Work-To-Work Relationships vs. Recommendations etc.
Talk Recommend Site Improvements
Join LibraryThing to post.
This topic is currently marked as "dormant"—the last message is more than 90 days old. You can revive it by posting a reply.
1jjwilson61
In the announcement thread in the New Features group we already discussed how the Contains relationship should affect your recommendations and other features like Members With Your Books (MWYB) and checkmarks but the other relationships should matter too.
For example, in this thread, http://www.librarything.com/topic/110645, it was mentioned that in books where later editions add material but don't rewrite the existing material that we generally combine them. But with the new Expands relationship maybe we should separate them but I'd still want there to be a social connection between owners of the different works.
Maybe somebody who owns an expanded edition shouldn't be connected as strongly to the owner of the original work, so perhaps for recommendations they should only be counted as half-strength. MWYB is more tricky. Should the book be shown as shared if it's really only connected with an Expanded relationship? Probably not, but what if it were in the MWYB list in italics?
The above should also apply to all the other relationships as well, I think, but maybe the Inspired relationship should be much more tenuous than the Expanded relationship, for example.
For example, in this thread, http://www.librarything.com/topic/110645, it was mentioned that in books where later editions add material but don't rewrite the existing material that we generally combine them. But with the new Expands relationship maybe we should separate them but I'd still want there to be a social connection between owners of the different works.
Maybe somebody who owns an expanded edition shouldn't be connected as strongly to the owner of the original work, so perhaps for recommendations they should only be counted as half-strength. MWYB is more tricky. Should the book be shown as shared if it's really only connected with an Expanded relationship? Probably not, but what if it were in the MWYB list in italics?
The above should also apply to all the other relationships as well, I think, but maybe the Inspired relationship should be much more tenuous than the Expanded relationship, for example.
2brightcopy
I don't think that topic has fully explored the issue. At some point, surely with enough expansion and you get a different work. And I'd say that bar is a lot lower when it comes to fiction than nonfiction.
I'm not sure Expands has really been pitched as a "revised edition" feature, though, has it?
I'm not sure Expands has really been pitched as a "revised edition" feature, though, has it?
3_Zoe_
For recommendations, I don't think it's worth doing more than applying a basic filter to exclude duplicates.
Likewise, I wouldn't bother including relationships in MWYB. Unless people are using LT in a non-standard way (e.g., cataloguing individual short stories), the percentage of books affected by relationships is probably going to be very small. I don't think it would have a noticeable impact on the MWYB list.
As to whether we should have a way to show "shared" expanded/whatever books in any library, not just in MWYB, I think that would be reasonable but not essential.
In general, I'd like to keep to the basic rule of "relationships don't change what constitutes a work". Otherwise it can get very messy very fast and I just don't think the benefits are there.
Likewise, I wouldn't bother including relationships in MWYB. Unless people are using LT in a non-standard way (e.g., cataloguing individual short stories), the percentage of books affected by relationships is probably going to be very small. I don't think it would have a noticeable impact on the MWYB list.
As to whether we should have a way to show "shared" expanded/whatever books in any library, not just in MWYB, I think that would be reasonable but not essential.
In general, I'd like to keep to the basic rule of "relationships don't change what constitutes a work". Otherwise it can get very messy very fast and I just don't think the benefits are there.
4andejons
>3 _Zoe_:
It's not just MWYB, it's also things like comparison with individual libraries, recommendations and reviews. In my library, it's something like 5-10 % of the books that are affected (and not becuase of short stories, but because of omnibuses and books split in several volumes). Sure, it works OK today, but I would really prefer not to have to keep pseudobooks to have the connection to the Hitchhikers' Guide to the Galaxy.
It's not just MWYB, it's also things like comparison with individual libraries, recommendations and reviews. In my library, it's something like 5-10 % of the books that are affected (and not becuase of short stories, but because of omnibuses and books split in several volumes). Sure, it works OK today, but I would really prefer not to have to keep pseudobooks to have the connection to the Hitchhikers' Guide to the Galaxy.
5_Zoe_
>4 andejons: For MWYB, my question is how much this will affect the overall list of similar libraries. Are the people who just have lots of your non-Hitchhiker's Guide books different from the people who have lots of your non-Hitchhiker's Guide books PLUS the Hitchhiker's Guide? I'm just not convinced that excluding a random 5-10% of your books changes anything.
What do you want it to do with reviews?
What do you want it to do with reviews?
