Noel has applied it to the case of the chemistry retractions in ACS publications. When I tried to follow up ChemBark’s story I had DOIs but no links. With the BO greasemonkey these DOIs get translated to links which I can follow. In this way Noel has constructed the complete set of linked correction/retractions – there appear to have been at least 6 papers which have been affected. I think you can see from this that the blogosphere pays a positive role in helping post-hoc peer-review in chemistry.
Noeal highlights the fact that anyone reading the original paper does not know that a retraction has been made. Obviously this is not possible in paper journals, but I would have expected a publisher to put up a note saying “this paper has been retracted/corrected”. I am now thoroughly confused as to what I am seeing at the end of a DOI. The fundamental questions are:
- Is a DOI and identifier to a static piece of information (which is what I would expect – as it stands for Digital Object Identifier) or
- Is a DOI a controlled addressing system managed by a purchaser of DOIs. IOW can a purchaser put different versions of the same information under the same identifier
If it is the former, then we have some chance of preserving the scientific record. If the latter, the purchaser of the DOI can rewrite history whenever they choose.
Of course that is true of much web information. For example I sometimes edit a post on this blog after the first publication. Mainly to correct typos or add clarification. But, unlike a publisher, I cannot change the record. When this post is published it gets aggregated by Planet BlueObelisk which effectively acts as a record. (It’s actually because certain typos and styles foul up the Planet that I edit the post later). The posts are also captured by Technorati – and are probably cached in many other places. And these caches have timestamps.
Admittedly this leads to a potentially confused set of versions but it is possible to reconstruct history. Unless publishing is made Open, history will be mutable.