Obscure library classification schemes

There's an interesting debate going on in the LibraryThing Thingology blog (the theoretical LT blog) about classification systems and their electronic availability. If I ever had a free hand to choose a classification scheme for a general collection, untrammelled by the legacy of any previous system, I'd be very tempted to use Colon, or Bliss.
But a mention of an obscure one, the Blegen classification, for classical studies libraries, reminds me of some of the more obscure ones I have used in my time. When at the Royal College of Veterinary Surgeons we used Barnard, for which there is no Wikipedia entry. I may write one. I used to visit the Marx Memorial Library, which has its own system, which I was once told was based on that used at the Lenin Library in Moscow.
Classification and taxonomy are in the news. See the Guardian:
In praise of taxonomy
Guardian 12 August 2006 (though I thought that the ICZN taxonomy had been available online for some time)
And in the London Review of Books a fascinating investigation of medical and psychiatric taxonomies by the philosopher Ian Hacking:
Hacking, Ian
Making up people
London Review of Books 2006 28(18): 23-26
(Online version for subscribers only)
Hacking gave the tenth British Academy lecture in April this year, Kinds of People: Moving Targets, on much the same subject.