Hi, Karen.

If I were viewing an OpenLibrary book webpage in my browser, and
wanted to download the source MARC record, I'd want a direct link to
the raw MARC (i.e. exactly the same results I would get by following
the curl instructions in your link below).  So the OpenLibrary site
would simply need a script to run the curl request and output to my
browser.

The case of multiple source records may be a bit trickier, since the
relevant records probably aren't adjacent in those massive .dat files
at archive.org.  So the OpenLibrary script would need to make separate
curl requests, and join them together with the MARC record terminator
character (hex 1D), before sending them back to my browser.

(Of course, others might prefer MARCXML records, but that should be
easy to add once the above is already in place.)

Keith


On Fri, May 16, 2008 at 11:49 AM, Karen Coyle <[log in to unmask]> wrote:
> It appears to be possible to link from the Open Library "book" page to
> the MARC record or records that contributed to the data. Right now the
> display of MARC records looks like what you see here:
> http://pharosdb.us.archive.org:9898/show-marc/marc_records_scriblio_net/part29.dat:504453:573
>
> Since clearly some people will want to download records into their own
> systems, I'm concerned that this display doesn't facilitate that. But I
> don't know what library systems expect or can handle. So could any of
> you advise me on this?