Jan Aerts
Nice! We need to keep pushing and it will happen. - Björn Brembs
Nicely done. I like dsi, but professions other than scientists will probably want to use the same service... dai, digital author identifier? dhi, digital human identifier? dimbl, digital identifier for a meat based lifeform? - Bill Hooker
This isn't the first (or last...) time someone proposes a centralized author identification scheme. I'm not too optimistic about this happening. Outside of scientific publications this might not even be a good idea (hello big brother). Another approach would be to set up services that collect information about the relationship between personal URIs such as web site URLs, OpenID URLs and e-mail addresses. There's even a standard for publishing such data: http://www.foaf-project.org/ - Eric Jain
Is it about the service, about its adoption by publishers or about tracking contributions which aren't publications? Sorry if that's a silly question, but I'm somehow lost what the whole discussion concerning unique researcher ID is really about. - Pawel Szczesny
@Pawel: The problem these proposals are addressing is that right now there is no reliable way to get a list of publications for a person -- especially if that person has a common name. I've even come across publications that had two authors with the exact same name! - Eric Jain
*sigh* Why does it always seem that the system that becomes the standard isn't necessarily the best, but the one is easiest to implement? - Mr. Gunn
Eric, thanks. But it doesn't going to work without publishers' support, right? Or we just make the list all by ourselves, using a service Jan has described? I've read Jan's post, but also I try to understand the real principle behind such idea. Are we going to calculate personalized impact factor out of such lists? - Pawel Szczesny
@Pawel: Publishers would be in a good position to ensure that their author names (at least in future) are associated with some kind of identifier. Life would be simpler but I don't see a problem if this identifier is not a special and agreed upon author identifier but just an email address or any other URI (incl OpenID) or combination. If the publishers don't do this then the publication to author mapping can be (and in fact has been) done elsewhere. But again, there is no need for a single, central server. - Eric Jain
I guess that what I describe in the post is more about a whole environment than a single thing. First of all there must be some way for researcher not interested at all in OpenID/Science2.0/... to get a unique identifier before they submit a paper. As that ID should not change it should not be the email address, and they will not want to go out on their own to find out how to get an OpenID. All the rest follows from that, basically. We can build such a system ourselves, as long as we have a way of ... - Jan Aerts
WRT sciD and IF: of course one could attach all kinds of info to a unique ID, preferably so many that computing a single factor would be absurd. See discussion at: http://friendfeed.com/e... and my summary http://bjoern.brembs.net/news... - Björn Brembs
...identifying everyone. - Jan Aerts
@Pawel: The reason why I'm interested in this and wrote the post is that (a) I've got papers with my obsolete contact details on, and (b) I'm tired of explaining to employers that I do other stuff than just write papers. (a) would be solved by the a unique ID and (b) by some way of contribution-tracking (active or passive). I am not interested in personalized impact factor: your impact can be huge in the specific area you work in but become minuscule if you look at science in general. - Jan Aerts
Hope I die before I get old ... - Henry Gee
Meet the new plos. Same as the old ... er .... where was I? - Henry Gee
I agree with the 'Big Brother' concern above - I don't see why the publication-focused author identifier necessarily has to be used for tracking of all-and-sundry information about a person. However, the point was raised on the other FF thread recently that the author/publication part of things really does need to be centralized for it to work (i.e. the CrossRef/DOI model). That's where things are headed anyway, which I think is fine. - 'Mummi' Thorisson
...but if people want to link up their (future) 'microcredit' history for e.g. db-submisssions and whatnot with their publication profile, they should free to do so by. This could then be done via the person's master online identity (i.e. OpenID) to aggregate info from various sources. Main message: the user should be in charge of what gets aggregated and what doesn't. - 'Mummi' Thorisson
Yes, the idea of linking your researcher id to some main id like an openid is a good one, but I reiterate what was said in a previous thread on the matter - the ability to selectively hide and show various aspects of your online personality would be very important. - Allyson Lister
