User talk:MartyD/ProposedDateHelp

< User talk:MartyD
Revision as of 13:26, 5 February 2022 by MartyD (talk | contribs) (Draft 3 available)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

Use this page to discuss the proposed revision of the date help text. Use one section per specific topic. Edit the topic's section to comment/respond. Start a new section to discuss a point not yet under discussion.

Draft 1

The original proposed text.

No Publication Date

We have one more special date:

  • 9999-00-00 - when the book is rescheduled for the future but does not have an exact date yet.

While rarely used, it is an option when a book is added early, looks stable and gets all weird just before publication. Annie 13:17, 18 January 2022 (EST)

Oops. I knew that. Good catch, thanks. --MartyD 08:39, 21 January 2022 (EST)
I put all of the future date stuff into a section of its own and also added 9999-00-00 into the special dates list. See if that looks better. --MartyD 09:58, 21 January 2022 (EST)
Let’s sync that future language with the ROA - 90 days, not 2 months for the future eligibility (and using the date per announcement). :)
I really do not like the last sentence in the future section. It implies that anything after 90 days still can be added as 9999-00-00. That’s not the case per the ROA. 9999 is really just for books that were added with proper dates but got updated later. I know that the very first line of the future section does make sure we are covered but the less ambiguity and more clarity, the better. Let me think on wording. Annie 11:56, 21 January 2022 (EST)
Easy enough to address that with a "do not use 9999-00-00". And will also fix 2 months -> 3 months in update. --MartyD 18:04, 21 January 2022 (EST)
Let's use 90 days? :) 3 months are 90 days only if it is January, February and March and it is not a leap year. Otherwise, it is 91 or 92 or 89... :) I know it is nit-picky but... I already had someone complaining that almost 4 months is pretty much the same as 90 days and using the same language everywhere helps. Of course that one was smack in the middle of the Plague Year when even 30 days ahead of time was too risky - but that is a different story. :) Annie 18:08, 21 January 2022 (EST)

Missing sources

Where will Goodreads and other bibliographic projects (FantLab for Russian (and Russian adjacent), for Czech and so on) fall into this list? The English language ones (bar Goodreads) are in our verifications list but there is nothing there for the non-English ones (yes, some may cover some books but...)

Same question for library sources which are not OCLC - due to their policies, some library sources may have better information than OCLC - DNB often has a month for the German books, some others may record even dates - none of that gets into OCLC even when OCLC links the record. Annie 13:27, 18 January 2022 (EST)

I would also add publisher social media accounts as a source, that in some cases can be more accurate than catalogues or websites. The former suffer from long lead times and so might not reflect reschedulings (e.g. the current Tor US catalogue looks to have been produced in May 2021 going off the filename); the latter may not be properly maintained (e.g. Penguin UK). In theory at least, publisher accounts on Twitter (and presumably FB, etc) should be pretty up-to-date; I usually check the UK ones every Thursday/Friday as they tweet anything they've published that day.  ErsatzCulture 14:27, 18 January 2022 (EST)
Yeah - media accounts and other official accounts on mass sites - I was thinking if these fall under websites/catalogs or if we want to call them out separately. Annie 14:36, 18 January 2022 (EST)
My thinking is that this list should only be providing guidance as to which secondary information to prefer and which secondary information to treat with some degree of skepticism. Buried in help about how to fill out a date field is NOT the place to provide an authoritative list of all possible secondary sources. Ideally, this section could refer to such a list, located elsewhere. But perhaps it's already overly specific and should be condensed. What do you think of something like this instead?
  1. The publication itself; this includes direct statements as well as dates encoded in number/date lines.
  2. Parties directly involved with the publication (publisher, author, artist, etc.)
  3. Later printings/editions.
  4. The ISFDB bibliographic Verification Sources (list and details)
  5. Other fact-oriented parties close to the publication (new book seller, review/announcement, bibliographer/database)
  6. Calculation based on codes, known announcement + publication timings, etc. (e.g., for book clubs)
  7. Sources with non-fact focus or removed from the publication (used book seller, interview, etc.)
Perhaps with a warning about relying on anything in Wikipedia. :-) It's still long. Maybe even better would be to list #1 - #4 and then just state that other sources may be used, but information from those four take precedent. And encourage editors to seek corroboration for more less reliable sources such as used book sellers and interviews. --MartyD 08:34, 21 January 2022 (EST)
That's better I think. I agree - we should not get too specific. A few notes:
  • "close to the publisher" and "removed from the publisher" can be... subjective. So how about:
    • Other fact-oriented sources: databases, bibliographies, library catalogs, contemporary reviews, announcements, listings in new book sellers and so on.
    • Other sites: used book sellers, non-contemporary interviews and reviews and so on
That still leaves some ambiguity but at least it is clear what we are trying to achieve.
  • I also think we need a "Sources from the country of origin of the book are usually more accurate than sources from other countries. However, that may vary for some specific languages, countries and situations." or words to that effect - for example a US based site is very rarely a better source than a Czech DB for a book published in Czechia; UK sites will usually be better for UK books (compared to the American sites) and so on. There are exceptions (sometimes due to common/similar languages, sometimes due to politics and so on) thus the last sentence - but that is generally always a good idea - "look for local source if one is available".
The only thing that still bugs me is that for non-English language books, these fact oriented sources may actually be better than our verification sources even if we put them lower on the list. So maybe the "country of origin" note should tie to the order note instead of being on its own.
Just listing 1-4 is perfect for US books and not so perfect for anything else - our verification sources are US ones for the most part. :) So we do need something after that to ensure that we don't discard valid information. Annie 12:31, 21 January 2022 (EST)
I would add a sentence encouraging editors to document any discrepancies in secondary sources in the Note field. Another thing that we may want to mention is that specialized bibliographies are, for the most part, more reliable than general purpose bibliographies. Ahasuerus 16:53, 22 January 2022 (EST)

