Following on from a conversation with @dshorthouse http://iphylo.blogspot.co.uk/2017/02/new-feature-for-biostor-extracting.html we could do the following. Imagine we have an article with a list of literature cited (generated by a CouchDB view):
For each list of literature cited we currently have a clickable lin that does a search for the reference and updated the DOM. The results of this search aren't stored. Imagine that we use a webhook to store the BioStor id (if found). We could also expand this to DOIs (via CrossRef), etc.
One potential problem is that the list of literature cited may change if the code for the CouchDB view extracts the citations is chaned, so how do we ensure identifiers are linked to the citations? We could store the citation-identifier link as an annotation (a la hypothesis.is) and so use a bunch of anchors (order in list, before, middle, after strings, etc.) and use those to help match the two together. Then we can imagine storing the citation-identifier link in CouchDB as an annotation object, and could have a view which generates the literature citation graph.
Following on from a conversation with @dshorthouse http://iphylo.blogspot.co.uk/2017/02/new-feature-for-biostor-extracting.html we could do the following. Imagine we have an article with a list of literature cited (generated by a CouchDB view):
For each list of literature cited we currently have a clickable lin that does a search for the reference and updated the DOM. The results of this search aren't stored. Imagine that we use a webhook to store the BioStor id (if found). We could also expand this to DOIs (via CrossRef), etc.
One potential problem is that the list of literature cited may change if the code for the CouchDB view extracts the citations is chaned, so how do we ensure identifiers are linked to the citations? We could store the citation-identifier link as an annotation (a la hypothesis.is) and so use a bunch of anchors (order in list, before, middle, after strings, etc.) and use those to help match the two together. Then we can imagine storing the citation-identifier link in CouchDB as an annotation object, and could have a view which generates the literature citation graph.