Difference between revisions of "EditBug:10065 Variant Titles"

From ISFDB
Jump to navigation Jump to search
(Some responses - this is confusing, I agree)
(Brainstorm)
Line 119: Line 119:
  
 
:::-- [[User:Mike Christie|Mike Christie]] [[User_talk:Mike Christie|(talk)]] 07:02, 14 Jan 2007 (CST)
 
:::-- [[User:Mike Christie|Mike Christie]] [[User_talk:Mike Christie|(talk)]] 07:02, 14 Jan 2007 (CST)
 +
 +
(unindent and <nowiki><</nowiki>hr<nowiki>></nowiki>as it's getting a little messy)
 +
<hr> From: [[User:Marc Kupper|Marc Kupper]] 15:22, 14 Jan 2007 (CST)
 +
 +
First, on Aldiss – You had deleted the VT title record for “Non-Stop” by Brian Aldiss. There are real-life publications that use that pattern.  See Aldiss’ [http://www.solaris-books.co.uk/aldiss/html/non-stop.html site] for images of a couple. I looked on Amazon, found one publication of Non-Stop by Brian Aldiss, and found we had a [http://www.isfdb.org/cgi-bin/pl.cgi?69890 publication] though by Brian W. Aldiss.  I fixed the author name and at the moment the publication is filed under the Non-Stop by Brian W. Aldiss [http://www.isfdb.org/cgi-bin/title.cgi?123 title].
 +
 +
This looks ok with me in that we have a [http://www.isfdb.org/cgi-bin/title.cgi?123 title record] called “Non-Stop by Brian W. Aldiss” and it shows publications for
 +
* Non-Stop by Brian W. Aldiss
 +
* Non-Stop by Brian Aldiss
 +
* Starship Brian W. Aldiss
 +
* Starship Brian Aldiss
 +
A thought is when we are dealing with small variations of an author’s name, rather than actual pseudonyms, that we don’t bother with variant title records but rather file these “variant publications” under the title record for the canonical version of the author’s name. We would need to figure out how to deal with people entering new publications under “Brian Aldiss” rather than going to the Brian W. Aldiss [http://www.isfdb.org/cgi-bin/title.cgi?123 title] and selecting “Add Publication to This Title.”
 +
 +
 +
As for [http://www.isfdb.org/cgi-bin/ea.cgi?Test%20Author%201 Test Author 1] – Right now the bibliography shows
 +
:*''[http://www.isfdb.org/cgi-bin/title.cgi?236071 Test Title 1]'' '''(2005)''' <nowiki>[</nowiki>'''as by''' [http://www.isfdb.org/cgi-bin/ea.cgi?Test%20Author%202 Test Author 2] <nowiki>]</nowiki>
 +
I agree it's misleading in that it's not apparent that publications exist for Test Author 1.  Here's an idea. The display logic would look at [http://www.isfdb.org/cgi-bin/title.cgi?236071 Test Title 1] and if there are any direct (not VT) publications it would display as:
 +
:''[http://www.isfdb.org/cgi-bin/title.cgi?236071 Test Title 1]'' '''(2005)'''
 +
:* '''Variant Title:''' [http://www.isfdb.org/cgi-bin/title.cgi?236091 Test Title 1] '''(2006)''' <nowiki>[</nowiki>'''as by''' [http://www.isfdb.org/cgi-bin/ea.cgi?Test%20Author%202 Test Author 2] <nowiki>]</nowiki>
 +
That would make it clearer that publications exist for both Test Author 1 and 2.
 +
 +
I thought about if there are no direct publications under [http://www.isfdb.org/cgi-bin/title.cgi?236071 Test Title 1] we would display
 +
:*''Test Title 1'' '''(2005)''' <nowiki>[</nowiki>'''as by''' [http://www.isfdb.org/cgi-bin/ea.cgi?Test%20Author%202 Test Author 2] <nowiki>]</nowiki>
 +
That means though that it would be hard for future editors to add publications of ''Test Title 1'' by Test Author 1. If there are no Test Author 1 publications though it would be nice though to be able to better indicate that ''Test Title 1'' was only written under the name "Test Author 2" other than it's on a single line in the bibliography.  Maybe this would do the trick:
 +
:*''[http://www.isfdb.org/cgi-bin/title.cgi?236071 Test Title 1]'' '''(2005)''' (no publications)<nowiki> [</nowiki>'''as by''' [http://www.isfdb.org/cgi-bin/ea.cgi?Test%20Author%202 Test Author 2] <nowiki>]</nowiki>
 +
and perhaps if publications exist to use
 +
:''[http://www.isfdb.org/cgi-bin/title.cgi?236071 Test Title 1]'' '''(2005)''' (1 publication)
 +
:* '''Variant Title:''' [http://www.isfdb.org/cgi-bin/title.cgi?236091 Test Title 1] '''(2006)''' <nowiki>[</nowiki>'''as by''' [http://www.isfdb.org/cgi-bin/ea.cgi?Test%20Author%202 Test Author 2] <nowiki>]</nowiki>
 +
 +
--[[User:Marc Kupper|Marc Kupper]] 15:22, 14 Jan 2007 (CST)

