Rules and standards discussions


Jump to: navigation, search

ISFDB Noticeboards
Before posting to this page, consider whether one of the other noticeboards might suit your needs better.
Help desk
Questions about doing a specific task, or how to correct information when the solution is not immediately obvious.
• New post • Archives
Verification requests
Help with bibliographic, image credit, and other questions which require a physical check of the work in question.
• New post • Archives
Rules and standards
Discussions about the rules and standards, as well as questions about interpretation and application of those rules.
• New post • Rules changelog • Archives
Community Portal
General discussion about anything not covered by the more specialized noticeboards to the left.
• New post • Archives
Moderator noticeboard
Get the attention of moderators regarding submission questions.
• New post • Archives • Cancel submission
Roadmap: For the original discussion of Roadmap 2017 see this archived section. For the current implementation status, see What's New#Roadmap 2017.

Archive Quick Links
Archives of old Rules and standards discussions.

1 · 2 · 3 · 4 · 5 · 6 · 7 · 8 · 9 · 10 · 11 · 12 · 13 · 14 · 15 · 16

Expanded archive listing
Rules and standards changelog

Every rule change than comes out of a discussion here should be added to the Rules and standards changelog.


Question 1: "uncredited" versus "unknown" for old manuscripts

Is there a strong preference, one way or the other, as to whether authors whose names have been lost to time, because the copyists of medieval manuscripts did not see fit to record them, should be in the database as "uncredited" or "unknown"? Most of these old works have their canonical author as "uncredited" but Sir Orfeo, for one, is "unknown." --Vasha (cazadora de tildes) 21:08, 7 January 2019 (EST)

Uncredited should be used if the original work was not credited. Unknown is for use when you are using a secondary source that doesn't specify the author, but doesn't necessarily mean the author is uncredited. -- JLaTondre (talk) 21:25, 7 January 2019 (EST)
If no one disagrees with this, I will go through and find all the "unknowns" and correct them to "uncredited." --Vasha (cazadora de tildes) 21:47, 7 January 2019 (EST)
My statement is per the help (see Template:TitleFields:Author, bullet 1). However, your "I will go through and find all the "unknowns" and correct them to "uncredited."" is ambiguous. I hope you mean only ones that match the case you mentioned above. Because there are valid "unknowns" in the database. -- JLaTondre (talk) 22:46, 7 January 2019 (EST)
Well yeah, I meant "Sir Orfeo" and any others that may have the same mistake. --Vasha (cazadora de tildes) 23:01, 7 January 2019 (EST)
Update: I spoke too soon—there are evidently a LOT of things that need correcting on the unknown summary page, and the task of figuring out which ones to correct is beyond me. However, at least now I know what should be done; and I did find and change a few obvious ones. --Vasha (cazadora de tildes) 23:15, 7 January 2019 (EST)
The bulk of the records are house names where the true writer is not known. That is a valid use. I fixed a couple of obvious errors. I'm sure there are more, but not convinced it's a lot of things. Most of the non-house names that I spot checked were valid uses. I do have it on my list to cleanup some of the 'Noname' house name ones as any of the Frank Reade Jr.'s are by Luis Senarens per SFE3. While the Frank Reade's were done by multiple people, the Jr.'s were all done by Senarens. -- JLaTondre (talk) 18:51, 8 January 2019 (EST)
Probably those Arabian Nights records should be "uncredited" but I'd really like to look at the publications to see exactly how they're credited. There are a couple dozen more folktales and old stories ... Yes, the house names make up 85% of it, but the rest needs checking. --Vasha (cazadora de tildes) 19:10, 8 January 2019 (EST)

Question 2: Dates of literature from manuscripts

What should be the date, in this database, of literature whose first extant copy is a manuscript that can't be precisely dated? The majority of them currently have the date "unknown" (0000-00-00) but some, like Clerk Saunders, are listed according to a date when they were printed; and Sir Gawayn and þe Grene Knyȝt has been given the date 1400-00-00 because that is the approximate date of the manuscript.

For the record, I'm in favor of continuing the practice of the majority. It's not unreasonable to regard the manuscript as a publication. Discussion of the date of composition and dates of manuscripts can be a paragraph in the notes, and a date of "unknown" lets visitors know to look at the notes for more information. --Vasha (cazadora de tildes) 21:46, 7 January 2019 (EST)

I guess it depends on how we define "publication". For example, we think of books printed by book clubs as "published" even though technically only book club members have access to them. Another example: back in the day some books were "privately published" for a select group of subscribers, kind of an early version of the "limited edition" concept. We consider them "published" even though 99.999% of the public had no access to them.
Prior to the invention of the printing press, manuscripts were arguably "published" if numerous copies existed. However, where do we draw the line? What if only 3 copies of a particular manuscript were made? 2? 1? And how can we be sure that the number of currently known copies is not a fraction of the total number of copies made? Ahasuerus 22:00, 7 January 2019 (EST)
Indeed, I guess "Clerk Saunders" is not very relevant to this discussion because it doesn't predate the printing press; it first appeared in a couple of 18th-century manuscripts (private notebooks) that were definitely not intended for circulation. But the case of medieval manuscripts is different: even if few copies were made (or even one!), they were usually still intended to be part of a library. I think we should gloss over the complexities of production and use by considering manuscripts prior to 1500 or so to be publications ... --Vasha (cazadora de tildes) 22:18, 7 January 2019 (EST)

Revisiting conflicting policies about nongenre works

Last year, in the course of a discussion about a somewhat different matter, I pointed out that two of the ISFDB policies conflict with each other: 1. ISFDB:Policy says that we exclude "Works that are not related to speculative fiction by authors who have not published works either of or about speculative fiction over a certain threshold." 2. We include the complete contents of genre magazines; I can't find where this is written in policies, but during last year's discussion, everyone agreed that this is our practice. Obviously, these two rules are contradictory: do we include or exclude a non-speculative work by a non-genre author that appears in a genre magazine? --Vasha (cazadora de tildes) 15:09, 17 January 2019 (EST)

I'll need to do more digging, but there is a relevant paragraph in Help:Entering_non-genre_periodicals#Genre_special_issues -- see the bolded part:
  • Sometimes, a non-genre periodical will devote an entire issue to speculative fiction and/or articles about it. This can be regarded as a genre publication and genre non-fiction should be cataloged along with the fiction (even though we do not normally catalog non-fiction from non-genre magazines).
Ahasuerus 14:33, 18 January 2019 (EST)

In the discussion I'm citing, no one had any suggestions as to what to do about resolving this problem. But I've since had some thoughts that might help add clarity.

An appearance in a genre magazine is not the only reason a non-genre story might legitimately be in the database. As you all probably know, some of our standard secondary sources, such as the Supernatural Index, included anthologies that contained a mixture of speculative and non-speculative stories, and indexed their entire contents without indicating which of the stories weren't actually speculative. Other writers then copied from them when creating lists of speculative fiction, not realizing that they were including non-genre works in the list. I see this as a good reason to include in the DB the complete contents of any anthology indexed by the Supernatural Index et al. (with NG items marked as such): we will be setting the record straight, and it will be useful for someone to be able to look up everything in one of these "speculative" lists and find it here. "Huh, why is this on the list? Oh, because of being in such-and-such anthology."

