An Improved Import Page
DiscussieRecommend Site Improvements
Sluit je aan bij LibraryThing om te posten.
1JoeB1934
I have been doing a lot of imports recently and I find the page reporting on progress inadequate. The main problem is that books being search for are identified only by an ISBN.
It would be very useful to provide the book title and author that goes with the ISBN.
A second valuable change would be to maintain the import screen that set up the process and simply to report details on the progress appended to the bottom of the screen. This step would keep in view all of the settings made during the initiation of the import. If the viewer could keep that information in mind, it could provide an understanding of what choices were made and why the import might be in trouble.
The viewer could initiate a refresh on the page periodically to monitor progress.
Finally, and I realize this might be difficult to implement, I often have wanted a way to mark a book record with a tag to indicate that it was just imported and could be identified as such.
It would be very useful to provide the book title and author that goes with the ISBN.
A second valuable change would be to maintain the import screen that set up the process and simply to report details on the progress appended to the bottom of the screen. This step would keep in view all of the settings made during the initiation of the import. If the viewer could keep that information in mind, it could provide an understanding of what choices were made and why the import might be in trouble.
The viewer could initiate a refresh on the page periodically to monitor progress.
Finally, and I realize this might be difficult to implement, I often have wanted a way to mark a book record with a tag to indicate that it was just imported and could be identified as such.
2lorax
have been doing a lot of imports recently and I find the page reporting on progress inadequate. The main problem is that books being search for are identified only by an ISBN.
It would be very useful to provide the book title and author that goes with the ISBN.
That's how the import works - if you provide an ISBN, it throws away the title and author you have provided in the file, and includes only whatever comes from your source with your chosen ISBN. If you care about your data, either don't use the bulk import and instead add books individually from your chosen source, or delete the ISBN (keep the column there and empty) and just add it in after the fact, in which case your file will import exactly as-is.
I often have wanted a way to mark a book record with a tag to indicate that it was just imported and could be identified as such.
Can't you just add this to the import file prior to uploading?
I mean, the bulk import process is broken to the point of uselessness anyway, so I'd rather see it rebuilt from the ground up than tweaked.
It would be very useful to provide the book title and author that goes with the ISBN.
That's how the import works - if you provide an ISBN, it throws away the title and author you have provided in the file, and includes only whatever comes from your source with your chosen ISBN. If you care about your data, either don't use the bulk import and instead add books individually from your chosen source, or delete the ISBN (keep the column there and empty) and just add it in after the fact, in which case your file will import exactly as-is.
I often have wanted a way to mark a book record with a tag to indicate that it was just imported and could be identified as such.
Can't you just add this to the import file prior to uploading?
I mean, the bulk import process is broken to the point of uselessness anyway, so I'd rather see it rebuilt from the ground up than tweaked.
3JoeB1934
>2 lorax: lorax
Good to see you here. Please look at Bug Collectors under how to import ISBN13 for my complete display of how badly bulk import is broken. I just put this here hoping to get anything out of this problem.
Is this obsession with 10 char ISBN a legacy software limitation?
My point in the bug collectors diatribe is that Goodreads works well for this problem and LT can't even use their output to solve the LT problem.
I called the import screen display pathetic but the same could be said about the ISBN import problem.
This is not a standard that LibraryThing users should accept and just do what you suggest of taking on the task of manually entering ISBNs.
Why can't they insert an ISBN for whatever book shows up after the blank ISBN process finds a book?
I don't own many books and all that counts to me is the title and author. But I don't like blank ISBNs around. GR, of course won't even accept such a thing.
I was a software developer for decades and to me the user must always be satisfied whatever the software engineers decided to do.
Good to see you here. Please look at Bug Collectors under how to import ISBN13 for my complete display of how badly bulk import is broken. I just put this here hoping to get anything out of this problem.
Is this obsession with 10 char ISBN a legacy software limitation?
My point in the bug collectors diatribe is that Goodreads works well for this problem and LT can't even use their output to solve the LT problem.
I called the import screen display pathetic but the same could be said about the ISBN import problem.
This is not a standard that LibraryThing users should accept and just do what you suggest of taking on the task of manually entering ISBNs.
Why can't they insert an ISBN for whatever book shows up after the blank ISBN process finds a book?
I don't own many books and all that counts to me is the title and author. But I don't like blank ISBNs around. GR, of course won't even accept such a thing.
I was a software developer for decades and to me the user must always be satisfied whatever the software engineers decided to do.
4shadrach_anki
>3 JoeB1934: Why can't they insert an ISBN for whatever book shows up after the blank ISBN process finds a book?
They aren't going to insert an ISBN essentially at random during the import process, as any given work will potentially have multiple ISBNs associated with it. An ISBN is edition-specific, and thus correlates to particular other pieces of data like publisher, page count, format, etc. Randomly choosing one of the ISBNs associated with a particular work upon import is...sub-optimal from a cataloging perspective. There would be no guarantee that it would accurately match the rest of the data entered.
>3 JoeB1934: But I don't like blank ISBNs around. GR, of course won't even accept such a thing.
