale.

After this, the next greatest opportunity is, I think, very close correspondences between titles and authors between works--cases where only a "The" separates the titles on a work. But this is tricky--authors will often publish a slew of books with very similar names. Most of this work will be non-ISBN works, which are quite the minority.

You've probably noticed that works without an author are a major factor. This is another field of combination--looking for works that have no author in any of their editions, but match the title of another work perfectly. Those without an ISBN, will often have to remain non liquet. Those with an ISBN will have mostly been resolved in this first 100k.

There's also some potential to use other clustering information we have--we have "work" models from library data, for example. But the data is worse than ours overall--LibraryThing's edition data is probably the best in the world. So it comes with problems.

Anyway, if others have ideas for finding potential opportunities algorithmically, I'm all ears.

When this is all over, I can certainly report on how many works were involved, and how much the daily separation increased over the average.

*I left off two categories:

1. Works that have no copies, unless part of a larger complex that does. If you've seen combination opportunity that was gray, this was that. There are a fair number of these.
2. Very very large works. There are only a few hundred of these, but I decided NOT to include the Hunger Games, Harry Potter, etc. There are only a few hundred of these.

Edited: Jan 19, 8:29am Top

>140 timspalding: "...ideas for finding potential opportunities..."
How about authors that are split and the works aliased to another author all because no one took the time to look at the editions layer (probably the best in the world) to find that the author to whom they wish to send the work is already the author?

most recent example:
4am-5am EST (9:00-10:00 GMT)
Guanhumara assigned Religion and national integration in Africa : Islam, Christianity, and politics in the Sudan and Nigeria (editions) to John Hunwick (1)
Guanhumara assigned Project: parks - an introduction to S.A. National Parks and their management (editions) to John Hunwick (2)
Guanhumara assigned Ecosystems, National Parks and Wildlife Preservation - a conceptual approach (editions) to John Hunwick (2)
Guanhumara saved John Hunwick with "Split the author."

Jan 19, 8:36am Top

I don't have a suggestion for combination potentials.

I do have a suggestion for another project like this:
Serve up works with multiple different authors in the editions, so we can check if they are the same work, and add/fix/check the Other Authors section.

Jan 19, 8:41am Top

>140 timspalding:
Thanks for taking the time to answer my questions!

As for "ideas for finding potential opportunities algorithmically": how about author names consisting of multiple words, one of which is "&" or "and" or "(ed.)" or "editor"?

Sometimes, all that prevents books from being autocombined is a disconnect between the title and the subtitle: either the subtitle is not there (and so one edition's title will be entirely included in another edition's), or the separator takes different forms (colon with or without additional space in front; dash; hyphen).

Edited: Jan 19, 8:42am Top

>142 JerryMmm:
You beat me to it!

Jan 19, 8:53am Top

>141 Collectorator:

I don't follow.

Serve up works with multiple different authors in the editions, so we can check if they are the same work, and add/fix/check the Other Authors section.

Yes, agreed. We could probably come up with a list of "co-author?" works to check.

Jan 19, 9:06am Top

>143 Petroglyph:

Or the subtitle differs between editions; I see this a lot for UK vs US publications of the same work. There are of course cases where

Title: A Subtitle by Author


Title: Another Subtitle by Author

are legitimately different works, but it's a well-defined case where lots are probably appropriate to combine.

Jan 19, 12:17pm Top

I was getting tons of new recommendations a several weeks ago but they have dried up recently, even after adding a couple of books. Guess it's luck of the draw.

Jan 19, 12:56pm Top

>140 timspalding:

- Books with the same name (removing all kinds of brackets and all kinds of "by YYY" statements from the titles where one of the books has an author and the other does not.

- I'd even say to get all books with the matching names and throw them into opportunities - so we can mark as ok the ones that are legitimate duplicate titles and combine the rest.

Jan 19, 1:56pm Top

>147 Collectorator:

No, really. I don't see your beef. I need to have things explained to me; I can't read minds.

Jan 19, 2:35pm Top

>150 timspalding: I guess it isn't really a combination issue after all, so never mind. I have no idea how a script could be invented that would uncover the misuse anyway. (nor any other script for that matter)

Jan 19, 5:13pm Top

>140 timspalding:
Books with similar or identical titles where one edition has an author's gender of n/a -- often an organization of some sort -- and the other edition can have anything else. (I'm seeing a lot of those cases: art books and museum guides and official reports and the like where some editions have a human author/editor and others have the institution listed.) I suppose such "institutional" publications, sometimes without isbns, and with multiple candidate authors and several ways of formulating the title, will be the ones that need Combiners' attention.