That's two potential reasons to make an exception to the principle "exclude ng works by ng authors." Perhaps we need to rewrite that policy; what do you think? It could be an "include only if" rule in three parts: "Include nongenre works only if 1. They are by an author above a certain threshold; or 2. They have been published in a genre magazine; or 3. They are listed in one of the standard secondary sources." (Or, whatever reasons for including ng works we agree on; the main point is that we can rewrite the policy so that "author above threshold" is no longer stated to be the only possible reason for inclusion of an ng work.) --Vasha (cazadora de tildes) 15:09, 17 January 2019 (EST)

The last time the issue came up, I made a comment that may be relevant:
(reformatted) One of the problems that we have run into when discussing what types of non-fiction to include is the sheer number of different ways to categorize them. Off the top of my head:
  • Length: books vs. short works. Short works are further subdivided into:
    • Essays
    • Reviews
    • Interviews
  • Type of publication where the non-fiction work appeared:
    • Genre periodicals
    • Non-genre periodicals
    • Genre books
    • Non-genre books
  • Type of non-fiction work:
    • Works about written SF
    • Works about non-written SF like comics, films, TV, etc
    • Works about non-genre issues
    • Works that are "sort of" related to SF, e.g. space exploration, history of myths, etc
This segmentation makes it difficult to ensure that everyone is on the same page when discussing proposals that cover only certain "cells" of this multidimensional matrix. Perhaps a graphical (tabular?) representation may help clarify what our policy currently says and where we may want to go from here. Ahasuerus 16:44, 16 December 2018 (EST)
Getting back to the issue of non-genre fiction, I agree that we need to -- at the very least -- clarify Help. Template:PublicationInfo:WhatToInclude, which is transcluded in ISFDB:Policy and Help:Screen:NewPub#Content_Information, was originally written based on the assumption that it would be used when entering genre publications. A significant part of the template primarily applies to genre magazines. We should make these assumptions explicit and also add a bidirectional link to Help:Entering_non-genre_periodicals in order to eliminate/reduce ambiguity. Ahasuerus 15:32, 18 January 2019 (EST)
I think that Vasha is talking about non-genre FICTION here though. While we need to resolve both, I think the bigger issue are non-genre stories and not non-fiction. Or I am badly misreading things :) Annie 15:56, 18 January 2019 (EST)
Rereading my comment above, I see that it wasn't very clear. It consisted of two parts. The first part was a repost of my December 2018 comment about non-fiction. The second part, which starts with "Getting back to the issue of non-genre fiction", was about non-genre fiction. Ahasuerus 16:18, 18 January 2019 (EST)
Now that you mentioned it - now i see that it is two parts - the second one looked almost as an afterthought so I think I got stuck on the first one. :) Annie 16:24, 18 January 2019 (EST)
I wasn't talking about non-fiction at all; I meant FICTION that is non-genre and has no speculative elements, and I guess I should have made that more clear in the title of the post. The way the Rules of Acquisition are currently worded, they lump fiction and non-fiction together. Both of them are referred to by the paragraph in "Exclusions" that says "Works that are not related to speculative fiction ..." Maybe this whole section of the Rules of Acquisition needs to be rewritten to discuss fiction and nonfiction separately? We do seem to have different standards for the two. --Vasha (cazadora de tildes) 16:46, 18 January 2019 (EST)
It looks like there are two issues here:
  1. Cleaning up the Rules of Acquisition and Help to bring them in line with our current practices (while making sure that everyone is on the same page re: what the current practices are)
  2. Using "the standard secondary sources" as the guide when deciding whether to include non-genre works
Naturally, I am all in favor of #1, but I have my doubts about #2. If nothing else, different bibliographers can use different inclusion criteria when compiling different bibliographies. There will be times when their criteria will be similar to other bibliographers' (including ours) but there will also be times when they will be different.
To take a step back and look at the big picture, the main reason why the issue of what's included and what's excluded is so complicated is that we use two different types of eligibility criteria. On the one hand, we have a list of criteria which determine where we draw the line between "speculative fiction" and other types of texts (or "titles" as we call them.) It can be tricky, but ultimately we can make it mostly consistent. On the other hand, we have another list of eligibility criteria based on publication data. The basic idea was that we wanted to:
Once we added publication-based criteria to the mix, things became complicated.
I think it would be best to clarify and clean up Policy and Help first. We made some progress last year when many "Debatable" lines were sorted out and I hope that we can make more progress in the coming months. Clarifying what counts as a "genre publication" and updating Policy/Help with guidelines re: handling genre and non-genre publications may be a good first step. Ahasuerus 11:25, 19 January 2019 (EST)
To the idea of cleaning up the Policy and Help, a thousand times yes. As you say, the publication-based criteria add considerable complexity. And that is presumably why they are not currently discussed in the RoI even though people are using them. We should have a comprehensive discussion about what the practices are.
It's interesting that you cite "Major Pugachov's Last Battle" as an example: that one's not even in a genre magazine, but in genre anthologies! During last year's discussion, more or less 100% of contributors supported indexing the complete contents of genre magazines, but were uncertain about anthologies. And I can see why: deciding what's a genre anthology is a problem that would puzzle the planet's best minds; magazines are a considerably more defined and constrained set (though naturally there are uncertain cases). At any rate, any revision of the RoI will have to mention this principle about the contents of magazines. --Vasha (cazadora de tildes) 13:02, 19 January 2019 (EST)
Let me quote Help:Entering non-genre periodicals to make sure that we are on the same page re: what Help currently says:
  • Anthologies that contain mostly general fiction, but also have some SF content, can also be entered using the data entry rules for non-genre periodicals.
The way it's phrased -- note the word can -- is somewhat ambiguous. Based on the principle "one headache at a time", I decided not to open this can of worms and leave the sentence "as is" when I was updating the page last year.
My current thinking is that it would be a good starting point for a discussion. If we can define "genre publications" and "non-genre publications", it would establish a solid foundation for future Policy/Help updates. Help:Entering non-genre periodicals already mentions the concept of "genre publications" under Help:Entering_non-genre_periodicals#Genre_special_issues, but it needs to be fleshed out. Ahasuerus 19:43, 19 January 2019 (EST)
That is a can of worms, all right. Note that "using the data entry rules for non-genre periodicals" might imply that we entered the editors as "Editors of..." and some people had formerly done so until a discussion last year (I wish I could find it) made clear that a solid majority were opposed to it. So that's one way that practices for anthologies differ from magazines. But, I guess I will move this discussion into a separate section to see what people think now. --Vasha (cazadora de tildes) 21:19, 19 January 2019 (EST)
One reason the Policy page is currently confusing is that it is not made clear exactly which works the "What to Include" section is referring to. Its description just says "This section will help editors to determine whether a contained work should be entered as a separate content record in a ISFDB publication record." But then it goes on to say, "All forms of fiction are always included," which is in contradiction to the sentence in the Rules of Acquisition one paragraph above, to exclude "Works unrelated to speculative fiction that are found in primarily non-genre publication [sic] that will be cataloged based on other criteria, e.g. ... a Playboy issue that include a single speculative story ..." So tacitly, the "What to Include" is not dealing with these "primarily non-genre publications". We need to make that explicit, and be more clear about what "primarily non-genre publications" are. --Vasha (cazadora de tildes) 13:02, 19 January 2019 (EST)

Anthologies with little spec fic: interpreting what "Help" says

(copied from above) Help currently says:

Anthologies that contain mostly general fiction, but also have some SF content, can also be entered using the data entry rules for non-genre periodicals.

The way it's phrased -- note the word can -- is somewhat ambiguous. Based on the principle "one headache at a time", I decided not to open this can of worms and leave the sentence "as is" when I was updating the page last year.