Handling pubs with different dates for different territories

It would be helpful to have explicit notes [*] on how to document pubs that have more than one date - typically for "transatlantic" publishers that put out the same editions in both North America and UK/Commonwealth territories. Examples: Titan Books, Severn House, Solaris/Rebellion/Abaddon. (I don't think I have any physical editions from any of these publishers - I am curious if/how they address the "Published in <month> <year>" info on the copyright page in cases like the ones I've linked to, where the pub dates are in different months?)

When I've seen such cases, I've generally entered the earliest date in the pub date field, and added the later one in the pub note (same for the price) - but I've gotten the impression that perhaps the more rules-compliant way would be to use the date associated with the country the publisher is based in, although I believe this isn't as clear-cut.

[ * As such cases are fairly rare, I'd be more than happy for this to be on a separate help page, and just have a link from the main pub date help ] ErsatzCulture 14:39, 18 January 2022 (EST)

Yeah, I don't know about that. In the only cases I specifically recall, we made two entries, one for each country (using that country's release date and that country's price). Of course, if everything's identical, there's no practical way to determine if a book in hand is one edition or the other. Earliest date seems most practical if not separating. --MartyD 08:38, 21 January 2022 (EST)
That whole thing had always been a bit of a mess in the DB though. For some we add 2 publications which are identical (thus making all statistics go weird) except for the currency and date - but they describe the same book... For some we add one book with notes explaining the duality (I do that when there is such discrepancy for books coming from Fixer). First date wins is usually my solution to those - anything else goes into the notes (and we hope that a future PV won't wipe out all the collected data because these notes are important for tracing editions sometimes... As long as things are documented, I think we are covered. Annie 12:45, 21 January 2022 (EST)
Sorry, I realize I missed this. When I get a chance I'll add a second bullet to the "Dating Books" section. Something along the lines of: If there is a single statement about simultaneous publication in multiple locales but with differing dates, use the earliest date if this record is for all of the locales, and use the locale-specific date if this record is for just one of the locales. --MartyD 15:50, 25 January 2022 (EST)
I have added this. It also occurred to me to add the companion case, explicitly telling people to use the latest date if a whole list is presented. I didn't want there to be any confusion between the simultaneous printing case and the historical publication list case. That led me to rework the whole book identification/dating to move all of the detail about reprints into a section of its own. See what you think. --MartyD 13:14, 5 February 2022 (EST)

Text shortening

I'm all for the longer text, as it is quite clear. ../Doug H 15:59, 18 January 2022 (EST)

Yeah - I think that the text is not really too long - dates are kinda important and making it as clear as humanly possible is a feature and not a bug :) —The preceding unsigned comment was added by Anniemod (talkcontribs) .
Be careful what you wish for. I've made it even longer. :-) --MartyD 13:11, 5 February 2022 (EST)

Dating Books

One small thing here bugs me a bit. "The publication date stated in the book should be used unless known to be/demonstrably for another printing." is technically correct but can be read as absolute statement by editors, discarding the next bullet point (that's what started this whole discussion again) and despite the earlier "using secondary sources to supply missing dates and date details.". So how about:

"The publication date stated in the book should be used when available (precised from secondary sources if incomplete) unless known to be/demonstrably for another printing."

Or something like that. And maybe flatly "Complete dates are preferred when available" thus ensuring that noone argues that 2020-02-00 is better than 2020-02-13 when we can find that "13"... Annie 12:39, 21 January 2022 (EST)

