auto-linking names
David Remsen
dremsen at MBL.EDU
Tue Nov 15 09:42:27 CST 2005
LinkIT reads a URL and identifies any scientific names on the page.
It then matches these names against indexes of authority lists or
collections we have mapped and, if the name exists in those indexes,
creates a URL to that site. Point it at a publication list of fish
articles for example, and auto-link all fish names to FishBase. A
list of mushroom pictures can be immediately linked to Index
Fungorum. Some journals link species names to ITIS as a matter of
course without checking if the names are actually recorded in ITIS.
LinkIT only builds links when they exist in the source index.
http://namebank.ubio.org/tools/linkit.php
We built this originally to as a way to play around with our name
recognition algorithms but increasingly are exploring it as a way to
narrow the gap between expert systems and more generalized content
providers. Instead of manually mapping one name at a time one can
simply drop in a button. The original content and links are not
impacted and the linking out is a dynamic option. Multiple buttons
or a drop-down menu could link out to different content. One button
links out to Fishbase for example, while another rebuilds the page
linked to NCBI or CBOL or TreeBase. Imagine an Index Fungorum button
that fungal authors can drop into their publication list and
instantly point to IF or a GBIF button that selectively adds a link
to any name for which GBIF has information.
An additional output parameter takes the array of names and displays
it's representative ITIS classification within a division. We will
add a Species 2000 display as well as we explore various interface
issues.
Perhaps even more interesting (at least to us) is how many names
exist in content but are NOT included within these collections. Our
plan is to continue to develop the application and try to deploy it
within library applications.
Any feedback would be appreciated. Yes, there are limitations.
Homonyms remain problematic although our name-recognition algorithm
does locate author names and can disambiguate some. Sites that
include external javascript files that reference images may not
display correctly.
We would also love to add additional indices to link out to. It
makes more sense to pass them on to expert content rather than our
provisional namebank record.
Thank you for your time,
David Remsen
_______________________________________________
David Remsen
uBio Project Manager
Marine Biological Laboratory
Woods Hole, MA 02543
508-289-7632
More information about the Taxacom
mailing list