On 03/14/2013 04:35 PM, Tom Johnson wrote:
> offering *structured (open?) data*  about every book.
>
> Other projects stop well short of this goal. Wikipedia, for instance, 
> has notability rules.

Yes, going beyond the notability rules of Wikipedia
in providing a better, free catalog, could be a good
cause. However, the notability rules have a good
reason in that it is hard to separate facts from
forgeries in very obscure subjects. If someone adds
information to OpenLibrary, how do we know that
this book really exists, and isn't just spam noise?

Wikipedia requires both notability and verifiable
source references. Normal library catalogs have
their holdings as their reference. If anyone doubts
any fact in the catalog, you can visit the library and
look at the physical book on the shelf. Even Worldcat,
that aggregates records from many libraries, can
point to each source library, where the holdings
contain physical books. The trace is not lost.
But for OpenLibrary, there are no holdings of
physical books, just the catalog record. So how do
we verify any knowledge, if we have no criteria
that need to be met before a record is created?

As a corollary, it would be pretty easy to fabricate
images that look like a scanned book, but contain
any kind of forgeries, and upload it to the Internet
Archive. I don't know if anyone has tried this, but it
could be an interesting exercise.


-- 
   Lars Aronsson ([email protected])
   Project Runeberg - free Nordic literature - http://runeberg.org/


_______________________________________________
Ol-tech mailing list
[email protected]
http://mail.archive.org/cgi-bin/mailman/listinfo/ol-tech
To unsubscribe from this mailing list, send email to 
[email protected]

Reply via email to