Some Bibliography links to PDF files seem to be rewritten by target server

[MattHeffron]

Some of the URL links in the Bibliography behave differently if clicked in the browser as opposed to pasting the same URL into the browser. Specifically, the Investigations into history tools for user support entry's link gets redirected to ResearchGate's parent page for this document when the link is just clicked on, but it opens the document directly when the URL of that link is pasted into the browser URL input window. (I'm assuming that this is the behavior for all of our links into ResearchGate.)

With the browser's (Chrome on Windows) Developer tools open on each of the pages for testing (the Bibliography, and a Chrome "New Tab"), both seem to send the same Request Headers (except for one cookie: _ga_4P31SJ70EJ). The clicked link returns "301 Moved Permanently" as the Status Code (and Response Headers Location of https://www.researchgate.net/publication/242625827_Investigations_into_history_tools_for_user_support) The pasted URL returns "200 OK" as the Status Code.

I cannot see how ResearchGate differentiates the two cases. In any case, although our Bibliography links to the correct document (as well as keeping a copy in Zotero), clicking the link from the Bibliography doesn't quite get to the document directly.

[MattHeffron]

I stumbled across this while trying to figure out how we could link to the documents stored in Zotero, on their server.

[masinter]

https://forums.zotero.org/discussion/49378/store-copyrighted-documents

You can't link to documents stored in Zotero -- access to file attachments online is only available in closed groups.

Storing items that you have accessed legitimately in Zotero is no copyright violation. Making them available to people who don't otherwise have access to them via a private group may be a violation of the terms by which you did access them, though (albeit one that's unlikely to cause you any problems).

[MattHeffron]

@masinter If we can't link to to documents that are stored in Zotero, then why are we considering Zotero to be our primary storage of such documents? Shouldn't we put our copies of documents (which we can legitimately publish) in a place that we can link to?

This is secondary to the original issue that the links behave differently when clicked, or copy-paste the URLs.

[masinter]

I think there may have been a misunderstanding. I think Zotero should not be our 'primary' storage for anything. They don't offer redundant global storage and unlimited access bandwidth. I think the Zotero server seems like a reasonable tool for building and maintaning the document metadata, altough I wish we could use git-based tools to back up, manage and version-control.

For long-term archive requirements I'll suggest:

https://larrymasinter.net/0603-archiving.pdf

is still valid.

[masinter]

was there a problem, or is this just a puzzle? We should have a copy but not serve it if it is reliably and quickly available on the net elsewhere .

[MattHeffron]

This issue evolved into discussing two issues:

1. Link to document on ResearchGate behaves differently if URL is entered into browser, vs clicked on from Bibliography. (This is with the identical URL.)
2. Where do we store documents referenced in the Bibliography that don't (reliably) exist elsewhere? I'll copy the comments related to that question to a new issue (#2098).