My current thinking is that it would be a good starting point for a discussion. If we can define "genre publications" and "non-genre publications", it would establish a solid foundation for future Policy/Help updates. Help:Entering non-genre periodicals already mentions the concept of "genre publications" under Help:Entering_non-genre_periodicals#Genre_special_issues, but it needs to be fleshed out. Ahasuerus 19:43, 19 January 2019 (EST)

That is a can of worms, all right. Note that "using the data entry rules for non-genre periodicals" might imply that we entered the editors as "Editors of..." and some people had formerly done so until a discussion last year (I wish I could find it) made clear that a solid majority were opposed to it. So that's one way that practices for anthologies differ from magazines.
The way it would look if we treated a not-primarily-genre anthology exactly like a magazine issue would be: it would be marked NG, we would only catalog speculative fiction contents even if there are speculative-related essays, we would not create art records for it unless the art is directly related to a piece of speculative fiction in it, and the editors would be listed as "Editors of ..." Does all that make sense? --Vasha (cazadora de tildes) 21:19, 19 January 2019 (EST)
We need to be very careful around that specific point - 10 years ago magazines and anthologies were easy to separate. These days? Half of what is being published by the small presses can be categorized either way - and sometimes what it looks like changes a few issues into a run. Just saying :) Annie 14:48, 20 January 2019 (EST)
Heck yeah, the line is blurred in all sorts of ways—a lot of annual journals are in the DB as anthologies, for example, and Foreshadow, which calls itself a "serial anthology" and plans to publish once a month for exactly one year, is in the DB as a webzine. It would certainly make life much easier if we applied exactly the same standards to the two, so that we wouldn't have to rethink anything if we changed the type from one to another. Or at least, if preferred, the same standards except for the "editors of" thing; that at least would be easy to convert from name to "Editors" or vice versa if changing the type. --Vasha (cazadora de tildes) 14:57, 20 January 2019 (EST)
I agree that standardizing the data entry rules for non-genre anthologies and non-genre magazine -- to the extent possible/feasible -- would be a step in the right direction. Currently we handle editor attribution differently. Anything else that we do differently for non-genre anthologies vs. non-genre magazines? (I am under the weather at the moment, so I may be missing something obvious.) Ahasuerus 18:42, 20 January 2019 (EST)

(unindent) Cover art. The non-genre periodical help page states that cover art should not be entered (credit or image) unless it illustrates genre contents. There is not a similar restriction for non-genre anthologies and common practice is to include it. -- JLaTondre (talk) 19:55, 20 January 2019 (EST)

Oh, right. Thanks! Ahasuerus 20:34, 20 January 2019 (EST)
Huh, I wasn't aware of that common practice, and I've been omitting cover art from NG anthologies. It's not much of a loss, is it? --Vasha (cazadora de tildes) 20:43, 20 January 2019 (EST)
Usually not, though I included cover art for those images that include speculative content or are by notable spec. artists. It is complicated. Stonecreek 03:37, 21 January 2019
Interesting... We have a guideline to not do magazine cover art for ng mags unless it illustrates a story, but a practice to usually include antho cover art. Those of you who make this distinction, how do you feel mags and anthos differ? What is it about ng magazines that makes you feel you should omit general speculative art, which you don't feel about ng anthologies? --Vasha (cazadora de tildes) 05:32, 21 January 2019 (EST)
This conversation left me with the impression it has to do to with the type of covers some magazines have (which is typically not an issue for anthologies). -- JLaTondre (talk) 08:01, 21 January 2019 (EST)
Not sure what you mean--are you referring to the comment about not wanting to host Playboy covers? There most certainly are erotic anthology covers, speculative as well as non. --Vasha (cazadora de tildes) 09:23, 21 January 2019 (EST)
I believe I have seen other similar comments. Whether it is the quantity difference or something else, you would have to ask those that espoused that opinion. But either way, it is something that should be thought through and dealt with (for both non-genre & genre, magazines & non-magazines) in a consistent way (don't allow, allow, allow with filter, other). -- JLaTondre (talk) 11:00, 21 January 2019 (EST)
Setting aside the question of whether to filter erotic images, I see no need to have cover images for nongenre magazines, and doubly no need for artist records. But there seems no reason to have them for nongenre anthologies either. Can we not unify the standards? --Vasha (cazadora de tildes) 13:16, 21 January 2019 (EST)

(unindent) And we still do need some standardization: when does an anthology earn the spec. fic. category? My rule of thumb was that more than half of the fiction items should be in that vein, but does that mean in number, in volume / page count, or in one of those? Stonecreek 03:37, 21 January 2019 (EST)

Either or both of those; in the doubtful indeterminate area, you can be influenced by what the editors state is the theme and emphasis of the work. --Vasha (cazadora de tildes) 06:27, 21 January 2019 (EST)
I find that in most cases the intent is clear and the clear majority of included stories is either SF or non-SF. The main exceptions are horror anthologies/collections which include both supernatural horror and psychological horror stories. Also, some "cross-genre" anthologies like Not to Be Taken at Night: Thirteen Classic Canadian Tales of Mystery and the Supernatural require additional digging to determine which stories are SF. The worst offender that I have come across recently is The Fatal Eggs and Other Soviet Satire: 1918-1963: the title novella is SF, but I have been unable to determine whether the rest are (some appear to be satirical fabulations.) Ahasuerus 14:52, 21 January 2019 (EST)
Yes! "mystery and the supernatural" anthologies are a headache; some editors who had their anthologies marketed in that category had a heavy emphasis on non-supernatural suspense, adventure, and melodrama, whereas others were mostly interested in speculative fiction with some non-supernatural weird fiction thrown in; and it can be really hard to know which is which without looking at the book itself. The famous example is Great Tales of Terror and the Supernatural (1944), the section labelled terror covers both spec and nonspec. I think that from about 1900 to the 1950s you had mixed spec/nonspec "mystery" anthologies and "terror" anthologies; I think the former term implies less supernatural fiction than the latter, but really the usage overlapped. --Vasha (cazadora de tildes) 15:30, 21 January 2019 (EST)
BTW, re: The Fatal Eggs and Other Soviet Satire; my local university library has it & I'll look at it tomorrow. --Vasha (cazadora de tildes) 16:22, 21 January 2019 (EST)
Thanks! Ahasuerus 16:36, 21 January 2019 (EST)

Help Cleanup: Proposed First Step

After reviewing various related Help pages, I think it may be best to start with Help:Entering non-genre periodicals. It's wordy, repetitive and duplicates many of the rules found in our main Help templates. I suggest that we start by eliminating duplication and compressing the rest to something more manageable. Ideally, we'll reduce it to a list of differences between the data entry rules for genre and non-genre publications. That should make it easier to understand what the currently documented differences are; then we can decide what to do about them. Ahasuerus 15:42, 21 January 2019 (EST)

How to reference pages preceding page 1?

(Moved discussion over from Helpdesk) Referencing pages

The current rules on page count and page referencing as explained here, here and especially here state that

  1. Total page count is calculated by adding all numbered and unnumbered page ranges, provided they have contents that need to be recorded. For example [10]+400+[15], where there are unnumbered pages before and after the body of the work. This is fairly straightforward.
  2. Referencing pages within the pre- or post unnumbered pages is done by means of the derived page number within that page range. This is not so straightforward and the rules are not easy to understand.

As far as I can tell the current rules are ambiguous in cases such as the [10]+400+[15] example above: if you refer to page [5], which one is it? Is it the 5th page of the unnumbered page range before or after the body text?

Solution applied by some editors (including me) for the 'after' bodytext case is to derive page numbering based on the last printed page number and then counting forward. My example would then be page [405] as referring to the 5th unnumbered page of the unnumbered page range following the bodytext.

