Amazon full-text book index [was: Re: a python book hint]

[John J. Lee]

claird at lairds.com (Cameron Laird) writes:

> In article <OkPsb.16752$9_.600184 at news1.tin.it>,
[...]
> I can't now do Amazon's text-search capabilities justice; they are
> making for a sea change of the scale of Google or VoIP or other
> "epochal" innovations.  Briefly, if there's a book that interests
> you, and Amazon has it, it's likely you can bring up any of its
> pages on-line.  You can sample it freely.
[...]

I'm told (bookpeople mailing list) they scanned and OCRed them all,
even though books are now always (well, almost always) in machine-
readable textual form at some stage in their preparation.  A bit sad
in terms of duplication of effort, though another recent bookpeople
post told the story of somebody's personal first "perfect" OCR of a
book: fewer errors in the scan than in the original!

The ability to search books is certainly interesting and new (I've
tried to find even a tables-of-contents database for books before, and
the university librarian told me that no such beast existed).  Maybe
it's even as important as Google, dunno.  It will be missing two
central parts of Google, though: instant, zero-marginal cost access to
the content, and linking information.


John