catalog system

thistlechick · Post by **thistlechick** » March 15th, 2006, 12:53 pm

Hath, have you been given access to the "Validator" yet? Let me know if you want to take a look at what we already have in place there.

alexfoster · Post by **alexfoster** » April 8th, 2006, 10:42 am

Hello, sorry to plough in after this has already been under discussion for months. Will the cat system be able to produce a progress summary page, eg something like this:

Ie for any given list of books (my projects, books I'm co-ordinating, all books, all incomplete books) it can produce an 'at a glance' table, preferably with funky colour coding rather than text, showing the status of each chap. My letters are for allocated, completed, verified.

What does the panel think?

thistlechick · Post by **thistlechick** » April 8th, 2006, 11:12 am

Alex, I think that's a great idea! It seems that once we have the data in a database, we could present the information in a format like you suggest.... or in many different formats.... it's just a matter of getting a database and getting the data into it... =)

alexfoster · Post by **alexfoster** » April 8th, 2006, 1:56 pm

Erm, yeah. Good point. So, good luck with that :)

Alex

Hath1 · Post by **Hath1** » April 10th, 2006, 5:11 am

thistlechick is exactly right. We are taking the time to attempt to create the entire system in an extremely simple and logical manner that will allow future "feature requests" to be added to the system quite easily in MOST cases. Movement is slow at this time on the project, but we're still working on it (I haven't been able to get to it the last couple weeks with working too much overtime and having things to do on weekends, but I'm sure I'll get back to it soon) (although it looks like I may be traveling with my family this weekend . . .)

Aldark · Post by **Aldark** » April 10th, 2006, 5:27 am

I have an issue at work where I'm the database admin for a database that was started before me by a husband of a wife who worked in our office. While the database is great it was also pieced together over a 5-yr period without great documentation.

Now we are trying to reformat the database into a web-based interface and it has been tough trying to break it down and follow all the connections between tables/queries/reports/etc.

FUN! (NOT!)

harvey · Post by **harvey** » April 15th, 2006, 8:03 am

How do cataloging systems handle the display of "messy" data?

I'm wondering how cataloging systems handle messy, real-world
information in terms of display to the user.

An example is different names, as in pseudonyms, or different forms
of the same name, for the same person. The LibriVox catalog already
contains at least one work, Spirits in Bondage, written under a
pseudonym, Clive Hamilton, by a famous author, C.S. Lewis. I know
from music that the same person or group sometimes uses a different
form of their name on different recordings, such as Ed Smith and
Edward Smith; I cannot image this doesn't happen for books.

The opposite side of this situation is different people with the same
name.

From music, I know of a combined case. There's a singer-songwriter
born Leslie Phillips. She recorded several albums under that name, as
well as several albums under her nickname, Sam Phillips (and acted in a
major movie as Sam). There was also a record producer named Sam
Phillips (a different person).

My query is this. If I search a cataloging system for C.S. Lewis, do
I typically get back everything Lewis wrote that's in the catalog,
including under pseudonyms? And how is it indicated that Lewis and
Hamilton are the same person?

Similarly -- and ignoring this is music, if I search for Sam Phillips,
do I get back works by Leslie Phillips, too? But I'm more interested
in how it is indicated that "Sam" and "Leslie" are the same person.
And in how the catalog system differentiates between the two different
Sam Phillips (especially if I don't know there are two and don't know
much about the one I'm searching for).

I ask this because I've given a lot of thought to designing a database
for music albums, and I could never figure out satisfactory ways --
particularly in the context of the display of search results -- of
indicating that different or similar names are the same person and
distinguishing between different people with the same name.

I expect these difficulties have been solved, but I've never before had
the opportunity to ask trained catalogers "how is it done?".

LibraryLady · Post by **LibraryLady** » April 15th, 2006, 9:08 am

It is done with authority records. This makes sure that all versions of a person's name will point users to other versions as well. Different people with the same name are usually distinguished by date of birth but sometimes other qualifiers such as "musical group" are added as well.