This approach breaks down for an unnumbered page range before the bodytext of course (unless we go with negative page numbers!). A solution editors have come up with for the 'before' bodytext case is to use roman numerals for both page count as well as page number, and adding a clarifying note. For the above example this would translate to: total no of pages [x]+400+[15] for the total page count, and the contents on page 5 of the unnumbered page range before the body text would then be referred to as [v]. See Senlin Ascends for an illustrative example.

Is there anyone else that has struggled with this and has come up with a cleverer solution? Can we improve or strengthen the current rules somehow?

And we also need to come up with an explicit rule for counting the total number of unnumbered pages before page 1. Do we start counting from the very first page (not counting the cover), or do we start counting backwards from page 1 until the relevant piece(s) we want to record the page number(s) of? I guess the former is the right way to count, but the rules don't really confirm that. (And I've used the latter in my earliest edits...) MagicUnk 14:34, 2 February 2019 (EST)

To my novice mind, using [bracketed] roman numerals for the unnumbered pre-pages makes the most sense, and it seems to me offhand that that wouldn't cause any problems. The only problem that would leave was if there were unnumbered pages before numbered roman numerals before regular numerals. It doesn't specifically say not to use roman numerals for unnumbered pages. gzuckier 17:22, 2 February 2019 (EST)
The purpose of the content page numbers is a) to differentiate publications and b) give a rough idea of size. I don't think there would really be any confusion with the [5] in the example above. It's insufficient information to recreate the publication, but certainly enough to tell if the pub you hold matches the description. ../Doug H 20:05, 2 February 2019 (EST)
I agree with Doug and don't see a need to change the numbering system. We don't worry about this with multi volume publications which can have multiple instances of any page. In those instances, we use the pipe (|) operator to force the numbering to the right place in the contents list. So for first page of the unnumbered pages preceding the pagination we could enter it as "[1]|0.01", page 1 in volume 1 would be simply "1" and page 1 of volume 2 would be "1|2001". There really isn't a need to introduce bracketed roman numbers which would necessitate changing those publications entered under the current system. --Ron ~ RtraceTalk 22:21, 2 February 2019 (EST)
I do not agree with Doug that we should be satisfied with insufficient or incomplete information. I believe that when providing information we need to be as specific and unambiguous as we can.
Anyhow, as for Ron's remark referencing page numbers per the current rules is unambiguous only when page order (by using the piping symbol) AND page count information is both available. In other words it only works for contents listed in the contents section that have page numbers & ordering added to them. It still does not work if you want to refer to pages that are not recorded as contents but one still wants to capture - see Stonecreek's example here. Granted, we could spell out in the notes field something like 'on unnumbered page [5] preceding page 1' or some such when we want to refer to a specific unnumbered page. It would be strictly following the current rules but it's quite elaborate that way. Nevertheless, we could do it that way and it would work.
So, do we agree that
  1. We consequently continue to use roman numerals for page count for unnumbered page ranges before AND after the body text, in conjunction with contents page number and page order (as the current rules dictate), and
  2. When referring in the notes to unnumbered pages from page ranges before or after the body text that are not entered as contents record we explicitly mention which page range that particular page is from?
If we agree this is the way to continue it would mean that moderators need to be more attentive to ensure the rules as they are are correctly followed by editors because they -are- confusing and its finesses not easy to grasp. It would also mean that a lot of entries are strictly speaking incorrect per the rules and are to be updated. Stonecreek's example is one to correct, and this one too. MagicUnk 10:32, 3 February 2019 (EST)
Neither of the two examples you cite follow that standards and in the case of Gypsy plus ..., it's particularly confusing. First off, since there are no brackets around the roman numbers in the pages field, I can't tell if there are 10 actual pages with roman numbers or not. There is no content listed on these pages, so they should only be included in the Pages field if the pages are actually numbered. The same goes for the "+[1]" page after the main page numbering. Again, we're not listing anything in the contents section, so they shouldn't be noted in the Pages field. Now, for the content indicated only in the notes preceding the 10 pages preceding the numbered pages, I think we would be better served by describing the pages. The note stating that the title is taken from page v, would seem to indicate that "page v" is the title page. If so, that's standard and need not be noted. If page v is something other than the title page, then the note should be expanded as to why the title is taken from that page instead of from the title page. I'm unsure of what is meant by "more feature contents" on "page vii". Is it a table of contents, or is "feature" a typo for "future" and it's really a list of forthcoming books? Neither of those items are significant enough that we need to reflect which page they occur on. Personally I wouldn't bother noting either, but stating that the TOC or a list of forthcoming books appear "after the title page" would seem sufficient. I'm also not clear on what "pp. 143ff" refers to. Assuming the "ff" is a typo, that's within the regular numbered pages. The bibliography and the About the Author essay could be listed in the main contents section. Adding the latter would make "+[1]" appropriate, in which case it should be entered as being on page [146].
The only non-standard usage in Senlin Ascends is the use of "[xii]" instead of [12] and the corresponding listing of content on page "[ix]" instead of "[9]|0.09". I just don't see any advantage to changing the standard from Arabic to roman numbers to indicate unnumbered pages. I wouldn't want to go back and revisit everything that has been entered according the existing standard without a good reason for doing so. --Ron ~ RtraceTalk 18:51, 3 February 2019 (EST)
Well, the "+[1]" page is caused by the section in the help of "Contents included with exceptions": it shows there is a content, but it's not listed as own entry but only in the notes. This, as the explanation for the Roman numerals is explained within the notes. At that time, it seemed better to me to have the preceding pages shown in the page count for the publication, but it would pose no problem for me to alter this. Christian Stonecreek 02:44, 4 February 2019 (EST)
OK, what I've done is update two pub records and applied the page count & page reference rules as they currently exist. Arm of the Sphinx has a before and after unnumbered page range, and Wonderblood has a before roman numeraled page range. Can you have a look and tell me if this makes sense and is per the current rules according to you? Pay special attention to the notes where I used the phrases before page 1, and at the end of the publication to refer to the relevant section. And looking at the contents of Arm of the Sphinx I still find it awkward to do it that way... Anyone a better/other proposal? MagicUnk 13:48, 5 February 2019 (EST)
The problems I see with Arm of the Sphinx is that you have started the numbering over for the post numbered page contents. Per the first bullet after "Pages without a printed page number" in this help section: "If the page is not numbered, and is within a range of numbered pages (i.e. the pages which follow the first numbered page within a publication), its page number can be derived from the nearest numbered page.", thus the interview can be entered on page 405 and the excerpt on page 411, assuming the novel ends on page 398. The rest looks good. For Wonderblood, I probably would have entered the pages as "vii+285" as the highest roman numbered page is vii. If you want to reflect the individual illustrations, they should be entered in the contents section and not in the notes. The titles would be "Wonderblood", "Wonderblood [2]", etc. Their page numbers occurring on unnumbered pages could be derived. If you only want a single INTERIORART record for all the illustrations it shouldn't be disambiguated.
For the former, you can see why I don't think we need to change to bracketed roman numerals for unnumbered pages preceding those that are numbered. There really shouldn't be any confusion as to where something occurs. --Ron ~ RtraceTalk 19:32, 5 February 2019 (EST)
Hi Ron. The issue, I think, we're having is is that it is not clear whether unnumbered pages at the end of the publication is to be considered within a range of numbered pages' or not.
I read here that may record the count of unnumbered pages at the end of a publication. For example, 320+[4].... So far so good.
Then I go on to read here that (2nd bullet) If a content starts on an unnumbered page within a range of unnumbered pages, its page number should first be derived and then entered in squared brackets., and ...if a content appears on the fifth page in a range of unnumbered pages, enter "[5]"., and finally Do not use brackets for unnumbered pages which fall within a range of numbered pages.
In my mind Arm of the Sphinx interview etc. are contained in 'Unnumbered pages within a range of unnumbered pages that come after the numbered range containing the text of the novel proper, so are covered by the second bullet, whereas you infer that the interview etc. are contained in Unnumbered pages within a range of numbered pages, which would require to continue counting, and using the page reference as you suggest.
So question is, how do we determine pages are either within the range of numbered pages, or are outside the range of numbered pages, hence require the second bullet to apply? MagicUnk 07:03, 6 February 2019 (EST)
The definition is in the first bullet of the same section. I quoted this above: "If the page is not numbered, and is within a range of numbered pages (i.e. the pages which follow the first numbered page within a publication)...". The second bulleted paragraph even refers back to the first for the definition: "Do not use brackets for unnumbered pages which fall within a range of numbered pages. (See the first bullet under this subsection.)" Even if we were to ignore these definitions, what would be the point of adding another series of numbered pages at the end of the book? We already know which pages are unnumbered from the pages field. Numbering them in the manner that you have only introduces the confusion between the defined unnumbered section (before any page numbers) and the rest of the book, which I believe is what your proposal is trying to address. --Ron ~ RtraceTalk 19:11, 6 February 2019 (EST)
So, what you are saying is that for unnumbered pages at the end of a pub the second bullet never applies as they willl always fall within the range of numbered pages preceding them, and consequently we are not allowed to use square brackets to refer to these unnumbered pages because they 'belong' to the preceding section as per the first bulllet? MagicUnk 23:17, 6 February 2019 (EST)
I've reread those help sections and I can see how you might interpret them as you do. Given the two interpretations, I believe that maintaining the Arabic numbering through the end of the book works best. Restarting the numbering sequence, or placing page numbers in square brackets at the end of the book servers no purpose that I can see. --Ron ~ RtraceTalk 19:33, 7 February 2019 (EST)

