Make it easier to see the date when a book came out/show green text data for books you don't own

{Repost from _Talk about LibraryThing_ https://www.librarything.com/topic/354163}

LibraryThing can be very unhelpful when you want to know when a work came out. If you are lucky, someone has put it in Common Knowledge for the work.

But otherwise, even if there's only been one edition of the work, the only way to get its publication date seems to be to add the edition. Consider this work: https://www.librarything.com/work/693691 I wanted to know if it was recent enough to be worth tracking down and reading.

No dates in Common Knowledge or Work Details. No "edit your book" section, which would most likely have the info for a particular book, *if* I owned it. (Of course in that case I could look at the title page; I mostly want this information for books I do *not* own.)

The editions page informs me that there are 5 "editions", all of them with ISBN 0195179390. I.e. there's actually one edition, but the LT system somehow created 5.

But there's no way for me to see the "book" page for any of these 5 presumably-identical "editions". I can't do anything with this list but separate some of these "editions" from the work. (Not a good idea; they're obviously the same work.)

The closest I've been able to get to a publication date is looking at the dates on the member reviews; the earliest is Oct 2, 2006.

Goodreads, on the other hand, tells me "first published in January 1, 2005". Moreover, it does it on the first page I see after a search by title. (I'm not counting GR's repeated attempts to get me to login or create an account, which I closed/ignored.)

This is, of course, a fairly common case - only one edition, with the publication and copyright dates essentially the same, so the date on the edit-your-book page would have given me what I wanted.

The other case - lots and lots of editions - is much harder, because there are many dates, and it's hard to know what the user wants. Fortunately books in that group tend to have a filled out common knowledge section.

Is there any way we could make this information available to people who might e.g. be looking at a recommendation and wondering whether to acquire the book, at least for the simple (single edition) case?

Go to the WorldCat page.
WorldCat shows editions from 2005 and 2006. It's one click. Very easy.

>2 MarthaJeanne: By the same logic:

Go to Goodreads.

Go to Goodreads is the deliberately provocative, non-answer. The real questions for me are where does the GR data come from and is it accurate and reliable? Also, there is no edition level here on LT despite a decade of begging, so there are other things going on that make this a difficult request. Having spent a lot of time putting OP dates in CK for books I don't even own, this is a sore spot with me since I like knowing that most of the time. Having a consistently accurate and available data source would be paramount for it to work and if that can't happen I'd rather have things the way they are.

I don't know that publication dates ever appear as green text data, so I suspect I have no idea what you're proposing should happen. Unless someone enters the original publication date, there's no way to tell. Earliest edition from members who have it catalogued isn't always OPD even for modern books, let alone old ones. CK has to be populated by members, so if you have an active interest in a book's OPD, spending a minute on search and entering it is what we do. That's why a lot of books have that data.

>4 Bookmarque: That wasn't an answer. It was a critique of the previous answer.
>5 SandraArdnas: It seems you're correct about the original publication dates. I agree that if we don't have the data and we can't reliably get it, this is probably a non-starter.

Oh now come on...most of the books entered and read are modern and thus have an OPD that is identifiable. Listing outliers like the bible and stuff by Homer is just stirring up complexity that won't matter much. If an OPD could be consistently sourced and was accurate, I would welcome it.

And you've deleted it so I assume you came to a similar conclusion. : )

When the Series system was updated in 2020, Tim mentioned that there is an estimated publication date for each work, which can be used to sort works in a series by publication date: https://www.librarything.com/topic/321038#7179244

A user asked: "When viewing a series in "Publication" order, where are the dates coming from? I assumed Common Knowledge, but I've found a few where the publication date on the series page is different from the Common Knowledge "Original publication date" field. (For example, The Claw of the Conciliator has a publication date of 1657 on the series page for The Book of the New Sun, but an "Original publication date" of 1981-03 on the work's Common Knowledge page.)"

And Tim answered: "The publication date comes from book data—Amazon, Bowker, Overcat, library records and CK. It's a very complex weighing of factors—the data always disagrees—which then "bubbles up" from ISBN to work. In its broadest outline, it's the date the most book records say it is. But it gets complex from there.

That said, CK should trump it. It isn't doing so here. It may be caching or a it may be a bug. I'll investigate."

That implies that there already exists somewhere an estimated original publication date for each work, and if there is I do think it should be made publicly visible (in green text, since it's an estimate).