See here for more info: http://en.wikipedia.org/wiki/Authority_control

thistlechick · Post by **thistlechick** » April 15th, 2006, 9:19 am

I've written in this thread about this topic .... and since I can write pages and pages about it because I love it so much (seriously), I think I'll save you from having to read it here and I'll just send you to some resources online and let you investigate it more for yourself... Have fun =)

In library systems we use authority control... you can view the Library of Congress Authority File here: http://authorities.loc.gov/

Here are some basic definitions to get you started:
http://ublib.buffalo.edu/libraries/units/cts/ac/def.html

And here's the Wikipedia article on the topic:
http://en.wikipedia.org/wiki/Authority_control

EDIT: hehe... and it looks like Annie managed to sum it up nicely while I was busy writing, editing, rewriting, deleting, and finally giving up and just giving the same link that Annie posted while I was doing that... hehe... Thank you Annie =)

LibraryLady · Post by **LibraryLady** » April 15th, 2006, 9:22 am

Jinx!

LibraryLady · Post by **LibraryLady** » April 15th, 2006, 9:25 am

LOL, I did the same thing, Betsie, started trying to explain it myself, rewriting, editing, deleting, and finally just posting the wikipedia link. You found some more links that I did so I'm sure Harvey will have plenty to keep him busy, learning the deep dark secrets of catalogers.

harvey · Post by **harvey** » April 15th, 2006, 1:37 pm

Annie and Betsie,

Thanks for your replies. I looked at the links you provided.
Reminded me of my first Internet programming job: merging 16
bibliographies stored in a dozen different software programs into a
unified format for on-line searching.

I looked up Sam Phillips in the Library of Congress (LoC) authority
control system:

It's interesting that the authorized form for Leslie Phillips is
not her given name, but her nickname, Sam Phillips. But it does list
Leslie Phillips as one of the reference names.
Either this system is inconsistent or there are arcane considerations
not readily apparent: The authority heading for Sam Phillips, record
producer, gives the date of his birth (1923) but not his death, even
though the MARC record lists the DoD. However, the authority heading
for Sam Phillips, singer-songwriter, gives no dates.
I even discovered a trivia fact I didn't know: Sam Phillips is
the producer who discovered Elvis.

However, these references do not seem to address the main thing I want
to know. They cover a mechanism for organizing the information, not
how it's displayed to the end-user. (On second though, perhaps it was
a mistake asking gung-ho specialists (:-) The issue is how to indicate
to users, especially in a set of results from a search (a set of catalog
records) which lists works by both Sam Phillips and Leslie Phillips,
that both names refer to the same person. I'm thinking of something
that's immediately clear, like the statement "Sam Phillips and Leslie
Phillips are the same person", rather than links to authority control
records. That is, something that's user-friendly.

Or do cataloging systems with authority control replace all references
with the authority heading version of the name. In that case, if the
user only knows that Leslie Phillips recorded some albums, but not
their names, and, from a search, gets works credited to Sam Phillips,
how do they know they actually got what they are searching for (ie,
works by Leslie Phillips)?

---

After writing the above, I went to the Library of Congress Web site
for some instructive searches. The first was with SONIC, a catalog of
sound recordings:

Searching for "Phillips Sam" returned 10 records:
- Half of items in the list contained no mention of "Sam Phillips".
- There was no unambiguous indication in either the partial or full records
  that there are two people named Sam Phillips.
- One of the records, for a work by the singer-songwriter (born
  in 1962), gives the name as "Phillips, Sam, 1923-", which is the date
  the record producer was born. (That is, this record gives the wrong
  Phillips.)
- None of the records are for works by Leslie Phillips.
Searching for "Phillips Leslie" returned 1 record:
- No records were returned for works by Sam Phillips.

So now I'm wondering what good the LoC's Authority Control is, if the
search system performs this badly?

Then I tried the main "Library of Congress Online Catalog", searching
for "Phillips Leslie". This appeared to have returned every record
for anyone with the last name of Phillips, starting with Leslie (what's
the point of this?). However, the one or two Leslies listed are not
the one I'm after. I had to click the "next" link about a dozen times
to get through the alphabetical listing to Sam Phillips. There I
found the works done by her as Leslie, but listed under "Sam". My
questions about this are (1) how is a user supposed to know to look
down the list for "Sam" when all they know about is "Leslie" and (2)
why should the user be expected to wade through pages and pages of
irrelevant records (ie, for hundreds of Phillipses whose first names
are neither Sam nor Leslie)?

Here's the brief record for one of Leslie's albums:

Code: Select all

                    The turning Leslie Phillips.

LC Control Number: 	91756401
 Type of Material: 	Music Sound Recording
Brief Description: 	Phillips, Sam.
	The turning [sound recording] / Leslie Phillips.
	Hollywood, Calif. : Horizon, p1987.
	1 sound disc (36 min.) : analog, 33 1/3 rpm, stereo. ; 12 in.