6lorax
I don't really care much about how it will affect MWYB; I just really want to stop seeing recommendations for things I already have cataloged because I have them as an omnibus and I get recommendations for the individual pieces, or vice versa. I think this is what _Zoe_ means by a "basic figure to exclude duplicates" in #3?
7jjwilson61
I'm just not convinced that excluding a random 5-10% of your books changes anything.
Maybe so much when you're talking about recommendations, but when you want to see what two libraries have in common it makes a difference.
Maybe so much when you're talking about recommendations, but when you want to see what two libraries have in common it makes a difference.
8infiniteletters
6. This. Yes.
9andejons
>4 andejons:
My point was that there are other connections than MWYB, and that I would expect LT to take advantage of the data (it's possible that I misread you and took you as more generally negative to using the data instead of just being opposed to using it for MWYB).
Anyway, if 10 % is a reasonable estimate of the number of extra books that could appear, I'd say that the potential 20 % difference (remember: two libraries) is not to be discounted. It will perhaps mostly reorder entries and change some of the libraries at the end of the list, but I don't see why the data shouldn't be used if it's there.
As for reviews, I frequently use the reviews page, as I like reading what other people thought about my books. It would be nice to be able to see reviews for contained books there as well.
Bottom line: while I think excluding books from recommendations based on some of the relationships is probably the easiest and most obvious thing to do and what will happen first, I think it would be worthwhile to use the data for other things as well. I can get the same effect by adding some "works" in addition to my books, but it's not a solution I really like.
My point was that there are other connections than MWYB, and that I would expect LT to take advantage of the data (it's possible that I misread you and took you as more generally negative to using the data instead of just being opposed to using it for MWYB).
Anyway, if 10 % is a reasonable estimate of the number of extra books that could appear, I'd say that the potential 20 % difference (remember: two libraries) is not to be discounted. It will perhaps mostly reorder entries and change some of the libraries at the end of the list, but I don't see why the data shouldn't be used if it's there.
As for reviews, I frequently use the reviews page, as I like reading what other people thought about my books. It would be nice to be able to see reviews for contained books there as well.
Bottom line: while I think excluding books from recommendations based on some of the relationships is probably the easiest and most obvious thing to do and what will happen first, I think it would be worthwhile to use the data for other things as well. I can get the same effect by adding some "works" in addition to my books, but it's not a solution I really like.
10jjwilson61
I agree that the Contains relationship should be used for all these things, but I started this thread to discuss the other relationships. Do you want to be informed that another library contains the expanded version or an abridged version of something in your library?
11brightcopy
10> Do you want to be informed that another library contains the expanded version or an abridged version of something in your library?
Personally, I don't care. For that level of interest, I think it's splitting hairs.
Personally, I don't care. For that level of interest, I think it's splitting hairs.
12jjwilson61
I was hoping that having relationships used in this way could resolve some of the controversial work combinations in a way that worked for both sides.
For example, I can see the argument that different translations of the same text can be very different (I know that Translation isn't a recognized relationship yet) and shouldn't be considered the same work. If there were a Translation relationship and things like Recommendations and MWYB used that relationship then they could be separate works but still be tightly bound.
For example, I can see the argument that different translations of the same text can be very different (I know that Translation isn't a recognized relationship yet) and shouldn't be considered the same work. If there were a Translation relationship and things like Recommendations and MWYB used that relationship then they could be separate works but still be tightly bound.
13Nicole_VanK
When I asked about something like that (dead language editions vs. modern translations) I got the answer that would be for a next phase (expressions).
14brightcopy
12/13> Right. Relationships are work-to-work things (and possibly work-to-series in the future, maybe even work-to-part-of-work - think short stories in an anthology). Expressions would be what Tim has talked about for translations, with all the translations of a work rolled up into a single work. So while there may be some interplay with different translations in the future, it won't really have anything to do with Relationships as far as I can see.
15jjwilson61
13,14> Right, but Translations was the clearest example I could think of. There are also edge cases where it's a close thing whether a later edition should count as the same work or not. In those cases being able to separate them but retain some connectivity through relations would be a good way to resolve it.
BTW, I believe Tim was talking about expression-to-expression relationships too. But despite Tim explaining it, exactly what an expression is has not stuck in my head and I don't think I'm the only one. Maybe he needs a new word, but as it is now I'm not sure that many people will understand the new system.
BTW, I believe Tim was talking about expression-to-expression relationships too. But despite Tim explaining it, exactly what an expression is has not stuck in my head and I don't think I'm the only one. Maybe he needs a new word, but as it is now I'm not sure that many people will understand the new system.