Perhaps CK can be leveraged algorithmically in other ways: get rid of duplicate series (where the same set of books has multiple series that share words); get rid of duplicate Series and Publisher Series.

Perhaps a push to separate identically-named books and movies, using the media field? I understand that many users add the book and then manually edit the non-isbn info to reflect their dvd, but some cleanup should be possible.

Jan 19, 6:48pm Top

One of the combination opportunities allowed me to stumble a couple days ago on (a?/the?) current Black Hole work.

Apparently it was not exterminated, just in hiding over the past few years.

Jan 20, 8:17pm Top

Moving the thread up.

Jan 20, 9:41pm Top

If i change my username to MrAndrew(4,500), will i top the All-time list?

Jan 20, 10:01pm Top

>155 MrAndrew:

Only if I don't change my user name to Lilithcat(5,000)

Jan 20, 10:02pm Top

>155 MrAndrew:, >156 lilithcat: The match is on!!

Jan 20, 10:11pm Top

That way madness lies. And Mersenne primes.

Jan 22, 8:30pm Top

I combined these works (https://www.librarything.com/work/5557071/editions), but now I am not sure about the author. Amazon has one spelling and WorldCat has another. Suggestions? Thanks.

Jan 22, 8:33pm Top

>159 krazy4katz:
The cover spells it "Joy Cowley", so I'd go with that.

Jan 22, 8:35pm Top


Edited: Jan 22, 8:57pm Top

>160 Petroglyph:, >161 Stevil2001: Both wise suggestions. Thank you! I will fill in the canonical name as Cowley. There are no other works listed under Coley. If you think that is the wrong thing to do, please let me know. Thank you!

Jan 23, 7:52am Top

Many of these Combining Opportunities are works where some copies are lacking any author or where the work has two or more legitimate authors and there are copies with the various names. I have been combining these without any other changes. What do others think about going in and editing the authors field? I can't really see the point, although I often do this if I come across a work where the 'author' field is actually filled with a combination of two or more names

Jan 23, 8:32am Top

How would you "edit" the author of someone else's record? There are some cases where author combination is clearly needed--a difference of a middle initial, a "Ph.D." added or misplaced, that sort of thing--but be careful with that. It's easy to make an author combination with an already-split author and create a tangle.

Jan 23, 8:46am Top

>163 abbottthomas:

What do others think about going in and editing the authors field?

I'm doing that when I see no author, or "smith john susie jones" and the like. Often, I'll see "John Smith" on some copies and "Susie Jones" on some, and then I'll often go in and add the other name as an "other author".

Edited: Jan 23, 11:03am Top

>163 abbottthomas:

If there are two legitimate authors, I confirm the other other author (most of the time anyway). This way the book does not just disappear from one of the two authors after the work records are combined.

>164 paradoxosalpha:

You do not. Technically you can force it by overwriting the primary author but in cases like this just leave it be. If there are authors to be combined, do that but don't change the names on the record - it causes a lot of interesting combination issues later on... So unless it is really bad (and having a Jr or PHD in the wrong place does not cut it for me), just let LT sort it out when more copies are added.

Jan 23, 10:57am Top

>164 paradoxosalpha: How would you "edit" the author of someone else's record?

You can edit the author of any work from its main page - scroll down to and click on 'Add Other Authors' and, as well as being able to add another author, editor, illustrator or what you will, you have the option to change the existing principal author. This option should be used with care.

>165 lilithcat: Clearly you are creating a more accurate record by adding the alternate name although it does seem easy to end up with one of the names showing up twice on the main page heading. Do you think you gain much else from doing it? I rely on the system deciding on the most popular name and displaying that.

Jan 23, 11:01am Top

>166 AnnieMod: This way the book does not just disappear from one of the two authors after they are combined.

Thanks. That is worthwhile.

Jan 23, 11:03am Top

>168 abbottthomas:

They being the books, not the authors of course - which is clear to anyone in this thread but is ambiguous on its own - I will update the post above. I need more coffee.

Jan 23, 3:21pm Top

I really hope tim will make >142 JerryMmm: a reality next.

