Talk:Requirements:Author Display

Title Structure

Question 1 - Omnibus Annotations

What should the omnibus annotation look like? At present there are two types of annotation:

1. The John Wenn style. In this case, a summary of the book types is listed in the omnibus annotation. For instance [O/3N] indicates that the omnibus consists of 3 Novels, while [O/2N/1C] indicates an omnibus consists of 2 Novels and 1 Collection.
2. The series style. In this case, the annotation indicates which books in a series are present in the omnibus. For instance [O/4,5,6] indicates that the omnibus contains the 4th, 5th, and 6th novels in the series.

I have been using the Series style notation and also updating the title notes to link to the sub-titles as a way of showing the relationships between titles. On Marion Zimmer Bradley's page see the "Darkover Omnibus" sub-series where I took advantage of this. Marc Kupper 16:53, 23 Dec 2006 (CST)

I don't think the two styles are mutually exclusive. I have been using numbers whenever an omnibus collects novels in a series and the novels' series numbers are known. It wouldn't be very informative to say "this omnibus collects three novels in Pratchett's 20+ novel series", would it? :) OTOH, if the omnibus is a random reprint of some unrelated novels, e.g. the old Thorne Smith three deckers or some recent GuildAmerica cheapies, then all you can do is use something like O/3N. Ahasuerus 18:26, 23 Dec 2006 (CST)

QUESTION 1 Disposition: Both styles of omnibus annotation are valuable. When an omnibus covers specific titles in a series, then the [O/4,5,6] style of annotation should be used. When an omnibus covers non-series titles (for instance, titles by different authors) then the [O/2N1C] style notation should be used.

Question 2 - Should Serial Data be Presented with the Title

Should serial printings NOT be displayed under the title? The series information printing would be restricted to the Series section of the author bibliography, and in the Series section of the title bibliography.

I'm all in favor of showing the relationships between titles and would want to keep the serial printings as so far they are not common enough that they are a distraction. I suspect at some point ISFDB will have +/- buttons to expand out the details or to remove the clutter depending on your POV. Marc Kupper 16:53, 23 Dec 2006 (CST)

It would be a question of personal preference that we could eventually address with the +/= button that Marc proposes or under User Preferences if it was't for one thing. Our current standard for novel dates is to display the date of first book publication. If we hide Serial information, then there will be no easy way of telling when the novel first appeared in print, which can be misleading. As I mentioned elseWiki a few days ago, I have seen two cases of confusion (on r.a.sf.w.) due to this issue in the last 6 months. If we were to remove the Serial data from the Long Works page, then I think we would need to change our date capturing standard. Not insurmountable, but something to think about. Ahasuerus 18:32, 23 Dec 2006 (CST)

regarding "Our current standard for novel dates is to display the date of first book publication." - Where is that standard implemented? Help:Screen:EditTitle says “The original publication date of this title.” and does not mention book vs. other styles of publication. Marc Kupper 02:13, 24 Dec 2006 (CST)

The next paragraph on the same page says:

If a work has been serialized in a magazine, there may be a difference between the first magazine publication and the first book publication. "Skylark of Valeron (http://www.isfdb.org/cgi-bin/title.cgi?1195)", for example, appeared as a magazine serialization in 1933-4, but was not published in book form until 1949. In these cases, record the first book publication date. A note in the title field can record the magazine serialization date, if there is one.

This is in line with what genre encyclopedias like Cluter/Nicholls typically do, except that they list the date of the first book publication in bold and the date of the first magazine publication in regular or italicized font. We may be able to duplicate this convention by deriving the first "book publication date" from the first "book" Publication for this Title and the first "magazine publication date" from the first serial on file. However, there will likely be many borderline cases and, besides, as we discussed earlier, our data is currently not comprehensive enough to make these derivations reliable.