Allyson - I totally agree, and that's one of the many cool things that OpenID promises: you can have multiple online personae, or profiles, associated with a single OpenID. E.g. work vs blogging vs shopping etc. Also, I think that many will find they in fact want completely seperate identities for certain things, ergo multiple OpenIDs. - 'Mummi' Thorisson
Gudmundur - can you have multiple online profiles right now with OpenId? I have one, but didn't know you could do that... Or is that just one of the "promises"? ;) - Allyson Lister
Allyson - depends on the OpenID provider: I think the major ones like MyOpenID and ClaimID and Verisign offer some sort of profile management (I use https://www.myopenid.com myself currently). Dunno about Google and Yahoo and others who not-so-geeky people are like to use to get their feet wet with OpenID (coz they already have a Google or Yahoo account). - 'Mummi' Thorisson
Gudmundur - thanks. I actually ended up with two open ids: first, I had one from myopenid (and I still do), but haven't used it in a year because my blog provider, Vox, made all of their blog urls open ids. this is great, and what I use most of the time, but I now realized looking back at my myopenid account that it is a very, very basic service. Thanks for the info! This leads to the question: can we link open ids to each other? :) - Allyson Lister
Allyson - that's the bit I don't know enough about. I agree that this would be a very useful thing to do, i.e. say that these 3x OpenIDs are really proxies for a single primary one. OpenID delegation is one way of doing this (see e.g. http://simonwillison.net/2006...), but then we're into geek-territory in terms of configuring the HTML header and whatnot. And not all relaying parties will accept delegated OpenIDs either, I believe.. - 'Mummi' Thorisson
Email addresses may change, but so do identifier systems over time... The previously mentioned FOAF standard allows you to claim ownership of multiple email addresses [http://xmlns.com/foaf...] (without revealing the addresses to spammers) and associate them with OpenIDs [http://xmlns.com/foaf...] or other information such as your DNA checksum [http://xmlns.com/foaf...] :-) - Eric Jain
If someone manages to set up a site that is targeted towards scientists and gets them to state a lot of previously unknown or ambiguous relationships between publications and email addresses and other information (and expose this information in a machine-understandable format) that would be great. But I don't think a new grand unified identifier system or any science-specific technologies are required for this. - Eric Jain
Eric is making a lot of sense here. The more I think about it, the more I like the idea of a distributed system. Since we already have openID and FOAF, how about a campaign to get scientists to use them as contact details on publications, and to get journals to encourage this? That might have more immediate benefit than trying to (in some sense) re-invent the ID wheel. - Bill Hooker
I like the thought of FOAF for sharing researcher contact details. I wonder if the in-the-works CrossReg service (see related FF discussion at http://friendfeed.com/e...) could do FOAF-RDF exports. I don't know much about FOAF - is there an author/publication (or more generic contributor/content thing) in FOAF, or could other schemas (Dublin Core, SKOS) be brought in to support this? - 'Mummi' Thorisson
..[forgot this, sorry] is there a way to somehow merge multiple FOAF profiles, say hypothetically my author profile from in-the-works CrossReg and my personal profile from mummirox.com/profile? - 'Mummi' Thorisson
@Gudmundur FOAF is 'just' a RDF file about a person ( a list of statements = subject/predicate/object). You can generate a basic FOAF file from many tools ( http://tinyurl.com/638buj , even from FF : http://friendfeed.com/mummi... ). A foaf could also be created from your publication in pubmed using a xslt stylesheet (I wrote one here: http://tinyurl.com/5gl4wo ... but it would be better with a... unique ID). I also wrote a tool for building a FOAF profile here: http://tinyurl.com/5pwujr - Pierre Lindenbaum
if you want to merge two FOAF files you need a... unique identifier... :-) the properties foaf:mbox or foaf:mbox_sha1sum could be used to merge two profiles. - Pierre Lindenbaum
Can also use any other attribute marked as an InverseFunctionalProperty at http://xmlns.com/foaf/spec/ (incl foaf:openid) or in any other RDF-based vocabulary for merging. The neat thing is that not everyone needs to agree on using the same kind of unique identifier(s). The challenge of course is collecting the identifiers and getting them mapped. Services such as FF that collect (and make available) such data as a side effect look like a promising approach... - Eric Jain
I like the way this discussion is going. If I were to submit a paper in the next couple weeks(and I am), what do you recommend I include and how do I do it? - Mr. Gunn