Revision as of 17:22, 14 January 2007

  • New Novel: Test Title 1 by Test Author 1
  • New Novel: Test Title 2 by Test Author 2
  • Make Pseudonym: Test Author 2 is a pseudonym of Test Author 1
  • New Novel: Test Title 1 by Test Author 2
  • Make Variant: TT1 by TA2 is a vt of TT1 by TA1
Unless I'm missing something, that should have caused the Test Author 1 biblio to show two version of Test Title 1; one by Test Author 1 and one under the pseudonym of Test Author 2. I'll leave these records out there in case they're needed; let me know if I can delete them. Mike Christie (talk) 10:47, 23 Dec 2006 (CST)

After each step:

  • New Novel: Test Title 1 by Test Author 1
 Author1:
   Test Title 1
  • New Novel: Test Title 2 by Test Author 2
 Author1:                         Author2:
   Test Title 1                     Test Title 2
  • Make Pseudonym: Test Author 2 is a pseudonym of Test Author 1
 Author1:                         Author2:
   Test Title 1                   Alternate Name Used By: Author1
                                    Test Title 2
  • New Novel: Test Title 1 by Test Author 2
 Author1:                         Author2:
   Test Title 1                   Alternate Name Used By: Author1
                                    Test Title 2
                                    Test Title 1
  • Make Variant: TT1 by TA2 is a vt of TT1 by TA1
 Author1:                         Author2:
   Test Title 1 (as by Author2)   Alternate Name Used By: Author1
                                    Test Title 2

Which is as intended. Are you saying you want to see:

 Author1:                         Author2:
   Test Title 1                   Alternate Name Used By: Author1
      Variant: Test Title 1         Test Title 2
      (as by Author2)

Alvonruff 07:00, 9 Jan 2007 (CST)

