---------------------- Information from the mail header -----------------------
Sender: American Memory Fellows <AMFELLOWS@RS8.LOC.GOV>
Poster: "Judith K. Graves" <jgrav@LOC.GOV>
Subject: Re: New LOC look
-------------------------------------------------------------------------------
Hi Everyone,
Here's a quick reply to part of Frances' post:
>
>Here's a naive question. Why is it that standard search engines don't pick
>up individual American Memory items? I'm just curious. Is there some sort
>of firewall? Or are individual items inaccessible to outside software in
>the same way online catalog records are inaccessible outside the the
>software they live in? I can see all the potential drawbacks -- the hit
>and miss rate would be so random and unpredictable. On the other hand, it
>would be nice if there were a way for American Memory records to be
>retrieved by folks who don't happen to know of the collections' existence,
>but need them. Does this make sense?
>
>Frances
>
The short answer is the software. The pieces of the item record (bib
record) that accompany each item are stored in massive databases and are
assembled for each search request. All the items that match the search
request get stored in a temporary directory (you'll see /temp/ in the URL)
that lives for about an hour or so before evaporating. (That's also why
you can't link to an item record by bookmarking it in your web browser.)
There is no way that all the millions of bits of data that are needed to
create a project of this size could be maintained. Can you imagine
changing the date on a collection (this has happened when new items are
added to existing collections) for 25,000 web pages in that collection?
Multiply that by the number of collections - which keep growing - and you
have a sense of the time it would take to maintain millions of pages.
Automating is the only answer. (There are other issues, on an
international scale, about how data is represented for searching by these
complex search engines.)
What can get indexed (and has) by outside sources are all the homepages,
special presentations, and other related info supporting the collections.
Judy
-------------------------------------------------------
Judith K. Graves
Education Resources Specialist
National Digital Library Program
Library of Congress
Washington, D.C. 20540-1320
jgrav@loc.gov
(V)202/707-2562 (F) 202/252-3173
http://lcweb2.loc.gov/ammem/ndlpedu/
This archive was generated by hypermail 2b29 : Tue Jan 04 2000 - 12:24:17 EST