I disagree with filling in the date from any secondary source. It's loading two different values into a single field. I find the idea that a Primary Verifier with the book in hand may be unable to enter a publication without checking secondary source(s) concerning. Also, said hopeful PV may have to look at ALL existing publications to find a match because the date in the summary COULD be listed differently than what they see in their copy (research or "known to be for another printing" could theoretically cause a printed date to be entered as 0000-00-00 or no printed date into a date). I believe the Notes are intended to document additional observations and the results of research. I do believe that the reference to the number line to determine if the date or 0000-00-00 is quite valid, as the information is part of the publication. A 'researched' date is something else, and is somewhat meaningless in the absence of the associated source. Augmenting the date by adding day or even month is a gray area, but I view it as 'researched' and would rather it not be done. ../Doug H 15:43, 21 January 2022 (EST)
Two quick things about this:
All that I am trying to accomplish here is document what was agreed to in the 2012 R&S discussion, which was: Use the book's statement as the basis of the captured date if it applies to the edition in hand, making the date more precise from other source is acceptable (with documentation), and discrepancies between what the book says and other sources should be recorded in the notes. Whether that should be the policy, and what to do about multiple dates, is beyond the scope of this.
I think there's also a misunderstanding, and I'll see about making the wording clearer. Nothing is meant to require the PVer to do any research of secondary sources or to confirm anything from a secondary source. The PVer's only responsibility is to make sure anything in the record supposedly coming from the book matches what is present in the verification copy. All of that text about everything not from the book itself is merely what's permitted to be used and how to use it. When properly documented, there should be no impact on the PV at all, as it will be clear what came from another source. Does that make sense, or am I misinterpreting what you meant? --MartyD 17:36, 21 January 2022 (EST)
So if a publisher only ever prints the year (and never prints a month or a date) as a matter of policy, we should not use month and day on their books even when it is readily available? And any undated book should be 0000-00-00, even if we can find a date for it? That's... not very useful and makes our DB less useful for research purposes. And it makes hell of any chronological ordering of publications in a title record for example. Plus that's not what the policy had been since I had joined the site (which is what we are trying to document here as Marty mentioned... We really need two fields - "stated date" and "actual date" but as we don't have them, we need to adapt. So let's get the language to what the CURRENT policy is and then go from there. Marty is right - if we want to change the policy, we need a R&S discussion and general agreement - which may be hard to achieve. Especially if there is a way to get the 2 separate fields soon-ish :) Annie 18:17, 21 January 2022 (EST)
Dating Books - The ISFDB documents a book's publication date as well as it can be determined.
  • The publication date stated in the book should be used unless known to be/demonstrably for another printing. Try to find a statement (often on the verso of the title page) that says something like "Published in June 2001"; the copyright date is often misleading, since works can be reprinted. Look out for signs that this is a reprint; indications often include a series of numbers (e.g. "3 4 5 6 7 8 9") at the bottom of the verso of the title page; this particular string indicates that this is a third printing. If you know you are holding a reprint, and there is no way to date that particular publication, enter the year field as 0000-00-00. Note that we are interested in recording each different reprint of a publication, since there can be some significant differences between them, such as cover art, or price.
  • If a more specific publication date consistent with the stated date is available, that may be used. The source and difference should be provided in the Note.
  • If the book's statement of publication date is missing or for another printing, a date from another source may be used. The source and publication's stated date (if any) should be provided in the Note.
  • If date is known to be incorrect based on an authoritative source, it may be replaced. The source and publication's stated date (if any) should be provided in the Note.
Here's something a little more accurate then (no, I don't intend to try changing the rule). It begs two questions - what's an authoritative source and how do you document the source? My concern on the latter is something like "Amazon" or OCLC which may have multiple conflicting entries. It needs to be specific enough for someone else to track it down to ensure it is a match for the specific publication. ../Doug H 22:25, 21 January 2022 (EST)
The more I think about it, the more convinced I become that we really need to add a "Stated Publication Date" to publication records and we need to do it sooner rather than later. A single field is just insufficient to handle the reality of what's out there. I plan to propose it on the Community Portal in the next day or two. Ahasuerus 16:58, 22 January 2022 (EST)
I have posted a proposal on the Community Portal. Ahasuerus 16:01, 23 January 2022 (EST)

Draft 2

I have made a major revision that tries to incorporate all of the above feedback. Please comment below, especially if I overlooked anything or made something worse. I mostly want to get all of the basic stuff covered and accurate before asking the broader community for comments, just so that there's less shifting and distraction by having to do major changes. --MartyD 15:40, 23 January 2022 (EST)


One thing you'll notice I tried to do is put a basic summary at the beginning. I hope that appeals to you all. --MartyD 15:40, 23 January 2022 (EST)

See below

Can this be a bit more specific, naming (or linking) to the section? ../Doug H 15:53, 23 January 2022 (EST)

Yes. I was wishing the help used headers instead of bullets, as then we could embed clickable links, but names are there now, so we can use those. When I get a chance, I will fix up the references. --MartyD 15:43, 25 January 2022 (EST)
I figured out how to do this and made those changes. See what you think (I am still making other edits, but I believe I got all of the cross-references cleaned up). --MartyD 10:52, 5 February 2022 (EST)

Draft 3

A revised draft is posted. Changes from Draft 2 are:

  1. I incorporated the comments about intra-text links/labels (vs. "above" and "below")
  2. I added how to handle multiple dates in a simultaneous publication statement.
  3. #2 led me to add explicit instructions for handling a list of publication dates.
  4. The combination of #2 and #3 led me to reorganize what is now Dating Books. I made it more bullet-oriented, tried to keep what-to-do at the top, and moved the detail about reprints down into a section of its own.

I will draw the general public's attention to this draft and ask that comments be added below. --MartyD 13:26, 5 February 2022 (EST)

Book Reprints

As I mentioned above, I pulled the book reprint information into a new Book Reprints sub-section. I added some more date-oriented content there that ought to be reviewed. I appreciate any comments about it one way or the other. --MartyD 13:26, 5 February 2022 (EST)