The current display seems to match my current understanding of ISFDB. Test Author 1 shows
And it does show both the version of Test Title 1; one by Test Author 1 and one under the pseudonym of Test Author 2. Test Author 2 shows
I personally would like it if Test Author 2’s bibliography also showed the variant title so that you could look at a pseudonym bibliography and see all of the works written under that pseudonym but believe that's unrelated to what you reported.
I also posted a question to Al about this as I’m still uncertain as to just what the step “Make Pseudonym: Test Author 2 is a pseudonym of Test Author 1” does to the database but think it’s buried in the queue User_talk:Alvonruff#Help_for_the_pseudonym_editor. Marc Kupper 18:36, 10 Jan 2007 (CST)
Here are two titles from Aldiss's bibliography:
  • Non-Stop (1958)
  • Variant Title: Starship (1985)
  • Variant Title: Starship (1960) [as by Brian Aldiss ]
  • The Primal Urge (1961) [as by Brian Aldiss ]
For "Non-Stop", Aldiss published the first edition as by "Brian W. Aldiss", and then some subsequent reprints in three forms:
  1. as "Non-Stop" by Brian W. Aldiss
  2. as "Starship" by Brian W. Aldiss
  3. as "Starship" by Brian Aldiss
Perhaps because the vts are a different title here as well as a different name, the biblio says "Nonstop (1958)" and does not say "[as by Brian Aldiss]".
For "Primal Urge", on the other hand, the first edition (and the only independent publication recorded in the db to date) was as by "Brian Aldiss". It makes sense for this to show "[as by Brian Aldiss]".
What is good about the Aldiss presentation of the two vts of "Non-Stop" is that I can tell from this that "Starship" was published both as by BWA and by BA. That's the information that is lost in the Test Author display, because it does not use the vt display with as separate line for each title/author combination.
Hence I think what I'm asking for is that a title section in a biblio should not imply (via the "as by") that a book was only published under that variant name. (Ideally it could also report differently depending on which name was used for first, or canonical publication, but I realize that isn't possible as we don't identify first editions directly.) How about if the biblio were to say "also as by" if the biblio's author name is found (along with the pseudonym) in the pub list for the title? For "Primal Urge", no change is necessary since we don't record a version of it as by "Brian W. Aldiss". For the first line "Non-Stop", no change would be necessary as there was no version of "Non-Stop" by "Brian Aldiss" under the original title. However, the second line could then be:
  • Variant Title: Starship (1960) [also as by Brian Aldiss]
For the original display question, it would read
   Author1:
     Test Title 1 (also as by Author2)
These are actually two changes; the change to vt display may not be even desirable since it doesn't add new info, and loses the separate date info that is there now. The change to main (non-vt) title display would add info, though.
This is probably a feature request, not a bug. Al, do you agree? Do you think this is desirable, and well enough defined for me to move over to the feature list? Mike Christie (talk) 05:10, 13 Jan 2007 (CST)
I got confused about what Mike wrote and so will do some thinking out loud:

>Perhaps because the vts are a different title here as well as a different name, the biblio says "Nonstop (1958)" and does not say "[as by Brian Aldiss]".