16Nicole_VanK
I'm sure I'm as puzzled as you are.
(Doesn't mean we won't unltimately come to an understanding though. And I applaud your efforts in at least throwing up the questions.)
(Doesn't mean we won't unltimately come to an understanding though. And I applaud your efforts in at least throwing up the questions.)
17brightcopy
15/16>
Work: Beowulf
The publication lines are what we see today on the "editions" (a misnomer, once these features show up) pages of works. This is made up of the title/author/isbn from various members Books. Then those are bundled up into a single Edition, such as all the different paperback copies of Seamus Heaney's Bilingual Edition. Then all of those are bundled into an Expression, such as "Seamus Heaney translations of Beowulf". That Expression is actually a subset of the Work Beowulf.
Does that clear any of it up?
Work: Beowulf
Expression: Seamus Heaney translations of BeowulfEdition: Beowulf: A New Verse Translation (Bilingual Edition) (Paperback)Publication Lines:
Beowulf: A New Verse Translation (Bilingual Edition) / Heaney, Seamus (ISBN 0571203760)
Beowulf: A New Translation / Anonymous (ISBN 0571203760)
Beowulf : a new verse translation / Heaney, Seamus (ISBN 0571203760)
Beowulf (Bilingual Edition) / Heaney, Seamus (ISBN 0571230415)
Beowulf / Heaney, Seamus (ISBN 0571230415)
Beowulf A New Verse Translation Bilingual Edition / Heaney, SeamusBook: http://www.librarything.com/work/2449742/book/64658676
The publication lines are what we see today on the "editions" (a misnomer, once these features show up) pages of works. This is made up of the title/author/isbn from various members Books. Then those are bundled up into a single Edition, such as all the different paperback copies of Seamus Heaney's Bilingual Edition. Then all of those are bundled into an Expression, such as "Seamus Heaney translations of Beowulf". That Expression is actually a subset of the Work Beowulf.
Does that clear any of it up?
18jjwilson61
What could sub-divide a Work into an Expression other than a translation?
19infiniteletters
Illustrations vs No Illustrations?
20lorax
18>
What could sub-divide a Work into an Expression other than a translation?
Illustrations, introductions, scholarly notes -- used properly this could resolve the age-old issue of the Norton Critical Editions; while the extra material deserves to be noted, and currently separation is the only way to do that, this could also accomodate that -- revised editions, British vs. American versions, etc.
What could sub-divide a Work into an Expression other than a translation?
Illustrations, introductions, scholarly notes -- used properly this could resolve the age-old issue of the Norton Critical Editions; while the extra material deserves to be noted, and currently separation is the only way to do that, this could also accomodate that -- revised editions, British vs. American versions, etc.
21brightcopy
19> Right. That was one of Tim's examples in the expressions thread (first post):
Expressions will also make it possible to differentiate between Margaret Wise Brown's Christmas in the Barn as illustrated by Barbara Cooney from that illustrated by Diane Goode.
Another point is that different expressions may overlap. So one expression might be "Beowulf translated into English". Another might be "Illustrated versions of Brown's Christmas in the Barn". It's all different ways to slice the same pie.
Expressions will also make it possible to differentiate between Margaret Wise Brown's Christmas in the Barn as illustrated by Barbara Cooney from that illustrated by Diane Goode.
Another point is that different expressions may overlap. So one expression might be "Beowulf translated into English". Another might be "Illustrated versions of Brown's Christmas in the Barn". It's all different ways to slice the same pie.
22_Zoe_
Returning to the earlier point about what to do with relationships, and particularly the "I don't see why the data shouldn't be used if it's there" idea, there's always a computational cost. Processing power isn't unlimited, and I'd rather see it used for things like Tag Mirror that we don't currently have at all, instead of just rearranging a currently-existing list to reflect nuances slightly better.
If there were no limitations on processors or development time, then by all means, relationships could be taken into account for everything. But given the trade-offs, I just don't see most of this as a high priority.
I just really want to stop seeing recommendations for things I already have cataloged
Yes, this is exactly what I meant. High value at relatively little cost.
If there were no limitations on processors or development time, then by all means, relationships could be taken into account for everything. But given the trade-offs, I just don't see most of this as a high priority.
I just really want to stop seeing recommendations for things I already have cataloged
Yes, this is exactly what I meant. High value at relatively little cost.
23brightcopy
22> Plus there's always the cost of whatever other ponies don't get fed because this feature ate all their food. :D