I think the bottom line is that both data elements are useful and that the current scheme, although rather space consuming in some cases (see Keller or even Heinlein), displays them both. Whatever we do, we need to be careful not to lose this functionality and especially not to confuse our users wrt publication dates. Ahasuerus 14:14, 24 Dec 2006 (CST)

Thank you – I had read that second part about book vs. magazine dates before but as it was not relevant to the titles I was entering at the time I just skimmed over it and forgot the details. A few thoughts:

• Many people will not have nor even be familiar with secondary sources such as Clute/Nicholls and bibliographic standards/conventions. They will be entering/verifying information based on copy of a book or magazine in their hands.
• I suspect it would help if the field labels on the forms/displays were a little more descriptive than just “year” or “date” so that someone that the average person will have an understanding of just what data ISFDB is trying to collect.
• I believe that ISFDB has enough publication data today that an accurate first-book-publication and/or magazine publication can be computed/displayed shown for most titles. I’m basing this belief on that it’s pretty rare that I’m adding new authors, titles, or first publications. What I am finding is that the basic information is already there but often details like the month, cover artist, are not in ISFDB. The field most likely to be wrong is the binding where I’ve changed tp to pb a number of times.

We may be working with different subsets of data here, Marc, and my experience is the exact opposite of yours. For example, I was checking David J. Lake's bibliography the other day, and of the 7 novel titles listed by the ISFDB, only 3 had Publication data :( Ahasuerus 17:42, 29 Dec 2006 (CST)

• With the idea that ISFDB’s existing data is already complete enough that you can compute first-book-pub or magazine then if someone comes along and sees the computed date is wrong they can fix it by adding a placeholder publication along with notes on the source of the information. For example, I may have a book whose copyright page mentions it’s a reprint and gives the original publisher name/date. That should be enough to create a placeholder record and the notes field will give the source of the data for that record. Marc Kupper 18:15, 24 Dec 2006 (CST)

QUESTION 2 Disposition: We should keep the current style of displaying serial information under the associated title.

Question 3 - Serial Data Label

Presuming that we want to preserve the printing of serial appearances with the parent title, what label should be used? We currently use "Magazine/Anthology Appearances", but I think that originates from a misty time in our history, and I can't think of a reason why an anthology appearance would show up here any longer. Is there a better label that would be clearer?

I have no problem with the existing labeling and as it is, just last night I ran across (Part 1) of a story in an anthology. Marc Kupper 18:30, 24 Dec 2006 (CST)

QUESTION 3 Disposition: We should keep the current style of labeling serial data.

Question 4 - Typography of Conjunctive Clauses

Many of the conjective clauses in the title strings ("with", "as by", "and") are displayed in a bold font. Recently Mike filed a display bug (20040) which discusses items which are unnecessarily bolded. Would people prefer these to be displayed in a normal, italic, or bold font?

I think we do want them to stand out, but probably not as strongly as bolding implies. Italicizing sounds like a reasonable compromise. (Besides, I am so much richer in liras!) Ahasuerus 21:25, 23 Dec 2006 (CST)

In the example (see Search the Sky on page 128 of the contents) cited by the bug report changing the and to a standard font (no decoration) should do as the author names on either side are hyperlinks and the text is black on white. I would not want to do an across the board fix for conjunctive clauses as bold may be appropriate for them. Marc Kupper 18:43, 24 Dec 2006 (CST)

I tend to agree with Marc on this one; I don't think any decoration is necessary in this context. Mike Christie (talk) 13:48, 29 Dec 2006 (CST)

Question 5 - Series Child Ordering

There are numerous series which are parents to both titles and children series. In what order should the series and titles be presented? That is - should it be titles first, series second, which is the current algorithm:

 Star Trek Next Generation
   * Star Trek, The Next Generation: The Continuing Mission (1997) by 
     Garfield Reeves-Stevens and Judith Reeves-Stevens 
   * Star Trek Next Generation Numbered
         o Encounter at Farpoint (1987) by David Gerrold
         o Reunion (1991) by Michael Jan Friedman 
   * Star Trek: Next Generation Giant
         o Death in Winter (2005) by Michael Jan Friedman 

Or should it be displayed series first, titles second, as in:

 Star Trek Next Generation
   * Star Trek Next Generation Numbered
         o Encounter at Farpoint (1987) by David Gerrold
         o Reunion (1991) by Michael Jan Friedman 
   * Star Trek: Next Generation Giant
         o Death in Winter (2005) by Michael Jan Friedman 
   * Star Trek, The Next Generation: The Continuing Mission (1997) by 
     Garfield Reeves-Stevens and Judith Reeves-Stevens

