Friday, May 23, 2008

BioOne (and/or CrossRef) sucks


<rant>
BioOne sucks. Really, really, sucks. I have lost count of the number of times they break DOIs. These are supposed to be the gold standard globally unique identifier, and BioOne continually buggers them. For example, take this URL:

http://www.bioone.org/perlserv/?request=get-abstract&doi=10.1600/02-14.1.

Note the doi=10.1600/02-14.1 bit at the end. If we go to the web page, we see this DOI displayed at the bottom of the page. Yet, when we resolve the DOI, we get the dreaded DOI Not Found error.

This.should.not.happen.

What is BioOne doing!? Now, it is possible that there's a problem with CrossRef, because Googling this paper I found it also lives on Ingenta with, wait for it, another DOI (doi:10.1600/036364404772973960).

This.should.not.happen.

BioOne and Ingenta are hosting the same paper, with different DOIs, only one of which is working. Will somebody please bang some heads together and sort this out!
</rant>

5 comments:

Charlie Rapple said...

Hi Roderic,
I agree that allocating more than one DOI to a paper sucks. It's usually happened by accident, maybe if the title has changed publisher or host. I don't know what's happened in this instance, but I do know how it can be resolved (the first DOI you cite can be deposited as an alias of the second DOI). I'll see if we have the necessary credentials to do that, or I'll get it passed on to the publisher to resolve if not.
All the best,

Charlie.
(on behalf of Ingenta).

Roderic Page said...

Charlie,

That would be great. There seems to be a major overlap in content between BioOne and Ingenta, especially for journals such as Systematic Botany. It's causing me major headaches as Google Scholar is serving the BioOne instances as the top hits, when these have broken DOIs (and don't get me started on what Allen Press are up to with their pseudo DOIs embedded in URLs).

Anonymous said...

The DOI displayed on the BIOOne page is not actually a DOI. I have no idea where BIOOne got it. The 'REAL' DOI is the one that points to Ingenta who appear to be the proper host for the society's content.

We have contacted BIOOne to have them address this (and other?) errors of this kind.

A search at http://www.crossref.org/guestquery/ does result in the one/correct DOI for this article.

It is unfortunate that Scholar refuses to take metadata from CrossRef with which they could certainly show the proper DOI (and article source) as the first search result.

Chuck (CrossRef)

Roderic Page said...

Chuck,

Thanks for following this up. It would be nice to get this resolved.

I'm a big fan of DOIs, but it seems some publishers are quite willing to publish DOIs that don't actually work. In some cases they don't seem to ever be registered, or there is a lag between when they appear online and when the DOI becomes live. For example, Here's a paper from Persoonia with a DOI proudly displayed in the PDF -- doi:10.3767/003158508X302212 -- and it doesn't resolve.

Anonymous said...

We've notified the 'Nationaal Herbarium Nederland', thanks for pointing out this problem.

Both of these problems are, unfortunately ,outside of our control since they don't involve real DOIs that we see come through the system.

Our only hope in these cases is to bring the issue to light for the publisher and get them to solve systemic problems that end up fixing a bunch of DOI errors in one shot. Of course there will always be one-off DOI problems that will forever be hard to find/fix.