PubMed Data Providers
Donat
required by the submitting publication metada and abstracts to PubMed
will be also included in citebank, GNA or other initiatives that
include bibliographic metadata.
http://www.ncbi.nlm.nih.gov/bookshelf/br.fcgi?book=helppubmed&part=publisherhelp
I could foresee a future where taxonomic publishers submit at least
their abstracts and metadata to PubMed, if not the entire publications
to MedLine. This would not only be an independent archive but also a
very strong link to other fields in genomics and beyond.
Donat
Richard Pyle
Thanks for the link, Donat. I'll have a look as soon as I return from Fiji.
Weirdly, my internet connection allows me to send/receive email through MS
Outlook, but I can't seem to get any web pages (opposite of what ususally
happens on these remote internet connections).
Very briefly, it seems there will be two components to "CiteBank" (not sure
which one gets the name, if not both). One component is what we've been
calling the "Dirty Bucket", which is analogous to GNI in that it will be a
repository for "text strings purported to represent citations". These,
then, will be used to develop reconciliation algorithms to help anchor them
to the "Clean Bucket", which is analogous to GNUB (well maintained and
curated citation record objects). After a Skype session with Chris F. at
BHL during our recent GNUB workshop, we decided that BHL would move forward
with the "dirty bucket" implementation, and we went ahead with the data
model for the "Clean Bucket".
I'll provide details on the latter when I get home; but basically we decided
on a very open-ended data structure, that allows one to define an unlimited
number of attributes ("fields") that can be applied to citations. In a
nutshell:
- A "Reference" object can be anything from a treatment to an article, book
chapter, book, correspondence, book series, journal, etc., etc. (there may
be an unlimited number of reference types, and this community will need to
define a set that we will use for our purposes.
- References are hierarchical, meaning that a child can inherit properties
from any member of the parent "pedigree" (parent, grandparent, great
grandparent, etc.).
- There is a dedicated structure for linking authors to agents (we also came
up with a simple Agent data model), which will allow discovering
publications by the same author, using various versions of their own name.
- There is a dedicated stricture for multiple titles, so that a Reference
can have multiple forms of the title (full title, standard title, translated
title, short title, abbreviated title, etc., etc.)
- All other attributes are stored in a structure modelled after what Jerry
Cooper has been using, which is sort of analogous to a triple-store. That
is, values are stored in a tall, thin table, where each record says
effectuvely "This Reference[ID] has this type of Field[ID] with this value".
Again, our community is going to need to define the various FieldType
values, and their mappings to ReferenceType values. But the point is that
the kinds of attributes that can be attached to individual citation records
can be expanded as needed to accomdate things like the PubMed metadata, or
any other standard set of metadata.
Time to go diving with some really big bull sharks....
Aloha,
Rich
> --
> You received this message because you are subscribed to the
> Google Groups "Taxonomic Literature" group.
> To post to this group, send email to taxo...@googlegroups.com.
> To unsubscribe from this group, send email to
> taxonlit+u...@googlegroups.com.
> For more options, visit this group at
> http://groups.google.com/group/taxonlit?hl=en.
>