For now I'd suggest the first approach is marginally better. Mike Christie (talk) 13:47, 29 Dec 2006 (CST)

For a mega-series like the Forgotten Realms or Star Trek, we probably want to display meta-titles that do not belong to any particular subseries first. One assumes that some of them may be important titles that anchor the whole fictional universe, although no examples come to mind at the moment. Ahasuerus 16:34, 29 Dec 2006 (CST)

Question 6 - Series List Ordering

When titles are printed in a series list, there are two types of titles: those that have a series number associated with them (implying a specific ordering), and those that don't (which are then printed in chronological order. At present the non-ordered titles are printed before the ordered titles, as in:

 Foundation Universe
   * Foundation
         o The Foundation Trilogy (1964) [C]
         o Foundation Series (1991) [O/1,2,3,4]
         o 1 Foundation (1951) [C]
               * Variant Title: The 1,000 Year Plan (1951) 
         o 2 Foundation and Empire (1952) [C]
               * Variant Title: The Man Who Upset the Universe (1952) 
         o 3 Second Foundation (1953) [C] 

It is my impression that there are those who would like to see the non-ordered titles follow the ordered ones, as in:

 Foundation Universe
   * Foundation
         o 1 Foundation (1951) [C]
               * Variant Title: The 1,000 Year Plan (1951) 
         o 2 Foundation and Empire (1952) [C]
               * Variant Title: The Man Who Upset the Universe (1952) 
         o 3 Second Foundation (1953) [C] 
         o The Foundation Trilogy (1964) [C]
         o Foundation Series (1991) [O/1,2,3,4]

I understand the desire to allow ordering all titles and give them alternate labels, but this is a different question. When some titles have not been given an order, what is the preference in where they are printed within a series listing?

If it's easy, I suggest we switch to the latter style, but not if it's a big change. Mike Christie (talk) 13:49, 29 Dec 2006 (CST)

I would suggest the latter style as well. For the most part, if series order is unknown for some entries in a series, chances are that it will be the later entries that are shaky. I was adding the previously discussed Russian "sequels to Hamilton by other hand" the other day and the fact that they were displayed first (in Cyrillic, no less!) was confusing. Ahasuerus 16:40, 29 Dec 2006 (CST)

I had initially suggested that un-numbered items get inserted in with the numbered ones by date. If you can do sort-compare functions then the code is ret = (both items are numbered) ? (title1_seriesnum – title2_seriesnum) : (title1_year – title2_year). Marc Kupper 23:48, 29 Dec 2006 (CST)

There are a quite few series entries that do not have and can't have series numbers, e.g. guides (Xanth, Kurtz, etc). If you have entries 1, 2 and 3 nicely cataloged, you probably don't want to insert an unnumbered book between entries 2 and 3 just because the date suggests that it might fit there. I think it holds generally true that explicit series numbers trump dates since our dates are usually less reliable than our series numbers and also since series can be ordered non-chronologically. Ahasuerus 00:49, 30 Dec 2006 (CST)

Until manual series ordering is implemented I doubt any of always on top, always on the bottom, or sort by date, for the unnumbered items will be deemed “intuitively correct” by editors. I was not concerned about ISFDB’s dates being unreliable as someone can always correct the dates if they upset a particular’s series’ ordering.

I had proposed “sort by date” for unnumbered items as it neatly takes care of cases where book one is unnumbered and someone decides to make a series out of it and also that it will tend to get omnibus and other titles in the correct place. If it’s a really long running series where something like the omnibus titles are a distraction then they can get dropped into a sub-series like I did with Darkover. Marc Kupper 03:52, 30 Dec 2006 (CST)

The point about omnibuses is well taken in that John Wenn's original online biblios tended to have series omnibuses listed after the series entries that they collected, and it worked pretty well. So the omnibus that collected the first three Xanth novels edappear between novels #3 and #4 in the series and so on. Unfortunately, omnibuses collecting series entries N through N+X are not always published until long after after series entry N+X+1 (or even N+X+3) appeared, so there is no guarantee that automatically inserting them in the series order based on their publication date will make any sense to the users. Also, omnibuses do not always collect series entries sequentially. The way John Wenn solved these problems was by arranging all entries/omnibuses in the series manually, but then he did everything else manually too, which we are in no position to do.