I suspect the reason it does not say "[as by Brian Aldiss]" is because the first publication in ISFDB had an author credit of "Brian W. Aldiss." Note - ISFDB does not have the first publication which was by Faber and Faber, London and credits "Brian W. Aldiss" but the point is - the current ISFDB display seems to be correct but you may have a point in that if the first publication did happen to be credited to "Brian Aldiss" that ISFDB may not display it correctly.
I did a title search to see if there were any entries for "Brian Aldiss." It turned out there were three title records for "Non-Stop" - all for Brian W. Aldiss. I merged one of them into the original novel and took the second which had no publications and made it a vt using the author name "Brian Aldiss." The bibliography now shows
  • Non-Stop (1958)
    • Variant Title: Starship (1985)
    • Variant Title: Non-Stop (1958) [as by Brian Aldiss ]
    • Variant Title: Starship (1960) [as by Brian Aldiss ]
and if I understand what Mike brought up - it should show
  • Non-Stop (1958) [as by Brian Aldiss ]
    • Variant Title: Starship (1985)
    • Variant Title: Starship (1960) [as by Brian Aldiss ]
Basically, the code needs to look for a VT with the same title and same date and if so display it on the first line. VTs with later dates or different names would get nested.
Or maybe this would be better as it would clearly show that both Non-Stop and Starship have been published (See Aldiss' site) using both Brian Aldiss and Brian W. Aldiss.
  • Non-Stop (1958)
    • Non-Stop (1985)
    • Pseudonym: Non-Stop (1985) [as by Brian Aldiss ]
    • Variant Title: Starship (1969)
    • Variant Title: Starship (1960) [as by Brian Aldiss ]
Mike - I got confused again by your summary. Could you please post a copy of the Aldiss Non-Stop/Starship bibliography as you would like it shown? I'll leave the extra title record in there so you can use it if needed though we should delete it when we're done with this discussion. Marc Kupper 23:49, 13 Jan 2007 (CST)
This is confusing, I agree. I started to answer some of your comments inline and then decided that would be even more confusing. So here's a fresh summary, with reference to some of your points above.
First, I think the biblio for Aldiss now looks as it should. It looked fine with the 169112 title; I've now deleted that and it looks fine now too. In the display you list above, it shows that there was a publication as "Non-Stop" by Brian W. Aldiss, and another publication under the same title as by "Brian Aldiss". That's accurate. Deleting the bogus title eliminates the vt line, as one would expect.
Second, as far as I can tell the Test Author display does not do quite the same thing. This biblio makes it appear that "Test Author 1" wrote "Test Novel 1" as by "Test Author 2", which is correct; but it's not apparent that TN1 was also written as by TA1. And I made sure that TN1 as by TA1 is 2005 (both title and pub record) while the version by TA2 is 2006. So this biblio does not show that. However, your example biblio for Aldiss does show that: you have a "Non-Stop" line and a "Non-Stop" (as by Brian Aldiss) line.
Before you posted, I thought the problem was that a vt that matched in title but not author simply caused the display of the first title to show "as by". That would have meant the "Test Author" display was as designed, but undesirable as it hides the fact that there was ever a pub as by TA1. However, for some reason the Aldiss display doesn't behave that way. It does show two lines for "Non-Stop". (At least it did before I deleted the title; maybe I was too hasty in deleting it. We can add another with a bogus pub and then delete that, if we need to.)
So now I am confused as to the reason why the Aldiss display appears to function differently to the Test Author display, but I believe the latter is misleading. I think Test Author 1's biblio should make it apparent that TN1 was published under two different names.
-- Mike Christie (talk) 07:02, 14 Jan 2007 (CST)

(unindent and <hr>as it's getting a little messy)


From: Marc Kupper 15:22, 14 Jan 2007 (CST)

First, on Aldiss – You had deleted the VT title record for “Non-Stop” by Brian Aldiss. There are real-life publications that use that pattern. See Aldiss’ site for images of a couple. I looked on Amazon, found one publication of Non-Stop by Brian Aldiss, and found we had a publication though by Brian W. Aldiss. I fixed the author name and at the moment the publication is filed under the Non-Stop by Brian W. Aldiss title.

This looks ok with me in that we have a title record called “Non-Stop by Brian W. Aldiss” and it shows publications for

  • Non-Stop by Brian W. Aldiss
  • Non-Stop by Brian Aldiss
  • Starship Brian W. Aldiss
  • Starship Brian Aldiss

A thought is when we are dealing with small variations of an author’s name, rather than actual pseudonyms, that we don’t bother with variant title records but rather file these “variant publications” under the title record for the canonical version of the author’s name. We would need to figure out how to deal with people entering new publications under “Brian Aldiss” rather than going to the Brian W. Aldiss title and selecting “Add Publication to This Title.”


As for Test Author 1 – Right now the bibliography shows

I agree it's misleading in that it's not apparent that publications exist for Test Author 1. Here's an idea. The display logic would look at Test Title 1 and if there are any direct (not VT) publications it would display as:

Test Title 1 (2005)

That would make it clearer that publications exist for both Test Author 1 and 2.

I thought about if there are no direct publications under Test Title 1 we would display

That means though that it would be hard for future editors to add publications of Test Title 1 by Test Author 1. If there are no Test Author 1 publications though it would be nice though to be able to better indicate that Test Title 1 was only written under the name "Test Author 2" other than it's on a single line in the bibliography. Maybe this would do the trick:

and perhaps if publications exist to use

Test Title 1 (2005) (1 publication)

--Marc Kupper 15:22, 14 Jan 2007 (CST)