(Unindent) True. Continuously numbering unnumbered pages that appear after the main text was what I've been doing up to now, but now I suspect that this is very much against (the intent of) the current rules. Shall we discuss rules clarification, or do we finish the front/back coverart discussion first? ;) MagicUnk 01:38, 8 February 2019 (EST)

Tagging Inactive users

I don't see any rules for if and when we should hat-notice another user's talk page with Template:Inactive user. For example, User:Dragoondelight was last active on 2010-12-01. On 2012-11-03, or 23 months later, Mhhutchins tagged the person as inactive.

User:Syzygy has been inactive since 2016-10-28. On that date, as part of a reply about a publication question the user included, "I am taking a small break from editing." The break is now at almost at 29 months.

What I would like to see is a note on Template talk:Inactive user with rules or guidelines on if and when an editor should add the template to another user's talk page. If a break is planned, the note should encourage self-tagging, such as what Hauck did on 2018-06-29 which was also that user's last-active date. --Marc Kupper 14:44, 24 March 2019 (EDT)

I see two separate issues here.
The first one is technical. There are three different types of "user activity":
  • database submissions
  • Wiki edits
  • publication verifications
The only way to find the last "user activity date" for a user is to use User Search accessible from the Advanced Search page. I guess we could create a nightly cleanup report that would:
  • find all users with a Talk page and no "Inactive" template in the body of the Talk page
  • for each identified user check if the last activity date is more than N months in the past, where N is to be determined
The second issue is procedural. Some people use the ISFDB on a regular basis, but don't edit it very often. Short of recording each user's last ISFDB access date -- which I don't think we would want to do for privacy reasons -- there is no way of telling if a user may be lurking and available to answer questions. Ahasuerus 15:18, 24 March 2019 (EDT)

Synopsis guidelines

I was asked about guidelines for writing synopis as someone pointed this out to me. Can it have a "judgement"? It's not supposed to be a review. --Mavmaramis 23:37, 6 April 2019 (EDT)

Template:TitleFields:Synopsis says that:
  • this is not a place for criticism or reviews, and should maintain a neutral point of view.
A number of statements in the linked synopsis, including "Fred Pohl should have known better", "The dilemma is solved by a stupid fortuitous event", and "By far the worst story by Pohl I have ever read", are incompatible with the Help text. Ahasuerus 23:43, 6 April 2019 (EDT)
Thanks Ahasuerus. Just to be absolutly clear I did not write that synopsis.--Mavmaramis 04:25, 7 April 2019 (EDT)
Understood. I have removed the spoilers and the commentary. Ahasuerus 10:25, 7 April 2019 (EDT)
I have identified the editor who has contributed 189 synopses. A number of them have similar problems, which I'll need to correct. I will leave a note on the editor's Talk page once I am done. Ahasuerus 10:38, 7 April 2019 (EDT)
Done. Ahasuerus 17:39, 7 April 2019 (EDT)

Variants for uncredited artwork (covers)

For known authors covers are made variants of each other when the have the same artwork (e.g. translations or cover reuse). This especially allows to see where artwork was reused and where and when the original was made. This is not possible for unknown authors.

A few years ago I already asked for a way to link two covers which have no known author, but my proposal to simply add "uncredited" as author and then link the titles like for known authors was rejected. One of the main reasons was the additional load on the "uncredited" author name. Currently a link in the description text is the only solution, but this fails to have all the benefits of "joining" the entries.

Inbetween ISFDB has been changed and "uncredited" is handled specially and no longer shown.

So I do propose again: For uncredited artwork allow the use of "uncredited" for artwork in these cases where needed to make varianted titles possible.


--Stoecker 12:16, 24 April 2019 (EDT)