Once thing than comes to mind is that if/when Al implements that extra ordinal for series entering, we may be able to use it to effectively insert omnibuses into the series order. I am not sure what the ramifications might be, though. Ahasuerus 14:26, 30 Dec 2006 (CST)

Question 7 - EDITOR ordering

If an editor is responsible for a number of magazines, all published at the same time -- a common practice in the 1930s-1960s -- then we end up with a spaghetti-like display, e.g. see Robert_A._W._Lowndes' summary display. Should we change it so that the magazines would be sorted alphabetically? Ahasuerus 22:14, 28 Dec 2006 (CST)

I think what we really need here is "Magazine support", whatever that means. I don't really like the way titles are removed from mag pubs for the benefit of biblio displays, though I understand why it's done. Mike Christie (talk) 13:55, 29 Dec 2006 (CST)

It does feel cludgy to me as well, but if we are going to continue this practice, how about taking the next step and merging all Magazine-specific EDITOR titles into one record that would look exactly like what Mike outlined below? Once you clicked on that megatitle, you would be taken to a screen listing all issues/Publications for that magazine/editor combination. The downside would be more Pubs to list for the megatitle, but the upside would be that the Summary Bibliography page would be easier to navigate. If an editor changed pseudonmy in the middle of a run (which has been known to happen, see Startling Stories), then we would have to break the EDITOR Title into two, one for each form of the editor's name/pseudonym, I suppose. Ahasuerus 16:47, 29 Dec 2006 (CST)

Ideally Lowndes' display would look like this:

Magazines edited

• Science Fiction Quarterly - 1941-1943, 1951-1958
• Future Combined with Science Fiction Stories - 1950-1951
• Dynamic Science Fiction - 1952-1953 (as by Robert W. Lowndes)
• Future Science Fiction Stories - 1952-1960
• Magazine of Horror - 1963-1971
• Bizarre Fantasy Tales - 1970-1971

-- Mike Christie (talk) 13:55, 29 Dec 2006 (CST)

Perhaps what we want instead are editor series. All the Magazine of Horror issues listed chronologically, then all the Science Fiction Quarterly issues. Alvonruff 17:03, 29 Dec 2006 (CST)

How would we enforce the chronological order? By assigning Series Numbers to each EDITOR record? Ahasuerus 17:13, 29 Dec 2006 (CST)

Example of what I had in mind. Current subset of John W. Campbell's magazines from 1937-1947:

 o Astounding Stories - 1937 (1937)
 o Astounding Science Fiction - 1938 (1938)
 o Astounding Science Fiction - 1939 (1939)
 o Unknown - 1939 (1939)
 o Astounding Science Fiction - 1940 (1940)
 o Unknown - 1940 (1940)
 o Astounding Science Fiction - 1941 (1941)
 o Unknown - 1941 (1941)
 o Astounding Science Fiction - 1942 (1942)
 o Unknown - 1942 (1942)
 o Astounding Science Fiction - 1943 (1943)
 o Unknown - 1943 (1943)
 o Astounding Science Fiction - 1944 (1944)
 o Astounding Science Fiction - 1945 (1945)
 o Astounding Science Fiction - 1946 (1946)
 o Astounding Science Fiction - 1947 (1947)

Would instead look like:

 o Unknown - 1939-1943
   o Unknown - 1939 (1939)
   o Unknown - 1940 (1940)
   o Unknown - 1941 (1941)
   o Unknown - 1942 (1942)
   o Unknown - 1943 (1943)
 o Astounding / Analog (1937-1971)
   o Astounding Stories - 1937 (1937)
   o Astounding Science Fiction - 1938 (1938)
   o Astounding Science Fiction - 1939 (1939)
   o Astounding Science Fiction - 1940 (1940)
   o Astounding Science Fiction - 1941 (1941)
   o Astounding Science Fiction - 1942 (1942)
   o Astounding Science Fiction - 1943 (1943)
   o Astounding Science Fiction - 1944 (1944)
   o Astounding Science Fiction - 1945 (1945)
   o Astounding Science Fiction - 1946 (1946)
   o Astounding Science Fiction - 1947 (1947)

