repository field should contain only data on actual archives, not online databases

[erazlogo]

Translators such as Amazon and ProQuest put "Amazon" and "Proquest" in the "repository" field. That is incorrect--this field should contain data on archives, not online article/book databases. Because of this right now Chicago CSL has to exclude archival information for books and periodicals, making it impossible to cite, say, rare books and newspaper clippings located in archives.

[erazlogo]

Okay, I'm changing Chicago styles to always display the archival info if the fields are no empty. There is no logical reason for excluding archival info for books, and the odd translator behavior described above really needs to be fixed.

[erazlogo]

simon and sean--
it seems i could easily delete all references to the "repository" field in translators by doing a search of scrapers.sql in bbedit myself--are there any objections to this? I'll wait for a day or two and do this unless I hear from you.

[simon]

Most of this is implemented in translate.js, not in scrapers themselves, which can be easily reversed. However, the repository field is also used to store, for example, the library catalog from which an item was retrieved, without which the call number becomes less useful. Additionally, this would require a script to change existing items, since otherwise, old items would still have the translator name in the repository field.

[erazlogo]

Yes, originally when we added the repository field we thought it could contain both library catalog name and the name of archive. However, now it turns out that this setup doesn't work with CSL formatting for archival data.

When citing a newspaper article, there is no way to check whether the repository field contains a translator/database name (and should not be cited) or the name of the archive where the newspaper clipping is located (and has to be cited). This happens often enough to make the current setup a significant defect. For example:

Darrel V. Martin in [Pittsburgh?] Post-Gazette, May [1937], scrapbook #12, Gosden-Correll Collection, Cinema-Television Library and Archives of Performing Arts, Doheny Library, University of Southern California.

I see three options:

1. Perhaps the Library Catalog name could be put in the extra field from now on? Or it could be a separate field eventually. As far as old records are concerned, people could erase the translator name manually for now. It would be nice to have the script too--I seem to remember you were considering a similar script for the URL field to enter the correct URI information.

2. Create a preference similar to the URL citing preference, where archival information is cited for magazine/newspaper articles only when publication info is incomplete--article title, publication title, and/or date are missing.

3. Set cite.js to cite archival info for certain sources (book, book section, journal/magazine/newspaper article, thesis) only if "location in archive" field is not empty. This will account for most cases, although not all.

Let me know what you decide. If you don't intend to change this, I will change the Chicago styles back to omit repository data for published sources.

[erazlogo]

Looks like Chicago note styles are the only ones that display archival data. MLA at least should also include it. I can add this to relevant styles, but it would be great to resolve this ticket one way or the other before I do it--no point in adding complicated conditionals excluding books, etc., to other styles if we're going to remove them later.

​http://forums.zotero.org/discussion/2008/location-in-archive/

It seems important to provide adequate support for archival citations, particularly since Zotero is published by the Center for HISTORY and New Media.

[erazlogo]

Here is an example of an electronic book that is now impossible to cite correctly for the above reasons:

​http://forums.zotero.org/discussion/2123/electronic-text-fields/

[erazlogo]

Plan for a fix: For 1.5, keep the current field for translator/library data, and move archival collection info into a new field (call the new field "archive" and rename the old one "library catalog" or something). The easiest solution would be to assume that anything in the current "repository" field for things other than journals, magazines, newspapers, books, and theses are actually archives and move those into the new field. Then map the new field to CSL "archive" in cite.js.

[erazlogo]

Dan, Any chance this could be closed now rather than later? That would get rid of a workaround code in Chicago CSLs that is specific to Zotero. Or is there a reason to wait?

[dstillman]

The above is still the desired plan, including the migration of data in item types other than the ones you list?

[dstillman]

Furthermore, which item types should get this new field (or which shouldn't, if that's easier)?

[dstillman]

(To answer my previous question: the ones with Location in Archive, presumably.)

[erazlogo]

The above is still the desired plan, including the migration of data in item types other than the ones you list?

Yes

Furthermore, which item types should get this new field (or which shouldn't, if that's easier)? (To answer my previous question: the ones with Location in Archive, presumably.)

Yes, at least those. But then there is also this ticket which proposes to add Location in Archive to all item types: https://www.zotero.org/trac/ticket/688

Which may mean that all these fields (Call Number, Repository/Library Catalog, Location in Archive, and Archive) should be added to all item types.

[dstillman]

(In [5701]) Closes #853, repository field should contain only data on actual archives, not online databases

• Renamed 'Repository' to 'Library Catalog' and added 'Archive'
• Moved 'Repository' data to 'Archive' for item types other than books, periodicals, and theses
• Mapped 'Archive' to CSL 'archive'

[dstillman]

#688 won't happen now, but the rest is taken care of.