I have experimented with the proposed change on the development server. Here is what the "Cover" lines on the two Publication pages look like after modifications:
It seems to look OK. As an aside, perhaps we should change "by uncredited" to "(uncredited)", but it would presumably need to affect all record types, not just covers, so we'd need a separate discussion.
The next issue is performance since we would eventually end up with tens -- possibly hundreds -- of thousands of "uncredited" COVERART titles. As Dirk pointed out, we no longer display the Summary page for "uncredited", so that shouldn't be a problem. Also, at this point we have almost 57,000 "uncredited" titles, which doesn't seem to be causing problems. Adding another 100,000+ "uncredited" titles doesn't seem to be likely to cause additional performance issues.
Are there other reasons not to enter uncredited cover art as "uncredited"? Ahasuerus 12:45, 25 April 2019 (EDT)
I very much agree it's important and useful to be able to link appearances of artwork as variants. But I am worried about using specifically "uncredited" to do it. A legacy problem with cover art "credits" is that we don't follow normal publication data element rule of "as it appears in the publication". We credit cover art directly if we are at all able to figure out who the artist is. And in most cases where there is no credit but there is a signature, we normalize that to a canonical name and do not record the signature as the publication's "credit" (although we might mention it in the notes). Given that practice, use of "uncredited" would lead to two inconsistencies: Inconsistency among publication covers -- some where the cover art is not credited in any way would have an actual artist credit, while some would have "uncredited" -- and inconsistency between title records using "uncredited" -- some where it means only "there is no credit in the publication where this appeared" and some where it means "there is no credit in the publication where this appeared, and we have not as yet been able to determine an attribution; this may change in the future."
Also, one other scenario to think about: If we have a tree of variants all for an uncredited piece of art, and for whatever reason we're suddenly able to determine the artist for the original/canonical appearance (easy example: say a later edition identifies the artist), now having anything in that variant tree using "uncredited" would essentially be a data error, as they should ALL be changed to reflect the newly-discovered credit. Of course, we sort of want to ("would", really) do that now, even without the variant relationships.
So I'd rather see something different used just for artists that reflects our artwork crediting practice. For example, "undetermined". Then we'd have three stock terms with well-defined and separate meanings: "unknown" -- we do not know if a credit is present in the publication (could use this everywhere); "uncredited" -- no credit present in the publication, and we will not use any other source in its place; "undetermined" -- no credit in the publication, but we will use another source in its place if we can find one. That would eliminate the inconsistencies. And there could also be some smart processing to auto-replace any "undetermined" occurrences in the variant tree when a determination is made. --MartyD 21:22, 25 April 2019 (EDT)
Not sure if I understand you correctly, so I'd like to add my two cents. You rightly point out that we are handling art credits differently compared to other titles. Because we determine artist credits using multiple sources (as opposed to just recording what's on or in the pub), we can only ever have two objective states for them: they are either credited, or undetermined as of yet (you can also have "unknown" for pieces of art whose creator has been for sure lost to time, for example)
As you define uncredited I understand it's a (subjective) decision by an editor because who's to say someone else isn't determined to find the artist, so state should be "undetermined" as per your definition instead.
To summarize: for art we can only have credited, "undetermined", or "unknown" (the latter requiring an explanatory note as to why we'll never be able to determine the artist). Correct? MagicUnk 02:38, 26 April 2019 (EDT)
That is pretty much what I was thinking, except we can have "unknown" when we do not have access to the book (or a suitable proxy for the physical book). For example, if our data comes from an Amazon listing, and that provides a Look Inside that shows front and back cover and copyright page, "undetermined" would be appropriate -- we are reasonably certain it is not credited in the book, and we have not yet determined who the artist is. If that listing did not provide Look Inside, but only the front cover (and no signature visible there), "unknown" would be appropriate. Maybe that's too subtle, but I was thinking it's useful to understand the difference between the book didn't credit it and we don't know if the book credited it. --MartyD 06:42, 29 April 2019 (EDT)
Sounds fine to me. When talking about policies, this case here should also be discussed: The original is uncredited, the variant has a credit, but very likely it is simply wrong. --Stoecker 02:05, 26 April 2019 (EDT)
One other rather obvious scenario came up this weekend: Two publications with identical, uncredited covers. It would have been useful to show the sharing, and there was also some material about the possible artist that has no single home without a COVERART record to put it into -- the information would have had to have been duplicated in each pub record. --MartyD 06:42, 29 April 2019 (EDT)
There is one other value to cover and one other 'state'. The value is blank, as so many are, and I'm not sure you've covered how you would deal with them - leave them blank (with what interpretations for searches and display) or default them. Which brings up another 'state' - covers for which an artist does not and cannot exist - such as pure text covers. For such covers, does the cover artist field get a special value, such as "none" or is it left blank, leaving it with unwashed masses? By the way, I'm all for being able to link uncredited covers and just want to be sure this doesn't muck things up. ../Doug H 08:15, 29 April 2019 (EDT)
I don't propose to add unknown/uncredited/undetermined (whatever gets decided) to any pub/cover which is not required for linking (i.e used by another cover or interior art). In my eyes that would only add title entries without any real meaning. --Stoecker 17:28, 29 April 2019 (EDT)
I do think the possibility to link (within a given publication entry) to another publication with the same artwork is pretty much what is enough. The complications of using uncredited/undetermined/unknown for artwork in my eyes are not our first business (just my 5 cent of thoughts). Stonecreek 01:50, 30 April 2019 (EDT)
We need to take care not to overthink this. Couldn't we suffice with only two states: either credited or 'unknown'? Credited would be the easy state to define: when we find the artist from whatever source, we credit the artist. 'Unknown' could then cover situations like
  • We honestly don't know, and will never be able to find out who'd the artist been - for example because the artist is lost in time, or
  • Not credited in the pub, and we can't find any secondary sources; which would cover 'uncredited'-as-not-in-the-pub and 'undetermined'-as-not-found-yet; both 'uncredited' and 'undetermined' can be considered substates of 'unknown' (as we're using any source -not just the pub- to determine the artist, uncredited doesn't really apply here)
Add to that that we would use 'unknown' only if we want to 'link' 2 or more records (the rest would remain blank), and we should be good to go. Thoughts? MagicUnk 11:07, 30 April 2019 (EDT)

(unindent) A side question here because I think we are getting into the details before we had decided what exactly we are trying to do (or so it feels to me). Here is an easy example: this set of books will never have a coverart entry under the current rules (no artist; designer does not get). Are we trying to allow covers like that to get connected? If so, we are looking at adding a lot of additional coverart records... Or are we looking only at the case where there is an artist eventually? Annie 11:43, 30 April 2019 (EDT)

No, I wouldn't think so. Adding a cover art record in this case wouldn't serve any purpose. Besides, the covers are already 'indirectly' linked via the common pub title, so no additional information would be conveyed with an additional art record. And I don't think we should add additional restrictions such as being eligible only when an artist could conceivably be identified. Using unknown coverart record when wanting to link should suffice I'd think. The rest can remain undetermined (left blank) MagicUnk 02:09, 1 May 2019 (EDT)
Until a foreign publisher uses that artwork for another book - then you lose the connection from the book level. And if the idea is to connect this covers, where do we draw the line? Annie 02:54, 1 May 2019 (EDT)
Well, yes. If I understand correctly, this is exactly what Stoecker is asking for. MagicUnk 17:04, 1 May 2019 (EDT)
Which is what makes me stop for a second here. What happens if the work that contained the cover get deleted (it does happen). Do we remove the covers from all the parents now? And how do you explain to a new editor that they can add the covers only if there is a cover that does not show up on that screen. And if at the time we now have a cover we need to link, we have 20 books on that screen, do we edit all of them to import the original cover (and if not, how do we chose?) :) We will fill the DB with inconsistencies. Thus my question for a very easy usecase. So we either should just say that "(almost) every book (that has an image linked anyway) needs to have a COVERART" or we will be dealing with art elements for a very long time. And just because they will appear on the same page is not a good reason not to connect the coverarts to start with...
I would love to find a way to link those covers - I just want to keep the DB as consistent as possible while doing it. And there is a difference between "we do not have a coverart because we have no idea who the artist is" and "we do not have a coverart because there is no artist at all (it is a collage or a photograph or whatever). The request solves the first issue; my question is about the latter. Just thinking aloud :) Annie 18:04, 1 May 2019 (EDT)
If we didn't have the legacy data and the very different standard used for cover artist credits to deal with, "uncredited" and "unknown" as they're used for other works would already address what we need. So maybe the simplest thing is to change the policy to make those credits consistent with all other publication data (well, most other publication data): Exactly as it appears in the publication, using "uncredited" if no credit and "unknown" for when we don't have access to the publication and don't know. An artwork-specific sub-policy would specify how to translate a signature into a credit when no other credit is present (following the review/interview model of avoiding creating new pseudonyms). Blank would be allowed as an uncredited equivalent. If we know a publication does not have a credit, and we determine the credit from some other source, then instead of inserting the credit directly, we'd require setting up of "uncredited" and then make that a variant of the credited record. We'd just have to acknowledge that there's legacy data not conforming to that approach that would need to be fixed up over time. Often, we'd be able to identify these situations from notes, but I don't think there would be any urgency to go clean them up. And anyone needing a title record for linking or notes purposes could add "uncredited" as needed and variant to other "uncredited" instances. Easy for me to say; I don't have to do the coding. --MartyD 19:26, 1 May 2019 (EDT)

Is there any conclusion? --Stoecker 08:33, 24 May 2019 (EDT)