Alvonruff 17:27, 29 Dec 2006 (CST)

Oh, I see! Yes, that would definitely be an improvement and address the spaghetti issue. Ahasuerus 17:33, 29 Dec 2006 (CST)

Question 8 - Numbering subseries within their superseries

This is a potentially broader issue, but it's mostly noticeable in Author biblio pages, so might as well discuss it here. At this time, there is no way to order subseries within their parent series. This can easily lead to confusion when their chronological order is reversed or otherwise violated. I suggest we add a "Number within the Superseries" field to the "Series data" record. At some point we will also want a repeating "Web page" field and a Wikipedia link field, but that's a different story.

Definition vs. Requirements

Are the current page contents defining how the display is currently done or is all of this “proposed work?” The reason I ask is I was having trouble visualizing some things and was going to ask for examples, or maybe draw up examples and ask “is this what you mean?” when I realized a lot of it seems close to what is already coded meaning we could link to existing live data and cite it as examples of things that look good or areas that need change and then someone could mock up an prototype page (by copy/paste/edit of the HTML from a live page) of the new & improved page. Marc Kupper 17:03, 23 Dec 2006 (CST)

Unless we're starting over from scratch (which I don't advocate), I'm going to document how it is (and yes we need some examples, as the grammar is pedantic, but obtuse), and use that as a starting point for documenting how we want things to be. Questions like how omnibuses are supposed to be annotated arise from trying to document the hows and whys of the current display. I'll put questions here as they arise, while I fill in the rest of the page over the next couple of days. But I did want to nail down how we feel the pages should look before tackling individual display bugs. Alvonruff 21:18, 23 Dec 2006 (CST)

Variant titles vs. pseudonyms

One thing that is moderately confusing is the fact that pseudonymous variant titles are not only stored as "variant titles" (which is as it should be), but are also displayed as vts, e.g.:

 Sixth Column (1941)
   * Variant Title: The Day After Tomorrow (1951)
   * Variant Title: Sixth Column (1941) [as by Anson MacDonald ]

Could we change it so that if the vt in question is the same as the parent title, then the label indicates that it's a pseudonymous appearance? That way our users won't have to hunt for the non-existing differences between the two titles :) Ahasuerus 21:33, 23 Dec 2006 (CST)

Feedback - stray publications

The new bibliographic displays look good. One question has to do with stray publications. I was looking at Dray Prescot’s page and saw a bunch of stray publications. As this was a pseudonym I added that and the list of stray publications was no longer visible.

Is there a way to get the stray publication list for an author that has titles or is a pseudonym? As it was – most, if not all of Prescot’s strays were because the books are recorded as being written by both Prescot and Alan Burt Akers and when the author order got flip/flopped it’s flagged a stray. Marc Kupper 00:03, 30 Dec 2006 (CST)

At one point Al added a tool to add Strays to the navbar. It didn't work too well and was quickly withdrawn. I don't know if Al has plans to revive it. Ahasuerus 00:52, 30 Dec 2006 (CST)

Pseudonym display on author page

I figure this is a good place to ask about this, and I'll make it a feature request if it's a good idea. How about a section near the top, perhaps in the metadata, that says "has used the following alternate names" which then lists the results of a pseudonym lookup? Unless I'm missing something, this is not there right now. Mike Christie (talk) 17:25, 31 Dec 2006 (CST)

Navigating Long Enough Number of Entries

For authors with enough entries can a short nav bar go between the author personal data and the list of publications?

bignum Novels  bignum+ Collections Bignum++ Shortfiction

where each links to a anchor in the page. It's possible that the code to check if it's big enough will consist of if author == Asimov but there might be others where the list is long enough that just scrolling is painful. Dana Carson 19:33, 18 May 2007 (CDT)