Print

Print


At 09:12 PM 1/27/2002 -0500, you wrote:
>We worry that this leaves "orphaned" a scanning phenomenon we encountered
>in American Memory: "the billboard job."  For very large posters, for
>example, we sometimes made several rows of images, representing tiles of
>the oversized original.  In a perfect display system, these would come up
>in a grid:
>image.a1   image.a2   image.a3
>image.b1   image.b2   image.b3
>image.c1   image.c2   image.c3
>
>Question: can we use the "area" elements to do this, or is there another
>trick we didn't notice?  If not, perhaps this is a METS version-two
>desideratum.
>

Well, you'd probably end up using <area>, but not in the way I think you want.

Since presumably in the perfect display system you'd want all of your images
to display at once to comprise a unified image of the poster, the structMap
for this would probably consist of a <div> for the poster, a <par> element
to indicate that all of the items identified in the <par> should be
played/displayed
simultaneously, and then <area> elements identifying each file that should
comprise the poster.  But the <area> elements would only contain the
fileid attribute; you're using the entirety of the image identified in
area, so you
don't need to use any of the other attributes to identify a section of the
image file.

What you really want is the ability to control layout, and we agreed to differ
that at the original METS meeting last year, so yes, I'd say this is a
version 2
issue.

One alternative would be to have a more complicated structMap, e.g.,

<div type=Poster>
         <div type=upper left>
         <div type=upper middle>
         <div type=upper right>

etc., etc., and then write your display software client with the smarts to know
that when it sees a <div> of type=Poster to expect a series of subsidiary divs
that will specify where each image should be placed.  This is obviously an
ungraceful hack, but it would at least provide what you want and correctly
identify the various pieces/locations.  Hopefully, if we update METS to
add layout facilities, you could do an XSLT transformation to this latter case
to move the layout information in this example encoding into its appropriate
location.

Jerome McDonough
Digital Library Development Team Leader
Elmer Bobst Library, New York University
70 Washington Square South, 8th Floor
New York, NY 10012
[log in to unmask]
(212) 998-2425