Jan 24, 3:21pm Top

I've been seeing some odd work combinations, but this takes the cake:

Under "Osmotic regulation in Chlamydomonas pulsatilla by Paul Valéry" (weird in and of itself), accompanied by this cover image: https://images-na.ssl-images-amazon.com/images/P/8880740431.01._SX450_SY635_SCLZZZZZZZ_.jpg, I found:

Opere poetiche/Valéry, Paul/ISBN 8877462744 (1 copy separate)
I Signori. Le nuove creature./Morrison, Jim/ISBN 8880740431 (1 copy separate)
L'abate C/Bataille, Georges/ISBN 8895249135 (1 copy separate)
Sonetti dal portoghese/Barrett Browning, Elisabeth/ISBN 8889299215 (1 copy separate)
Il racconto d'inverno/Shakespeare, William/ISBN 8811364361 (1 copy separate)
Poesie 1952-1973/Evtušenko, Evgenij A./ISBN 8811638003 (1 copy separate)
L'abate C./Bataiile, Georges/ISBN 8895249135 (no current copies separate)
Oper poetiche/Valéry, Paul/ISBN 8877462744 (no current copies separate)
L'abate C/Bataiile, Georges/ISBN 8895249135 (no current copies separate)
Osmotic regulation in Chlamydomonas pulsatilla/Otto, Tunde A./ISBN 0612191095 (no current copies separate)


Jan 24, 3:39pm Top

>171 lilithcat:
Whoa, that's a mess. None of mine have been that disparate.

Jan 24, 3:48pm Top

I've got "A Man Called Spade", https://www.librarything.com/work/1975202, and "The Adventures of Sam Spade and Other Stories", https://www.librarything.com/work/1513568, which both contain "Mies nimeltä Spade ja muita kertomuksia / Dashiell Hammett / (ISBN 9510193909)" which translates as "A man named Spade and Other Stories."

I almost pulled all of the latter editions out of both works to combine them into a new work but it occurred to me that its possible that "The Adventures of Sam Spade and Other Stories" and "A man named Spade and Other Stories" could contain the same stories. Should I just pull "A man named Spade and Other Stories" out of "A Man Called Spade" since they're definitely not the same and combine them with "The Adventures of Sam Spade and Other Stories" since it already contains the other matching editions?

Edited: Jan 25, 5:19pm Top

Longtime combiner here. Just learned about this feature. Love it, particularly b/c it looks for ISBN matches for those "A by B (ISBN C)" and "A (ISBN C)" examples that Tim mentioned above. Thanks!

Jan 25, 4:38pm Top

Okay, take a gander at these graphs, showing growth.

I put them on twitter, where you can expand to see the large size. There's no easy way to do that here.


Jan 25, 5:05pm Top

>175 timspalding:
Thanks for those graphs! Exactly what I wanted to see!

Edited: Jan 25, 9:22pm Top

So I am a long time LT member and have been doing combining for a long time now, but only for the books I catelogue or look up. Not sure about doing any complicated ones. But what I have been wondering for a very long time is how to prevent this problem in the first place. Why not create some up front edits. No title without an author for one. People put all sorts of weird things along with the title that are better fit for tags or data elements that can be found under "edit my book." Remind them not to do that or disallow it. Even messaging, rather than edits would help, reminding people to search for their book or work first. Is any of that possible?

Jan 25, 11:22pm Top

BTW, I'm encountering quite a few works with the description/review as part of the title, such as:

Nelson Mandela: Long Walk to Freedom-You may know that Nelson Mandela was a South African anti-apartheid activist who was imprisoned 27 years, elevated to President of South Africa in 1991 and subsequently won the 1993 Nobel Peace Prize. He was a champion of children’s rights and once said “Education is the most powerful weapon which you can use to change the world.” He wrote about his extraordinarily inspiring life in his 1994 autobiography “Long Walk to Freedom.” Chris Van Wyck later edited the / Chris van Wyk / (ISBN 1596435666) (separate)

Annoying, but I suspect there's not much that can be done about them.

Jan 25, 11:38pm Top

>177 Micheller7:

No title without an author for one.

I have a lot of books with no author credited. Why shouldn't I be able to catalog them?

Jan 26, 3:37am Top

>179 lilithcat: Conceivably there could be a checkbox with "No author given", but I agree, I also have a number of such books.