>8 Bookmarque: There's a world of books between the Bible and contemporary ones. Many are highly represented on LT (think early SF and Fantasy, for instance) but not in their first edition. Ergo, it is not possible to get even remotely reliable OPDs from member data. First editions of many, many books are expensive collector items, not something habitually entered here.

>2 MarthaJeanne: Are you saying there's a link to worldcat from the LibraryThing page, that I haven't noticed?

>10 SandraArdnas: ok then, perfect will be the enemy of good. No further discussion needed.

>11 ArlieS: Yes. I use it a lot. You may have to set it, but it is in the Quick Links, upper right, under the covers.

>10 SandraArdnas: This problem is why I limited the request to the simple case, and suggested (implied) that any halfway relevant date is better than nothing.

What I imagine for the complex case is an ability to see dates on the Editions page. Real editions usually have dates. If there's an ISBN or LC number, there should be only one publication date associated with it. Of course there may be errors in those dates, including the presence of multiple dates for the same book/edition. But it would be better than nothing.

With regard to errors users may notice - make it possible to report/correct them, as usual with uncertain data.

If there's only one date across all editions, it could (should) bubble up to the main page as well. Possibly if there are several dates, they could all bubble up ("editions published 1997,2001,2002,2003, 2010, 2022").

Obviously this is no use for books/editions that predate LC and SBN, and/or were entered manually without any publication data.

But some of them will have been entered with publication data, which could still be bubbled up just like data found for books entered via the various catalog sources on the add books page.

And plenty of others will have dates on their common knowledge page, which could also be bubbled up to the main page for the work.

It's nothing like a perfect solution. But AFAICT, nothing in LibraryThing is perfect - there are always errors and missing information, as well as ongoing attempts to correct them. Can we find a way to avoid having the perfect be the enemy of the good enough?

>13 MarthaJeanne: Doh! Now that I actually look, it's right there under my nose. Thank you.

>14 ArlieS: I still don't understand where do you expect to see that data in green? The fact that I don't isn't particularly relevant, if those who are supposed to implement it don't, it is. So my suggestion is to clarify your RSI, where exactly and what data would you like to see.

Personally, if it's CK (which normally doesn't have such green text feature, but is the only place with OPD) I'd be dead set against it since it would lead to people confirming it is as legitimate data without checking all too often.

I'd also reemphasize that without users entering CK there would be none. So, instead of arguing about perfect vs good enough, I'd much rather focus on the fact that members should enter that data. That is good enough practice that has benefited the site for all these years. If you have a reasonable RSI, by all means go forward, but that does not stop you from entering OPDs. Verified, correct OPDs are obtained only that way.

>16 SandraArdnas: Err ... I'm afraid many of your TLAs are not in my lookup table, or in one case are in that table with an obviously irrelevant meaning.

The only one I think I recognize is CK, which probably stands for "Common Knowledge", and refers to a section of info LT presents about a work

I should perhaps not have mentioned the "green data" thing; that came from someone else's attempt to translate my suggestion into LT terminology, and I didn't fully understand it. I think it means something like "data LT flags as untrustworthy because of its source". I don't know which sources are flagged unreliable, and which are treated as reliable, but I think site users are in the "unreliable" category.

Talking of reliability - the publication info *sometimes* gives the copyright, sometimes gives a publication year for the edition, sometimes gives 2 years - one of them labelled with some abbreviation for copyright. I get the impression it contains whatever the data source (=? publisher) wants, with no actual agreed-on meaning. It took me a _long_ time to determine that consensus on LT is that this field is NOT supposed to contain the work's publication year/copyright, but the year the specific book was printed/published.

Given the inconsistent info we get from publishers etc., I'm not sure we can display anything 100% reliably. Even the title is routinely modified by some of our sources to include e.g. a series name, that does not show as part of the title or sub-title on the physical book.

We somehow muddle through, most of the time, with work titles that aren't too misleading, even if the titles of the books associated with the work aren't all the same. (No, I'm not talking about works translated into other languages, or issued under different titles in the UK and the US.)

We don't seem to have problems making it easy for a user to see all the titles we have for a given work, at least in the case of translation. Scroll to the bottom of the work details page, e.g. https://www.librarything.com/work/5674/workdetails/189410313

Though in fact if you look at the Editions page, you'll find that is NOT _all_ the titles: https://www.librarything.com/work/5674/editions/189410313. Even in English, this example is sometimes titled "Alanna: The First Adventure", and sometimes "Alanna: The First Adventure (Song of the Lioness)", plus many other variations on both themes.

