On 07/30/13, john g marr wrote:
> What's the underlying "working hypothesis" of your proposed system? Not the purpose of the system, but the rationale for it.
My working hypothesis is that exploitable document structures and internal metadata are of more use to members of the literary community than the mainstays of the metadata we provide, i.e, author, title, and subject. It's easier to define "internal metadata" by example. I'm using Aesop's fables (and a little from LaFontaine) as a test bed. The internal metadata would be things like marking every mention of an animal, every instance of a speech by a character, and similar elements. This would allow searches such as, "Show me every fable in which a wolf is a speaker."
This is based on my research into the use of information technology by members of the literary community. They have little interest in a librarian asserting that, "This fable is about Patience, this one is about Loyalty...," and so on. Either they already know that, don't care, or want to build their own interpretation. Sorry, but that's what the evidence suggests. Much of what they want from metadata is either above the level of a BBLO (Book or Book-Like Object), or below. So they might want different versions of the same fable by different authors, or all fables in which a moral is a speech by a character, as opposed to a statement by the narrator.
I'm working from a base of XML documents, and automating as much of the workflow as I can through various programming tools (XSL, XQuery, and some Java applets). I actually have a fair amount of working code, but nothing to put online yet. I may need to switch hosting services, or re-write some of the code in PERL.
As for the rest, I re-invoke Godwin's Law and the concept of bikeshedding, and opt out of that part of this thread.
Steve Paling
Received on Tue Jul 30 2013 - 15:32:47 EDT