Edited: Jan 26, 5:44am Top

>177 Micheller7: You have a lot of people that just get frustrated with the detail oriented nature of the site. If you start making them do more just to add a book (which already has many new users throwing up their hands in defeat), we'll never grow.

You also have people who come in, try the site, don't like it after their first attempt and never come back. They aren't going to alter their data. Then what?

And when you say remind people to search for their book or work first, do you mean on site or through the add book tab? Because they already do the former, which is what makes them frustrated that they can't just click a button to add it to their own library. If the later, we've got notes all about that get ignored. (On a site about books and reading. How Ironic...)

Jan 27, 3:49am Top

>10 davidgn: Half way in 13 days! Good estimate.

Jan 27, 5:48am Top


What should be done here? There are multiple editions within one work, should each edition, 1st 2nd 3rd etc. be a different work? Or should they all be combined as they are all the same work?

Jan 27, 8:15am Top

>183 OwenRochester: I struggle with this too. It possibly depends on knowing how different the various editions are, which revolves on way more familiarity with any given work than the average combiner.

This leaves us with either always choosing to lump them all together, or always separating into individual works, neither option seems totally satisfactory.

Jan 27, 9:34am Top

I'm sure people are philosophically apposed to this, but my position with textbooks is to lump together, based on the "cocktail party" test: obviously there will be some differences, but each version is essentially serving the same social function. Especially if it's not a very common textbook, so that if it were split, you'd have a bunch of works just owned by a couple people. (If someone with actual knowledge of a particular textbook thinks differently, tho, I am happy to bow to them. But as a college prof, most editions just seem to push chapters around a little bit.)

Edited: Jan 27, 12:04pm Top

Unless someone split specifically and added disambig notes explaining the reason, I'll combine them. The other day there was one with 90 editions and a nice note explaining how 45 changed the idea of the thing, dropping half the content.

Edited: Jan 27, 12:31pm Top

For the purposes of this exercise I'd just go with what's already been done. If they're mostly separated into different editions already, maintain that split, otherwise lump them.

Jan 27, 1:04pm Top

Different volumes, cataloged as separate books, should be separate.
Different editions, as stated above, unless noted as fully overhauled from previous editions to be a completely different work, go together.

Jan 27, 1:15pm Top

>187 jjwilson61:

There ye go...

Jan 28, 3:04pm Top

Jan 28, 7:50pm Top


According to the title, the work with the matching ISBN in the second set belongs where it is. However, I separated it and according to the ISBN, the cover selected for that book by the owner and WorldCat, it belongs with the first group.

What to do?


Edited: Jan 28, 8:48pm Top

Title beats ISBN. Mark it as "Okay as it is" and move on. If it was missing the subtitle as opposed of having it where it does not belong, it can go either way but in this case, it looks like someone added the wrong book and fixed the title only...

Jan 28, 11:04pm Top

>192 AnnieMod:

Title beats ISBN.

I think this is true. Would you agree that ambiguous title gets trumped by ISBN? So, for example, if you have

Book (ISBN A)
Book, Volume 1 (ISBN B)
Book, Volume 2 (ISBN C)
Book (ISBN B)

That "Book (ISBN B)" should go with "Book, Volume 1 (ISBN B)"?

Jan 29, 12:36am Top

>193 timspalding:. Yeah, in most cases - especially if this ISBN belongs only to volume 1. If it is a book that is rarely split though (a novel for example) I may decide to leave it alone - there is a better chance that they added volume 1 and took out the numbering than the opposite. Not too many of those had been found though :)

In multivolume known sets with ambiguous titles as above - yes, I'd go with the ISBN.

Jan 29, 3:40am Top

What's going on with ISBN 0833211269 - see https://www.librarything.com/combine.php?combination=method1:6445570

Is it some sort of example or placeholder ISBN? I hit skip!

Jan 29, 5:25am Top

>195 rodneyvc: It looks like a publisher (possibly a self publisher) reused the same ISBN for multiple books. Would take some work to weed out, admittedly.

Jan 29, 5:34am Top

>192 AnnieMod:, >193 timspalding:
Yup. If the distribution of the isbn leads me to suspect that it might belong to that particular volume (i.e. if a few "Volume 2" titles have isbn B, and that isbn doesn't occur with unambiguous "volume 1" and "volume 5" titles).