You may not like having a blank ISBN field, but there are plenty of books that simply do not have ISBNs associated with them in the first place. Older titles from before the days of ISBNs; special book club editions; proposed but still unwritten works in a series; some self-published works; and things like Kindle books and Audible audio books all fall into this category.
I will note that your statement about GR is incorrect. The only required pieces of information for a book record on that site are title and author. Everything else is technically optional, and there are plenty of records that have no ISBNs on them. Things like Kindle books and Audible audio books will be entered with an ASIN, not an ISBN (and ASINs are more prone to changing than ISBNs).
They aren't going to insert an ISBN essentially at random during the import process, as any given work will potentially have multiple ISBNs associated with it. An ISBN is edition-specific, and thus correlates to particular other pieces of data like publisher, page count, format, etc. Randomly choosing one of the ISBNs associated with a particular work upon import is...sub-optimal from a cataloging perspective. There would be no guarantee that it would accurately match the rest of the data entered.
>3 JoeB1934: But I don't like blank ISBNs around. GR, of course won't even accept such a thing.
You may not like having a blank ISBN field, but there are plenty of books that simply do not have ISBNs associated with them in the first place. Older titles from before the days of ISBNs; special book club editions; proposed but still unwritten works in a series; some self-published works; and things like Kindle books and Audible audio books all fall into this category.
I will note that your statement about GR is incorrect. The only required pieces of information for a book record on that site are title and author. Everything else is technically optional, and there are plenty of records that have no ISBNs on them. Things like Kindle books and Audible audio books will be entered with an ASIN, not an ISBN (and ASINs are more prone to changing than ISBNs).
5AnnieMod
>3 JoeB1934: "But I don't like blank ISBNs around. GR, of course won't even accept such a thing."
https://www.goodreads.com/book/show/22934429-solaris - No? :) Not all books have ISBNs. ISBNs and their older sibling SBNs are relatively new invention. Old books simply never had one. Modern books from small (and not small) presses often don't have ISBNs either - Folio Society and Easton Press for example never use ISBNs.
Here is a modern book in GR with no ISBN: https://www.goodreads.com/book/show/60033196-david-copperfield
Or magazines: https://www.goodreads.com/book/show/58866729-ellery-queen-s-mystery-magazine-sep...
You can complain about a lot of things if so you wish but claiming that GR won't accept books with no ISBNs is just incorrect. :)
What GR has and we don't have is an edition layer - a single entry for all books which are the 1962 edition of that specific book for example. So when you pick the edition, you get the edition's ISBNs. LT has only 2 layers - individual books and works (all languages, all editions). When you add a book, it has no idea if it matches something it already has because we just do not have the middle layer.
>3 JoeB1934: "Why can't they insert an ISBN for whatever book shows up after the blank ISBN process finds a book?"
Take for example a book such as 1984. Unless you tell it somehow, how is LT to even guess what ISBN to pick? The Swedish 2022 edition one? The Russian one? One of the many English ones? How would it guess what language your book is in if there are multiple languages in the work (with the same title) or which ISBN (if there are multiple ISBNs)? So it does not try to guess.
https://www.goodreads.com/book/show/22934429-solaris - No? :) Not all books have ISBNs. ISBNs and their older sibling SBNs are relatively new invention. Old books simply never had one. Modern books from small (and not small) presses often don't have ISBNs either - Folio Society and Easton Press for example never use ISBNs.
Here is a modern book in GR with no ISBN: https://www.goodreads.com/book/show/60033196-david-copperfield
Or magazines: https://www.goodreads.com/book/show/58866729-ellery-queen-s-mystery-magazine-sep...
You can complain about a lot of things if so you wish but claiming that GR won't accept books with no ISBNs is just incorrect. :)
What GR has and we don't have is an edition layer - a single entry for all books which are the 1962 edition of that specific book for example. So when you pick the edition, you get the edition's ISBNs. LT has only 2 layers - individual books and works (all languages, all editions). When you add a book, it has no idea if it matches something it already has because we just do not have the middle layer.
>3 JoeB1934: "Why can't they insert an ISBN for whatever book shows up after the blank ISBN process finds a book?"
Take for example a book such as 1984. Unless you tell it somehow, how is LT to even guess what ISBN to pick? The Swedish 2022 edition one? The Russian one? One of the many English ones? How would it guess what language your book is in if there are multiple languages in the work (with the same title) or which ISBN (if there are multiple ISBNs)? So it does not try to guess.
6Felagund
>3 JoeB1934:
Many of us agree that LT import is not satisfying, and some have been asking for improvements for a loooooong time (LT staff, if you are reading this: https://www.librarything.com/topic/219575#n7757999 please please please).
Regardless of our opinions and wishes, it works as described here:
https://wiki.librarything.com/index.php/Adding_and_importing_books#How_does_the_...
If you care a lot about your bibliographic data, it feels better to use a spreadsheet with blank ISBNs (respecting all the exact columns and headers in the example) and then add ISBNs manually. This is unfortunate, and not intuitive at all, but pathetic feels a bit strong.
