On 6/15/2011 5:43 PM, Peter Noerr wrote:
> And it is available - in our commercial software (not a plug - we don't sell it, just noting that it is not the sort of thing to try yourself on any scale - it takes a lot of resources).

I wouldn't go that far -- I _have_ done it myself, at the examples we're 
talking about in this thread! My implementation certainly isn't perfect, 
but it's 'good enough' that we're getting lots of value added from using 
it. ("My" implementation is software originally written by Ross Singer, 
5 years ago or more. It hasn't received 5 years worth of development, it 
hasn't gotten much development in the past 2-3 years, it's mostly just 
been continuing to work--although I've got plenty of ideas for 
improvements I haven't had time to do. But yes, it took a couple smart 
people a buncha time to implement, but we did it.)

But I agree it's tricky, because the devil is in the details, having 
done it myself too.

The Umlaut software is designed as a 'just in time'/'last mile' service 
provider, providing access points and services from local and remote 
platforms, for known items. You find an item of interest in some other 
interface, that has structured citation metadata, this is passed to 
Umlaut to provide those just-in-time last-mile services.

Umlaut is intentionally designed to make it as easy as possible to embed 
Umlaut discovered services on local interfaces, like the catalog for 
instance.  If you can add javascript to your catalog, you can figure out 
how to make the javascript obtain structured citation data from the 
current page (esp ISBN/ISSN/LCCN/OCLCnum), and you can write JQuery 
selectors to describe where each section of Umlaut content should be 
placed on the page (and maybe some CSS to style it appropriately) -- 
then you can have Umlaut content added via Javascript to your catalog 
item detail pages.

Umlaut is open source.  But Umlaut isn't as easy to install/configure as 
I'd like it to be and know it can be.  It is nevertheless currently in 
use not only at JHU (where I, the lead developer reside), but to one 
extent or another also at NYU and Vanderbilt.  But I've got plans to 
make it a lot easier, hopefully sometime in the next 6 months.  If this 
interests you, keep an eye out.  No matter how easy I make it, it'll 
probably never be Just Works Out of the Box, it's going to require a bit 
of local technical development to integrate it with all your local 
systems in the way you want.

So anyway, it's very tricky -- but I've got open source software that's 
_doing_ it, which while not perfect works well enough to provide 
significant improvement to your library offerings.  I've been frustrated 
at not being able to explain to people why what Umlaut does is exciting 
and important in the first place, perhaps this thread will help.  (Or 
perhaps I am under-estimating people's interest, maybe they are 
interested, but Umlaut is currently just too hard to to install and 
setup. I've got plans!).