I don't know what we did to decide on a canonical english title for https://www.librarything.com/work/5674. I also don't know why we can't do the same thing to produce potential publication dates (presumably in some cases we have several for the same ISBN), or pick canonical publication dates (per ISBN, when there is an ISBN). That doesn't get us the original publication *or* the copyright date, except that in most cases, the earliest date will be the approximate copyright date.

Sure, it's fuzzy info, based on heuristics and best guesses. But it's only a bit more fuzzy than the title.

There's also the most common case, where there's only one title and only one publication date and one ISBN, Or perhaps even more common, one title, 2 or 3 ISBNS (hardcover, paperback, ebook) with publication dates within a year or two of each other.

If it's the simple case, display what is most likely correct. If it's not, maybe display a link to a pop-up with details, or information about why we can't give a reliable date.

p.s. I'm aware of the first-20-character heuristic for title matching. Obviously we can't use that _specific_ heuristic for publication years, though it (plus case insensitivity, plus ignoring punctuation) might have been enough to cause the system to equate all the english titles I saw for that example work.

OPD = Original Publication Date
RSI = Recommended Site Improvement

Green data is data that doesn’t come from the particular book you’re looking at but is estimated from other books in the work. It’s not necessarily any less reliable.

You can already fill in OPD with the original publication date if you've determined what that is through your own research which is most of what you say you want

>17 ArlieS: Even in English, this example is sometimes titled "Alanna: The First Adventure", and sometimes "Alanna: The First Adventure (Song of the Lioness)"

The latter is chiefly from Amazon data, and one of the great failings of that source. No conscientious publisher, bibliographer, or librarian includes a parenthetical series name like that. It's an abomination.

>19 jjwilson61: My problem is that for a random book I don't own - in particular, anything recommended to me that isn't obviously uninteresting - I'm likely to want to know the OPD to determine whether it's something I'd want to read. Much of the non-fiction I read is in fast moving fields; a book 20 years out of date is only marginally more useful than a book 200 or even 2000 years out of date. In the case of certain genres of fiction, on the other hand, my sweet spot is 20 or 30 years ago; recent works are likely to have features that will probably annoy me.

Ideally, LibraryThing would have everything I wanted to explore potential reading, except for the books themselves. I'd be able to easily get from recommendations to OPD, as well as publisher blurbs and any reviews written by other LT users.

Once I have the book in hand, with a probably-accurate copyright date on the reverse of its title page, I've already decided to acquire the book - or I'm at least in a library or brick-and-mortar bookstore considering it. At that point all I can do is possibly help others by providing the information I wanted to have available much earlier.

What really frustrates me is that LT has the information I want. Most of the records for individual books include publication data. But I can only see book records for books I include in my LT catalogue. I can't get LT to show me dates, let alone publishers, for each of the editions on the editions page of the work.

Actually, I'm oversimplifying. There's a 1 to many relationship between works and books, aka editions. Then there's a 1 to many relationship between editions and individual LT users' records of _their_ book. The publication data for a book in my LT catalogue has usually been updated to add copyright date; someone else with the same book probably won't have done that. The title for a book I own may have been modified to NOT have its first 20 characters identical to the title associated with some other WORK, to work around the 20 character heuristic. I may well have updated other fields as well, such as indicating that the book is hardcover or paperback. (I also always add the translator for books read in translation, if that name isn't already present; I'm surprised how often that information is missing from the starter data I get from Overcat or similar sources.)

In a perfect world (that still organized data this way) I could look at other people's "my book" records, and also get summary data about them.

Unfortunately, I can't even look at their book records. (If I could, I could in principle write a web scraper to gather summary data for me - though such a thing would have to be used _very_ sparingly lest it create excessive load on the LT site.)

>18 AndreasJ: This suggests that what I want could be phrased as:

LT should guesstimate the OPD from the publication data for all the books in the work. (This would be "green data".) It should then be displayed somewhere that users can easily see without having the work in their LT collection - such as on the Main page for the work.

>9 norabelle414: I missed this the first time round. If LT already has this, then it's just a matter of displaying it, and I can stop trying to explain how such an estimate could be made. (Tim's algorithm is sure to be much better than any heuristics I could design.)

In that case, my proposal is "display this estimate on the work's main page". Or possibly, "display the common knowledge OPD, if there is one, and otherwise display this estimate. Also, since this is complicated, explain how this works on a help page easily found when a user clicks 'help' while on the work's main page. Bonus - also explain how to give feedback when you think the estimate is incorrect."

Closing the loop on this - publication year (whether estimated or CK) is now shown on the author page

And it allows for sorting by publication year as well! That's great.

Thank you