It was a good discussion which put a lot of options on the table, but I don't see a consensus. Sometimes it takes a few iterations until we find an approach that addresses all (or at least the majority of) the concerns and can be feasibly implemented. Ahasuerus 10:30, 24 May 2019 (EDT)
It's too bad this died without solving Stoecker's immediate problem. Here's a limited proposal cobbled together from the things above:
  • Allow -- but do not require -- "uncredited" and "unknown" to be used for cover artist, the same way we use those credits elsewhere:
    • "uncredited" = no credit + no signature, and no attribution from another source.
    • "unknown" = we don't know if there's a credit (e.g., we have a listing from some secondary source that is silent on the credit)
  • Create a COVERART record using "uncredited" or "unknown" only for purposes of linking variant appearances of the artwork. Otherwise, omit the COVERART record as is current practice when the artist cannot be determined. We'd also tolerate this use in the absence of linking/variants (i.e., we would not consider such a credit in the absence of variants a data error).
  • Appearances of "uncredited" or "unknown" should be adjusted if more information becomes available.
    • Maybe "we" could make a clean-up report that finds sets of variant COVERART records where one or more variants uses "uncredited" or "unknown" and one or more others use something other than one of those.
Is that something we could agree to? Other than the candidate clean-up report, I think it's something that could be implemented only through policy, with no need for code changes. The software's current handling should be adequate. --MartyD 07:43, 14 June 2019 (EDT)
That would be an enormous help. I have probably hundreds of cases now where the linking is done via the note field, which is error prone and you always have to check the same entries again and again to know if they are missing artists completely and searching original makes sense or if original is already known, but cannot be linked ATM. Maybe that isn't such a big issue in English world, but a lot of German books use uncredited artwork from earlier English books and and proper linking would make maintaining this a lot easier. --Stoecker 09:04, 15 June 2019 (EDT)

Dates on COVERART variants

The Help:Screen:AddVariant's Date bullet states that variant titles should be given the date of that variation's first appearance. This applies whether the variant is due to differences in author credit or title wording, as well as to translations. Handling of a recent submission to alter the date of a COVERART variant revealed that there is a practice of making some COVERART variants use the date of the artwork's first appearance, rather than the date of the first appearance of its variant use, at least in some cases. As I understand it, the idea behind that practice is that unless the artwork itself is titled, there is nothing "variant" about art used for two different publications -- the difference is in the illustrated work's title, not in the artwork.

What do we think? Should the current practice continue and be codified, or should that practice be altered to follow the broader variant dating scheme? One practical problem I see with the current practice is that there would be an inconsistency among some COVERART variants, since presumably we would use different dates if the explicit artist credits were different (just as we would use different dates if an author credit were different). I don't know if requiring more careful thought about what date to use is a good thing or a bad thing. :-) --MartyD 07:21, 14 June 2019 (EDT)

More on the background: Most pieces of art in the database are - contrary to the first sight on a page of ISFDB - not titled at all. It is we who decided to give an illustrative work the title of the piece it is used as an illustration for. I see the immediate merit to give an illustration the title it is first used for: in many cases it was commissioned for that, and it is the most practible way for us to handle pieces of art. However, if it is re-used, there seems not to be much meaning in assigning a different date than that of first publication: the artwork is the same, and it is quite discussable if the 'title' should be altered or remain the initial one. The respective editor uses graphics from out of a pool available at that time and place.
Moreover, what is interesting about a piece of art in respect to its creator for me (and many others I'd think) is its title (mostly unknown to us), when it was created (also mostly unknown to us), and when it was first published (in most cases known). However, we have pieces that were not commissioned for a piece of speculative fiction, and I do advocate that they should bear their date of first publication or becoming publicly available (i. e. pieces of fine art), and it is because for reasons of consistency that those pieces should be catalogued with their contemporary date. That are my thoughts on the matter, anyway. Stonecreek 11:00, 14 June 2019 (EDT)
I am not if favor of a differing standard for dating artwork. As with other title types, having the date of an original publication on the canonical title makes sense and there is nothing to be gained by repeating that date whenever the artist credit or title under which the artwork appears changes. The original publication date is still preserved just as it is with other title types. Myself, I find it useful, or at least interesting to know when a piece of artwork was first published under a given title. I'll also point out that not all titles are our invention, since variants can jump from COVERART to INTERIORART, sometimes the title is determined from the caption where the work is published. My last reason for opposing a change to the existing standard is that multiple standards is more difficult to explain to new users and adds complexity to the rules with no real benefit. --Ron ~ RtraceTalk 18:50, 14 June 2019 (EDT)
I agree with Rtrace. ···日本穣 · 投稿 · Talk to Nihonjoe 20:03, 14 June 2019 (EDT)
That pursue leads to the consequence that interior pieces like - for example - 'Amazing Stories, June 1926 (cover)' haven't got 1926-06-00 like they should (and anyone interested in art should expect) but the date of the publication they got included in. Anyone interested in the chronology of publications a piece of art was published in only has to look at the list. And, I repeat, the piece of art usually was not published under that title, it is only by our assignment that those titles are displayed in the list. Stonecreek 01:11, 15 June 2019 (EDT)
It also doesn't take into account the pieces that are in the database but weren't published initially on a speculative publication: classical / fine art, and works by 'speculative' artists but published for example with a crime fiction or a nonfiction item. We have loads of them from both fractions and it would be of interest when they were first published / publicly available. It would be most comforting to see in the title & in a given publication from what time they date. Stonecreek 02:58, 15 June 2019 (EDT)
I think, the points made by Ron and Christian are both valid. Using Christian's approach for cover art dates adds complexity to the rules and moderation, which I'd rather find undesirable, on the other hand it'd indeed be strange for casual user to see only the date 2002-10-00 for Amazing Stories, June 1926 (cover), and not 1926-06-00. Moreover, in general it'd indeed be interesting to see classic artworks which have their own name listed with this name as the title (and their first date of publication). As for the latter, there are already data like this in the database, see this artwork by Vincent van Gogh or this by Hieronymus Bosch: the painting's title and publication date were used, if known, and the publications using it as cover art are variants of it, using the usual rule "date of variant = first apparance of the variant title".
As a short-term improvement, what if the software also displays the date of the canonical title on a variant's title page if these dates are different? Not only for cover art, but for all title types. Example for "Amazing Stories, June 1926 (cover)" (see the red text):
Variant Title of cover art for: Amazing Stories, June 1926 • (1926) • by Frank R. Paul
That way the user can immediately see the original date (though in this special example the date is already contained in the title itself and therefore redundant).
To sum it up:
  • I'd rather keep the current rule regarding the date of variant covers as it is, but add the software improvement mentioned above.
  • We could add a rule which allow the creation of canonical titles like it was done in the Vincent van Gogh and Hieronymus Bosch examples mentioned above: "If the cover or interior art has been created and published before standalone (e.g. as a painting) and its original name is known, a canonical title for this art can be created, using the artwork's original name and date. If the original date is not known, enter 0000-00-00. All appearances of this art in publications are then variants of the original artwork's title record". This, however, also adds some complexity to the rules, but it's an new rule, not an exception from an existing rule and therefore easier to handle, I think... Jens Hitspacebar 05:44, 15 June 2019 (EDT)
I'm in agreement. For dates, artwork should follow the current rules. There is no reason to hide information (date first appeared under a variant/translation/type) & add rules complexity. For titles, artwork should also follow the current rules. If an artwork has an independent title (whether a classical or modern piece), it can be varianted to a new record with that name. This consistency would make for a better user experience. We already have enough exceptions that make it difficult for new users. As for the software suggestion, if a parent has a different date, the software will display that in the publication view when the publication has the variant. Seems reasonable to do the same thing at the title level. -- JLaTondre (talk) 09:26, 15 June 2019 (EDT)
Sounds good to me: it's a compromise that I hope we all could live with (provided its possible to install). Stonecreek 12:34, 15 June 2019 (EDT)

(unindent) Here's proposed wording intended to cover the above discussion:

Title records for artwork (COVERART and INTERIORART) should follow the varianting and dating rules used for the titles of the works illustrated. If the cover or interior illustration reproduces a piece of art that was originally produced as a standalone titled work -- e.g., such as one of Renoir's paintings or one of Ansel Adams' photographs -- a separate title record may be made to represent that original work and its creation date. In such a case, use the COVERART or INTERIORART title type corresponding to the work's first known use for speculative fiction publication purposes.