Jan 29, 6:48am Top

>195 rodneyvc:
That example has been taken care of. Here's a similar one: https://www.librarything.com/combine.php?combination=method1:2406832

Jan 29, 6:55am Top

I agree with the ISBN/title principles laid out in this thread.

Jan 29, 7:17am Top

>195 rodneyvc: Good job. I've come across situations where an ISBN was used on two different works, but 16 otherwise unrelated works is a whole order of magnitude messier. I can more easily understand people using the same ISBN for individual volumes of a 22 volume set.

Edited: Jan 29, 11:25am Top

This message has been deleted.

Feb 7, 8:38am Top

>140 timspalding:
if others have ideas for finding potential opportunities algorithmically, I'm all ears

Zero-copy works that share a title and an author with a non-zero-copy work -- or part of the title/author name (I've just combined a bunch of zero-copy works by S. Carmiggelt into ones by Simon Carmiggelt, but that kind of case crops up regularly.

Feb 7, 10:04am Top

It always seem a pity that zero copy works can't just slowly shrivel up and die. They do seem to cause an undue amount of trouble. I have read, and think I understand, the reasons given for keeping them in their un-dead state, but would they really be missed?

Feb 7, 10:35am Top

>203 abbottthomas:

They help getting other copies combined automatically - especially when their exact name/author/ISBN combination is coming from a source and it is repeatable....

Feb 7, 12:56pm Top

>204 AnnieMod: I've heard the argument before and I'm sure it is right but is it worth the downside of the collection of non-existent crud lying at the bottom of most works with a large number of copies?

You say:
....especially when their exact name/author/ISBN combination is coming from a source...
Particularly in this Combination Opportunities thread we are seeing large numbers of zero copies which do not meet that criterion.

Feb 7, 1:05pm Top

>205 abbottthomas:

But we should not just ignore all we have just because some of it is bad...

We need a "ratty data" flag. :) This will allow the legitimate 0-copies to be separated by the riff-raff - and then the riff-raff can be safely cleaned out. Plus this can also allow things with mixed ISBNs to be prevented from attracting more copies. As it is, I would much rather deal with the 0-copies than end up with a lot more uncombined works - especially in smaller authors, there are a lot of uncombined works.

Feb 7, 1:27pm Top

>206 AnnieMod: I'm with Voltaire, or rather Evelyn Beatrice Hall, on this ;-)

Feb 7, 2:37pm Top

Can we at least get rid of zero copy works and editions which were entered manually. These have in general been corrected by their originators and have no value whatsoever.

Feb 7, 4:40pm Top

>208 ringman: Everytime I make a dumb mistake when manually entering something it pains me to think of the ridiculous 0-copy I just caused to happen, especially because the kind of mistakes I make are often ones that would cause problems later. (For instance, adding several books in a series, and copy/pasting things into the wrong fields) I don't think I've ever added anything without a typo I immediately had to fix.

This is also why I hate to edit from the edit page and always do it from catalog view where the information is denser and I don't have to scroll.

Feb 11, 7:48pm Top

Oh crap. I go away for a month and look what happens. Tim, why do you hate non-LT productivity? lol This has already completely pulled me in and I only discovered the thread like ten minutes ago. This is SO much easier then poking around at random works to see if anything needs combining (which I used to do a lot when bored).

Feb 15, 2:16pm Top

Lots of them are cycling 'round again. Some I fixed else-page and just need confirmed as okay. Others are just too messy for me to tackle. And then others are do-able because I'm on laptop rather than iphone.

Feb 15, 3:22pm Top

Don't be afraid to mark things as fixed or already done. I didn't make a link for "insolvable." It wouldn't be the end of the world if an unsolveable mess was marked as such, even if one person out of 50 could solve it.

Feb 19, 8:52pm Top

>140 timspalding: timspalding:
if others have ideas for finding potential opportunities algorithmically, I'm all ears

Is there a big enough dataset of works with identical titles (or essentially identical, just with different title casing) within the same author record?

I see this problem commonly on new books, where the print and ebook editions have different ISBNs and thus are separate works until combining.

Feb 21, 1:15pm Top

>214 krazy4katz: Mark OK as is. That's one screwy ISBN.

Feb 21, 1:44pm Top

Will do. thanks.