If I understood your posts here and there https://www.librarything.com/topic/342032#n7851427 , you are concerned about 13 vs. 10 digits ISBNs. I don't think it has anything to do with which books are found or not during the import: 13-digit ISBNs starting with 978 can be safely converted into their 10-digit equivalent and/or back (see https://isbn-information.com/convert-isbn-10-to-isbn-13.html if you are interested). So it really doesn't matter, unless the ISBN starts with another prefix - I have a few 979 ISBNs in my collection, and I assure you that they are proprly stored in the LT database. As far as I know, what matters for the import is this: if the systems sees an ISBN, it will try to find this ISBN (13 digits or the equivalent in 10-digit format if applicable) in one of the source databases (library catalogues, Amazon...) that the user has selected. If one ISBN is not known to these sources, the book will be skipped - which is probably the safe thing to do.
I hope this answers some of your questions?
Many of us agree that LT import is not satisfying, and some have been asking for improvements for a loooooong time (LT staff, if you are reading this: https://www.librarything.com/topic/219575#n7757999 please please please).
Regardless of our opinions and wishes, it works as described here:
https://wiki.librarything.com/index.php/Adding_and_importing_books#How_does_the_...
If you care a lot about your bibliographic data, it feels better to use a spreadsheet with blank ISBNs (respecting all the exact columns and headers in the example) and then add ISBNs manually. This is unfortunate, and not intuitive at all, but pathetic feels a bit strong.
If I understood your posts here and there https://www.librarything.com/topic/342032#n7851427 , you are concerned about 13 vs. 10 digits ISBNs. I don't think it has anything to do with which books are found or not during the import: 13-digit ISBNs starting with 978 can be safely converted into their 10-digit equivalent and/or back (see https://isbn-information.com/convert-isbn-10-to-isbn-13.html if you are interested). So it really doesn't matter, unless the ISBN starts with another prefix - I have a few 979 ISBNs in my collection, and I assure you that they are proprly stored in the LT database. As far as I know, what matters for the import is this: if the systems sees an ISBN, it will try to find this ISBN (13 digits or the equivalent in 10-digit format if applicable) in one of the source databases (library catalogues, Amazon...) that the user has selected. If one ISBN is not known to these sources, the book will be skipped - which is probably the safe thing to do.
I hope this answers some of your questions?
7JoeB1934
My statement about Goodreads (GR) not accepting blank ISBN applies to their import file only. They say:
File Import Tips
All columns in our sample file are supported, but only ISBN is required.
The discussion about 10 char ISBN is very interesting, but the fact remains that LT was unable to find them for any of my Libby imports. Of great significance to me is that GR found ALL of the Libby books and they only use Amazon for their searches.
Even more frustrating to me is that when I provided the GR export to LT the finding of the books was very poor (see my discussion of these results in Bug Collectors). Why is it so hard to find the books when given the relevant 10 and 13 char ISBN's?
Another question pertains to my original statement in this RSI. Why not improve the import page communication to the users, whatever the process LT is using?
I wouldn't be bringing up these issues if I didn't greatly appreciate LT. My main benefit from using the system revolves around the Talk features and all of the options for dialog about books, searching for them and our reactions to them. These statements of mine are in the same spirit.
Finally, I want to reference the statement 'Felagund' made:
"Many of us agree that LT import is not satisfying, and some have been asking for improvements for a loooooong time (LT staff, if you are reading this: https://www.librarything.com/topic/219575#n7757999 please please please)."
That statement being necessary is an incredible indictment of what seems to be a one-way discussion. Any world class software system like LibraryThing would have a two-way dialog with users and developers where answers would be given.
How in the world can it be necessary to ask "(LT staff, if you are reading this:"
File Import Tips
All columns in our sample file are supported, but only ISBN is required.
The discussion about 10 char ISBN is very interesting, but the fact remains that LT was unable to find them for any of my Libby imports. Of great significance to me is that GR found ALL of the Libby books and they only use Amazon for their searches.
Even more frustrating to me is that when I provided the GR export to LT the finding of the books was very poor (see my discussion of these results in Bug Collectors). Why is it so hard to find the books when given the relevant 10 and 13 char ISBN's?
Another question pertains to my original statement in this RSI. Why not improve the import page communication to the users, whatever the process LT is using?
I wouldn't be bringing up these issues if I didn't greatly appreciate LT. My main benefit from using the system revolves around the Talk features and all of the options for dialog about books, searching for them and our reactions to them. These statements of mine are in the same spirit.
Finally, I want to reference the statement 'Felagund' made:
"Many of us agree that LT import is not satisfying, and some have been asking for improvements for a loooooong time (LT staff, if you are reading this: https://www.librarything.com/topic/219575#n7757999 please please please)."
That statement being necessary is an incredible indictment of what seems to be a one-way discussion. Any world class software system like LibraryThing would have a two-way dialog with users and developers where answers would be given.
How in the world can it be necessary to ask "(LT staff, if you are reading this:"
8kristilabrie
>7 JoeB1934: With a little digging you can see that staff, including developers, have posted numerous times in response to the multiple requests for a closed import-export loop (I myself am in strong support of this but am not a developer and don't know the technical hangups for such an improvement). I've also replied to your thread in the Bug Collectors Group addressing your recent import attempts and explanation of why they failed where they did. Please let me know if you have further questions!