One gotcha that occurred to me is there is no generic "ARTWORK" title type, only COVERART and INTERIORART. So I made up a recommendation for what to use. Another thought I have is whether there should be a prohibition against making a separate title records for works commissioned for the SF publication where it first appeared. The ISFDB is not an art database. --MartyD 10:47, 17 June 2019 (EDT)

I was also thinking about the COVERART vs INTERIORART problem and had come to the same conclusion as you did. I'd make a few additions to your proposal which I think makes it a bit clearer (green text by me):
Title records for artwork (COVERART and INTERIORART) should follow the varianting and dating rules used for the titles of the works illustrated. If the cover or interior illustration reproduces a piece of art that was originally produced as a standalone titled work under its own, original name -- e.g., such as one of Renoir's paintings or one of Ansel Adams' photographs -- a separate title record may be made to represent that original work and its creation date. In such a case, use the artwork's original name as title and the COVERART or INTERIORART title type corresponding to the work's first known use for speculative fiction publication purposes. Example record: this artwork by Vincent van Gogh.
As for your last two sentences: I don't understand what exactly you propose to be prohibited, because when a work was, as you wrote, "commissioned for the SF publication", then logically "it first appeared" in genre (because it was commissioned for it) and should, as usual, become the parent title. Either there's something missing there, or my Babelfish has hiccup :) Jens Hitspacebar 15:37, 17 June 2019 (EDT)
What I mean is: We do not want someone to create a record for a Chris McGrath or Chris Moore (for example) painting just because it also exists as a separately titled work on one of their websites. How we distinguish between the two different cases, I do not know. --MartyD 22:01, 18 June 2019 (EDT)

(unindent) Too much travel lately so late to the party. I do not see why we should be treating art differently than the rest of our titles - threat the title as a "title, author, language" trio, date goes to when it was published for the first time in that combination (except when a parent needs to take a date from a child due to weird publishing order). Anything else both hides information that we have and makes it harder to use the DB to find anything (even if art does not change per language except for cartoons, hiding information deep into the lists when we have it is a bad DB design... and makes the DB less usable). Not to mention that it complicates things even more for new editors.

Can someone point out one reason why are we trying to do something funny with dates here besides the "well, they are the same work" argument? We are not an art DB after all, we do not really look at copyright for anything, let's not go down into rules that will be hard to enforce (and research) just because... Annie 14:22, 19 June 2019 (EDT)

The immediate problem I am trying to resolve is that I have an edit on hold that would change the date on variant coverart record to the date of the parent. I would have rejected it, but upon investigating, I found there is a de facto practice, at least among some editors and moderators, of using a single original date for all variations. So I need to accept or reject the edit. If it should be accepted, the help ought to be changed to match. If it should be rejected, we have a consistency problem. My take is that the above discussion is trying to find a way to have the "standard" variant rules apply while also providing a way to capture and present a very different origination date. --MartyD 20:52, 23 June 2019 (EDT)
A few thoughts now that I am feeling better and can do more than mindlessly process Fixer's data:
  • The proposed software improvement looks eminently reasonable and doable. FR 1285 has been created.
  • Early on, the policy was that variant titles shouldn't have separate dates. It's possible that many variant COVERART titles that have the same date as their parent titles had been entered that way before the current data entry rules were adopted.
  • I believe I understand the distinction that Marty and Jens are trying to make in the second part of the proposed language. I appreciate the desire to capture it, but I am concerned that the proposed rule may add a certain amount of complexity and uncertainty.
My suggestion would be to:
  • Change the software as per FR 1285
  • Update Help to clarify that the current data entry rules for variants' date apply to art variants
  • Move the issue of non-genre art titles to a separate discussion
Ahasuerus 11:09, 24 June 2019 (EDT)

(unindent) FR 1285 has been implemented -- see the "Variant Title of" line of this VT page for an example. At this time only the 4-digit year of the parent's date is displayed since that's what we do in the "Other Titles" section of Title pages. Ahasuerus 17:49, 24 June 2019 (EDT)

Notifying the PV - rule update needed?

I'd like to get confirmation from the community whether the implicit rule of not having to (no longer have to) notify the PVer if a clear statement is added in the 'note to moderator' field can continue to be applied or not.

I've now had a price update rejected twice (which was clearly an error as I was the one who entered the wrong price in the first place, and which correction could be unambiguously confirmed to be correct by checking the online source to boot) with the rejection reason being that I didn't notify the PV first. This rejection is not justified imo because

1. I've done quite a few similar non-trivial updates before, where I provided a clear and unambiguous reason in the notes to moderator field, and that passed without any objections whatsoever; not from the approving moderator, and not from the PV, and

2. There have been previous discussions where the outcome pointed to no longer have to adhere to the 'notify the PV first' rule as long as a clear reason is provided, that being the current consensus, or so I thought

Formally documenting current practice in the rules has the benefit of clarity for any and all, and thus remove the inconsistent rules application within the moderators team of said rule. Note that if an explanation is unclear or ambiguous, the moderators remain entitled to still reject these submissions of course. Thoughts on formalizing current practice? MagicUnk 14:03, 3 September 2019 (EDT)

This was for one of my verifications. I have corrected the price, and adapted the message on my talkpage (I don't want to be informed of changes like this). I'm in favor of changing the rules, but not for every change. In the not too distant past we had a moderator who even deleted primary verified pubs without informing the verifier. There should be some value in (primary) verification. --Willem 15:13, 3 September 2019 (EDT)
It really comes down to who the PVs are. When I am approving, I do look at the page of the PV to see what notes they have on notifications. So the same non-trivial update may go through in one case and be held for notifications in another. Which may be a bit confusing for an editor but I do not think that we can change anything around that.
Notifications are still expected for non-trivial edits where editors do not say outright not to notify them. The moderator note field is a great one but... think about some of the more prolific PVs. When I go on one of my cleaning frenzies, some of them can see hundreds of "changed PV" in their lists per day. 10 days vacation can make a mess of their lists (we do not have that many left but when I was doing LCCN and OCLC moves, the numbers were not exaggerated). A major update will get lost in the shuffle - so most people would expect that the non-trivial changes are communicated in other ways as well. The agreement we were veering towards was "do not notify if it is a trivial change" (and noone picked up the definition of trivial although it was tried), not "no need to notify ever" I think.
Prices are especially problematic when they are not physically printed on a book - list prices change and it gets... complicated. And different editors react differently when not notified. :) Outside of English, we rarely have too many editors working in a language - knowing their patterns and expectations helps.
One small note - when adding an explanation, add also what changes to what so we have the old value saved in the notes. That makes it easier for everyone (because once approved, the old value is lost) :) Annie 13:56, 4 September 2019 (EDT)
Personal tools