[tei-council] remaining EEBO TCP issues

[Paul Schaffner]

On Mon, Oct 21, 2013, at 15:26, Sebastian Rahtz wrote:

> does that mean you'll change your sources? or do you expect
> people like to me to make the change afterwards?

Could do. Haven't yet. But yes, changes (at present only a few
changes, such as the corrections that Martin Mueller sent me this 
morning) happen to the earlier files all the time.  Every new release
is a re-release of all the earlier files, including some with (usually
very minor) changes. And no, these are not (yet) being seriously
tracked.

Dealing with a similar issue this week and last (and next too,
probably),
which is why I've been silent lately: head down, focusing on the
matter of re-generating headers for all 131,000 EEBO items*, without
losing any of the corrections that we've made to ProQuest metadata,
or any of the de-duping and correcting that ProQuest itself has 
made to the MARC that it releases. An 'overlay' problem, as they say
in the library world. 

*That is, all the unique combinations of bibliographic record and
image-set identifier: only about half of those will ever have to map
to a TCP file, but all of them are mapped to a TCP identifier, since
we create the headers and the IDs first for everything we might
ever transcribe, rather than transcribing first and then describing
them.

pfs
-- 
Paul Schaffner  Digital Library Production Service
PFSchaffner at umich.edu | http://www.umich.edu/~pfs/