A place for your own papers

[fgdorais]

Adding my own papers to my reading list doesn't sound quite right. The reading list sounds like temporary storage where the goal is to read the papers and discarding them from the list. I generally don't plan on reading my own papers and I also don't plan on discarding them.

It would be great if I could associate my arXiv author id to my spnetwork profile in some way or another.

[cjlee112]

This is definitely needed. The reason we avoided it so far was the lack of mechanisms for validating a user's authorship claim. I couldn't find an obvious way for arXiv to do this. That is, even if a user has an arXiv author identifier (and it looks like a large fraction of arXiv users haven't bothered), how can a selectedpapers.net user prove that they "own" that author identifier? Of course, a user can give us an arXiv author identifier, and we can retrieve their list of papers from arXiv's atom API. But proving that they own the author identifier (without giving us their arXiv password) would require that arXiv provide some kind of OAuth (or Persona) style login process. It would be great if we could link up with people at arXiv to discuss such possibilities.

Without that, all we could do is let users claim authorship, with a stern submission form warning ("I have reviewed the identity of this manuscript and affirm, on pain of banishment, that I am the following listed author of this manuscript (choose one of the author names)"). Then you provide the community with ways to report apparently dishonest claims.

What do people think?

[fgdorais]

I'm not too concerned about people claiming papers they didn't write. Academia has a long tradition of working on a honor system so I think this would be very rare, so rare that it wouldn't be a bother to deal with it on a case-by-case basis. I would contact an arXiv insider to see whether they could help with this kind of authentication mechanism.

[cjlee112]

Am I right in thinking a large fraction of arXiv users have not bothered to create Author Identifiers (and also prove to arXiv that they are an author on each of the papers on which they are a co-author using the emailed code from the original submitter)?

[jdchristensen]

Is there a way for you to get the arxiv author's e-mail address, and send a message to them containing a link that when clicked will link the arxiv account with the selectedpapers.net account?

[cjlee112]

hmm, I think arXiv tries hard to protect email addresses, and does not want robots requesting them (to prevent spam). I guess we could ask arXiv whether some standard procedure like Persona could be set up, to enable arXiv users to authenticate their identity to us without revealing their password.

[cjlee112]

interface proposal: for any paper, the list of paper authors will be hyperlinks that take you either to that person's page (if an spnet user has staked a claim to be that author), or else a simple page that gives you a few options, e.g.:

• claim that you are this author and this is your paper (with a warning that dishonest claims will be banned);
• send invitation to the author (if you know them) to claim their authorship on this paper (and thereby get alerts for any discussion about it that occurs).
• search for other papers by this author
• anything else?

Please comment.

[bkloeckn]

I would like to be able to get notification of new claimed papers by authors I follow. This could help getting new papers some echo, with people marking as interesting some of the papers disclosed to them that way, which should trigger a notification for other people following them, and so on.

But as I understand things, it would mean that there should be a "following" status that should be transversal to G+, twitter, blogs, etc. which sounds interesting to me but may be orthogonal to the spirit of SPNet.

Alternatively, authorship claims could trigger a proposal to write a post on one's social network, possibly with an already pre-written post including identification of the paper, #spnetwork tag and a tag corresponding to the arXiv subject classification of the paper.

[cjlee112]

@bkloeckn getting notifications of new papers by authors you're subscribed to is definitely part of the spnet design. We just didn't have any easy way to get from arXiv that author to userID mapping. But if we provide an interface for people to claim authorship on their papers, we can do what you suggested.

[ketch]

@cjlee112 That interface seems like a good start to me.

Also, we could use ORCID identifiers rather than arXiv IDs, looking ahead to the possibility of including commentary on papers outside the arxiv. ORCID allows one to connect other accounts (like LinkedIn and Scopus) to it from their end. See http://orcid.org/.

[cjlee112]

ORCID looks like the way to go for identifying authors long-term. But I suspect their database is mostly an empty box at this point. We should still use it, but I suspect it won't help us identify a lot of arXiv authors.

[cjlee112]

I suggest two interfaces (both on the honor system, since we have no way to verify identities in a deep way):

1. author list for any paper should be hyperlinked, either to the record for an author ID if known, or to a form where a user has the option to claim to be that author.
2. user can submit a list of paper IDs (same syntax as for posts, e.g. arXiv:1234.5678) that s/he has authored, and system will get those records and automatically match author by name / initials if possible. If a paper author can't be unambiguously matched, report that exception with a link redirecting the user to claim authorship via method 1 above.