This is hardly very useful. For example, what in the world is the
meaning of "The turning Leslie Phillips" for what appears to be a
title of the catalog record? The full record is slightly better:

Code: Select all

                    The turning Leslie Phillips.

LC Control Number: 	91756401
 Type of Material: 	Music Sound Recording
    Personal Name: 	Phillips, Sam.
       Main Title: 	The turning [sound recording] / Leslie Phillips.
Published/Created: 	Hollywood, Calif. : Horizon, p1987.
      Description: 	1 sound disc (36 min.) : analog, 33 1/3 rpm, stereo. ; 12 in.
          UPC/EAN: 	7502107571
 Publisher Number: 	SP-0757 Horizon  WR0757 A&M Records
         Contents: 	River of love (2:43) -- Love is not lost (3:57) -- The
                        turning (3:36) -- Libera me (3:09) -- Carry
                        you (4:18) -- Beating heart (2:56) --
                        Expectations (3:15) -- Down (3:25) -- Answers
                        don't come easy (4:10) -- God is watching you (3:50).
            Notes: 	Horizon: SP-0757 (on container spine: WR0757). Lyrics on inner liner.
         Subjects: 	Gospel music.
LC Classification: 	A&M Records WR0757
 Other System No.: 	(OCoLC)17241540
  Geog. Area Code: 	n-us---
     Quality Code: 	lccopycat

None of what I saw at the LoC satisfies my wish for a clear, explicit
method for informing the user that Sam and Leslie are, indeed, one and
the same person. It seems to me the user is expected to guess. Nor
does it appear that the LoC system provides a simple, direct way to get
all works by a particular person, regardless which of their alternate
names are used for searching.

If it isn't clear already, I'm pretty dismayed by what I encountered
at the Library of Congress. I do sincerely hope the LibriVox catalog
/ database / search system works rather better.

thistlechick · Post by **thistlechick** » April 15th, 2006, 2:59 pm

forgive me for not reading every single word of your verbose message... so it's possible that i will again not actually answer your question but just make some librarian-esque comments:

The Library of Congress Authority File performs exactly as it is meant to perform.... it is not meant to be used as a front end search interface. So, you are looking at it out-of-context. It is simply a tool used by catalog librarians and is incorporated into the actual catalog.... if you go to the actual Library of Congress catalog and search for those author names, the results will point you to works by the authors where appropriate.

There are established rules/standards used in the library community as to which names will be used and how they will appear... for more information on those, you'll need to contact a cataloger at the Library of Congress. Also, note that the LOC authority file is only ONE EXAMPLE of an authority file. Any group can create their own authority file for their own purposes... this is just how they do theirs and libraries tend to consult it since it is already established.

Also, take some time to look at how Project Gutenberg directs users to the approriate author names for an application nearer to that of Librivox. Example: Search for these two names to see how the "See" reference works:

Nicolson, Adela Florence Cory

Hope, Laurence, 1865-1904

... if you're interested in pursuing your Graduate degree in Information and Library Science, I'd be happy to recommend a university program or two.

harvey · Post by **harvey** » April 15th, 2006, 6:27 pm

Betsie,

Thanks for your efforts to educate me in the intricacies of library
science.

>The Library of Congress Authority File ... is incorporated into the
>actual catalog.... if you go to the actual Library of Congress
>catalog and search for those author names, the results will point you
>to works by the authors where appropriate.

But that's exactly what I tried to point out in my too-long post that
the Library of Congress catalogs do not do. The two I used either did
not return all works by an individual, regardless of what name I used
(from the authority file record for that person), or required me to
know the authorized name to locate the works when I searched for a
reference name. Neither are what I expect based on what you say.

I looked at the Project Gutenberg site. It has some of the same
problems as the Library of Congress catalogs: mainly, that it doesn't
make explicit the relationship between different names for the same
person. In fact, in one way it's worse: at least on this one page,
the record for Spirits in Bondage doesn't mention that it was
published under the pseudonym Clive Hamilton, listing instead only
C.S. Lewis as creator.

I'm beginning to feel like I'm missing something significant in this.

Greg McMullan · Post by **Greg McMullan** » April 17th, 2006, 10:27 am

I don't have a lot of time, but it might be interesting for folks to look at http://www.qwikly.com/2006/02/wikipedia-for-cataloging-books.html which seems to be talking about a place where folks can catalog their own libraries of books in a shared environment, with an API that relies on Library of Congress data. I will for sure be looking more later, when I'm not running out of lunch break.