Feb 24, 4:25pm Top

This is an example of my comment in >213 feldamundo:: Grace Bonney
3 records for "In the Company of Women: Inspiration and Advice from over 100 Makers, Artists, and Entrepreneurs", each with identical titles but with different ISBNs under the same author record.
Are there a lot of other examples like this that could be solved algortihimically?

Feb 24, 7:46pm Top

We're making remarkable progress in what may turn out to be one of the shortest three weeks on record. ;-)

Nonetheless, I suppose it's worth thinking about other algorithms that might be useful to find things needing combining.
From something I'm working on: perhaps situations like this (and similar situations) might be generally findable?

(With my pseudo-tags. Imagine angle-brackets)

(ti)Diego Rivera: His World and Ours(/ti) by (au)Duncan Tonatiuh(/au)
(ti)Diego Rivera: His World and Ours by Duncan Tonatiuh (2011-05-01)(/ti)

To generalize:
(ti)TITLE(/ti) by (au)AUTHOR(/au)

Feb 25, 6:43am Top

>217 feldamundo: Those are already in place, but won't start to work until after the first time the ISBNs get combined.

Feb 25, 3:49pm Top

On the home stretch! Fewer than 10,000 remain to be done.

Feb 25, 5:46pm Top

What do you do when the ISBN is the same, but the title is not. Then you separate that work, look it up on WorldCat and get a third title? The only think I could do was let it drift into space… I feel as though I am doing more harm than good sometimes.

Feb 25, 6:23pm Top

>221 krazy4katz: Skip to next, I'm afraid ;-(

Feb 25, 6:55pm Top

>221 krazy4katz:

Title beats ISBN except for multi volume sets where you do educated guesses - so if the title is somewhere in LT, conbine there. If not, leave it where you found it or separate it out and let it float - if the title ever get added, they can be reconciled.

Edited: Feb 25, 8:36pm Top

Ok, I messed up. I was working on works by Marc L. Miller, who is also Marc Miller (3) on the disambiguation page. I combined the 2, but now all I get is the one work on the disambiguation page. The real Marc L. Miller is here: https://www.librarything.com/work/1132845 and here: https://www.librarything.com/work/9614119 etc. He wrote a lot of books but I can't get to them anymore. Help! I am stopping before I do anymore damage.

Feb 25, 9:10pm Top

>224 krazy4katz: Don't combine Marc L. Miller with Marc Miller when Marc Miller is a split author.

Feb 25, 9:11pm Top

>224 krazy4katz:

I think that someone was working while I was trying to untangle.

Do you see them here: https://www.librarything.com/author/millermarc and https://www.librarything.com/author/millermarcl here now?

In such cases don't combine the authors, you need to alias the proper (3) into the actual author.

Edited: Feb 25, 10:06pm Top

At least Marc L. Miller is back. Thank you!

Marc Miller (3) - who is the same person - has disappeared. I think there was only one work under that name I don’t know if it is now under Marc L. Miller.

I’m not sure I ever understood aliasing, although I might have at one point. I have stayed away from it.

Unfortunately I separated a singleton from a work under Marc L. Miller, whose author was listed as Marc Miller, and that started the problem.

Thank you again. k4k

Mar 9, 7:00pm Top


Edited: Mar 9, 8:47pm Top

>228 krazy4katz: Someonw used other authoring to send the book under the author I would have aliased it in (I do not remember if I did and someone undid it) - so there is no book in the system that belongs to Marc Miller(3) anymore. That's why you are not seeing it.

All should be looking fine now. Or are you still missing books?

PS: If you want to try your hand at aliasing - open the page where you assign books to the different split sections. There is a tab there for aliasing. :)

Edited: Mar 9, 9:15pm Top

Oh, sorry! I bumped to get this back on the first page. But thank you for the advice on aliasing! I never realized there was a tab for aliasing. I will try it next time. k4k

Mar 9, 11:06pm Top

>230 krazy4katz: Now its on everyone's first page. Couldn't you just star it instead?

Edited: Mar 16, 10:46pm Top

>231 jjwilson61: Sorry. Will do that next time.

Mar 10, 7:41pm Top

Only 5000 to go.

Group: New features

44,651 messages

This group does not accept members.


This topic is not marked as primarily about any work, author or other topic.

About | Contact | Privacy/Terms | Help/FAQs | Blog | Store | APIs | TinyCat | Legacy Libraries | Early Reviewers | Common Knowledge | 123,919,996 books! | Top bar: Always visible